Shewanella oneidensis and Methanosarcina barkerii augmentation and conductive material effects on long-term anaerobic digestion performance

Abstract

This study explores the use of conductive material in scaling up anaerobic digestion for enhanced biogas production. Focusing on Direct Interspecies Electron Transfer (DIET), the research employs a syntrophic DIET-able consortium formed by Shewanella oneidensis and Methanosarcina barkerii in 3.8-L experiments utilizing reticulated vitreous carbon (RVC) as conductive material. In short-term tests with acetate the syntrophic co-culture with RVC resulted in 86% higher maximum velocity of methane production, while in long term with real feed 13% increased rate was observed: the addition of 1.77 (S/m)*m² RVC resulted in a faster methane production of 2.39 mL/gVS*h compared to 2.08 mL/gVS*h of the reference. The experimental conditions of syntrophic inoculum and RVC as conductive material gave a benefit in terms of process rate compared to the reference, considering the inoculum fate, Methanosarcina barkerii was among the dominant taxa at the end of the experiment, while Shewanella oneidensis was outcompeted. Among the methanogenesis production pathways, an increase of hydrogenotrophic methanogenesis has been observed in presence of conductive material. Further research is needed to understand the role of RVC in sulfur compounds production. Utilization of RVC to augment methane production yielded interesting results for real-scale application. As an added carrier, RVC remains unaltered and can be readily recuperated and reused multiple times.

Introduction

Anaerobic digestion (AD) is a microbial-based process well known and applied for centuries, to convert organic material in the value-added byproduct biogas. The initial complex organic compounds, by the action of diverse microorganisms, are first hydrolyzed in smaller molecules and then are converted stepwise through acidogenesis and acetogenesis in the metabolic precursors of methane [2].

In the challenge of decarbonization policies, AD is still one of the most promising technological options to displace coal: comparing the same energy output of coal, methane emits in fact roughly three times less CO₂, less than a tenth of sulfur oxides, a quarter of nitrogen oxides, and essentially no particulate matter or heavy metals compared to coal [29].

Therefore, there is high interest in optimizing the process performance to improve the overall energy balance by selecting industrial residues as feedstock with specific pre-treatment technologies [5, 16, 33], phase separation [14, 18, 36, 37, 39] and optimizing the process management [35].

In AD, syntrophic association between Bacteria and Archaea plays a major role in the metabolic conversion of organics to methane [4]. Different microbial taxa are involved in the methane production by many sequential reactions. The overall process performance is in fact, based on the equilibrium between fermenting bacteria (acid- and aceto-genic) and methanogen archaea. One mechanism supporting the syntrophic association is the exchange of electrons between fermenter taxa and methanogens known as Interspecies Electron Transfer [3, 27]. DIET acronym for Direct Interspecies Electron Transfer, refers to this sharing of reducing power, without any molecular carrier, but through physical connections among different microbial taxa. The two functional groups involved are the electron donors and acceptors, defined, respectively, as exoelectrogens and endoelectrogens [18]. Among the many microbial taxa potentially involved in the first part of the process, i.e., acidogenesis and acetogenesis, many Bacteria have been reported as exoelectrogens: Geobacter sp., [26, 31, 45], Shewanella sp., [41], Clostridium sp. [23] and Thauera sp., [42]. For the methanogenesis phase, there are three different pathways possible to produce methane: the acetoclastic, the hydrogenotrophic and the methylotrophic. Out of the three methanogenic pathways, the methylotrophic is the less found in biogas plants and the acetoclastic is usually the most reported by AD operators [7]. High interest is placed on the hydrogenotrophic pathway as by reducing carbon dioxide, could contribute to atmospheric CO₂ reduction mitigating climate change.

In the last decade, several scientific papers have reported the addition of conductive material (CM) to increase the performance of AD [28, 30, 48]. Among these studies, many focus on direct or mediated electron transfer detailing the process conditions, the microbial strains involved and the benefit in terms of process performance and methane production [18, 19].

Although the mechanism of electron exchange is not yet fully clarified [40], many different CMs have been shown to have a positive impact on the process performance [8]. Among the benefits cited, increased cumulative methane production, shorten of lag phase, increased rate of conversion to methane were reported [43, 46].

It must be noted that most of the studies published are however based on small scale lab tests, using alcohols as feedstock, and monitored for short-term periods (usually less than 15 days). Although recent published results are based on real feed (inter alia [38], DIET-based AD optimization experimental data from upscaled applications and combined with real and continuous feed are relevant for the scientific community. Among the questions still not fully addressed and of interest from an industrial application point of view, are the feasibility of using DIET-able microorganisms as bioaugmentation and the maintenance of the benefit of the CM presence over the limited experimental time period (Q. [47]).

In this manuscript we tested a co-culture of Shewanella oneidensis and Methanosarcina barkerii for syntrophic methane production and evaluated the feasibility to use it as syntrophic DIET-able consortium in AD with real feed and reticulated vitreous carbon (RVC) as CM addition over 100 days process. Shewanella oneidensis has been selected as known and model electrogenic organism [6] while Methanosarcina barkerii for its high versatility in methanogenic pathways. Shewanella sp. involvement in DIET mechanisms has been chosen as already proved by [20, 44] and Methanosarcina sp. is a taxon reported in DIET studies in AD [12]. In a real case application, the conductive carrier would have a biofilm of an electrogenic bacterium involved in the first steps of AD, and the methanogenesis would have been completed by the methanogens already present.

The novelty of the present work lays in the confirmation of the advantages of DIET-based AD in real conditions, in particular real sludge feed and for 100 days monitoring and the fate of a DIET-able inoculum. These aspects are not yet fully addressed in literature and are of importance for the assessment of the Technology Readiness Level (TRL) of DIET-AD optimization.

Material and methods

Experimental design

In the present paper therefore, we start from short-term test data to confirm the two pure cultures’ ability to produce methane syntrophically with RVC as CM addition. Then we evaluate the CM addition effect in terms of process performance and microbial community dynamic, and the fate of the syntrophic inoculum in medium size reactors on long-term (100 days) process. Thickened sludge was collected in a 2-week interval and used as feed in the experiment to mimic substrate variability. Biogas production tests were carried out in 3.8-L capacity reactors, the headspace was replaced with pure nitrogen to assure anaerobic conditions. Two sequential experiments were planned.

First experiment set: syntrophic methanogenesis

To evaluate the suitability of Shewanella oneidensis and Methanosarcina barkerii as DIET-able inoculum and therefore the ability of to convert syntrophically acetate in methane by exploiting the CM presence, three identical reactors were set with synthetic medium (Medium DMS120 modified by substituting lactate with acetate and methanol) as substrate and 14 days monitoring. The reactors were named SM, SM RVC and SM RVCbio and designed as follows: SM test had 2000 mL of inoculum prepared with Methanosarcina barkerii and 60 ml of Shewanella oneidensis cultures, SM RVC test had the same inoculum and sterile RVC carrier, SM RVCbio was equal to SM RVC, but for the use of carriers with grown Shewanella oneidensis biofilm on them. RVC was chosen as conductive material as it has been already used in the growth phase of Shewanella both in this experiment and also in the past [36, 37].

Second experiment set: thickened sludge feeding

To evaluate the interference of indigenous microorganisms of the digestate toward a DIET-able inoculum and the stability of the improvement of the process due to CM addition, a second experiment was designed in a real condition setting.

Three replicate control reactors (named AD) were set with digestate as inoculum and sterilized thickened sludge as substrate in a ratio 1:1 on organic matter basis; three replicate test reactors with RVC carrier bioaugmented with Shewanella sp. biofilm and 50 ml of pure culture Methanosarcina barkerii, referred as AD MRVCbio.

The digestate consisted of 29.5 ± 0.1 g/L total solids (TS) from which 59% of organic origin expressed as volatile solids (VS). The thickened sludge contained 24.2 ± 0.1 g/L TS from which 81% organic matter. Reactors capacity was 3.8 L with 2.1 L of headspace.

Reactors were run in fed-batch mode with feedings every 14 days for more than 100 days. The digestate as inoculum and the thickened sludge as substrate and feed were sampled from the municipal WasteWater Treatment Plant (WWTP) of Chiasso (CDA, CD, Vacallo, CH).

Syntrophic inoculum

Methanosarcina barkerii (DSM800T) was cultured in anaerobic conditions with 80% N₂ and 20% CO₂ on the DSM120 medium modified by substituting sodium acetate and methanol with lactate for the first experiment and DSM120 medium with no modification for the second experiment. Methanosarcina barkerii was incubated at 37 °C for 5 days before inoculum, reaching OD₆₀₀ = 0.6.

Shewanella oneidensis biofilms were grown on RVC as a support for later insertion into the methane cultures. RVC is a redox resistant conductive foam and the pore size used was 100 PPI (purchased by ERG aerospace corporation, CA, USA) [10]. To grow a biofilm of Shewanella oneidensis MR-1 on it, a 1-L dual chamber microbial fuel cell with Nafion membrane N117 (Ion Power, Germany) was used [36, 37]. A preculture of Shewanella oneidensis MR-1 (ATCC 700550, Gescher Group) was developed using Luria Bertani (LB) medium, which was prepared from 1 L demineralized water, 10 g tryptone, 5 g yeast extract and 10 g NaCl. This preculture was used to inoculate the anolyte of the MFC. TSB medium (Tryptone Soy Broth) was used as anolyte. It was a modified version in which dextrose was replaced by L-lactate. 1 L TSB medium was prepared from 1 L deionized water with 17 g tryptone, 3 g soytone, 2.24 g lactate, 5 g NaCl and 2 g K₂HPO₄. Lactate was added during cultivation without replacing the present anolyte. A 50 mM potassium ferricyanide (K³Fe(CN)₆) solution was used as catholyte. This solution was made from 1 L DI water, 16.8 g K₃Fe(CN)₆, 5 g KH₂PO₄ and 10.9 g K₂HPO₄.

Four rectangular solid shaped carriers were added for the first experiment (SM RVC and SM RVCbio) for a total volume of 273 cm³ for each reactor. Two carriers with the same total volume were added to the second set of tests (AD MRVCbio). The RVC used for both the experiments had an electrical conductivity of 0.99 S/m, an available surface for microbial attachment of 1.79 m² and concentration of 8.78 g_RVC/L for both experiments.

Samples and sampling

For the digestate feeding test (second experiment), samples were collected from each replicate reactor. Sampling for microbial community and physical chemical characterization was carried out before every feed in sterile conditions.

Biogas production and composition

Biogas production was assessed manometrically in the reactors recording values every hour (Keller Leo2, Winterthur, CH). Subsequently, the recorded pressure values were converted to biogas volumes under standard conditions (273.15 K, 1 bar) applying the ideal gas law. Data were normalized on VS-basis for comparison.

Methane content and gas composition were measured at fixed intervals during the experiments. For each cumulative curve at times corresponding to plateau phase and flex point, CH₄ and CO₂ volumes were reported. Flex points were identified by zeroing the second derivative the fitted cumulative biogas production curve. Methane and carbon dioxide concentration in biogas was determined using infrared (IR) analysis with the Gas Analyzer (Biogas5000 Geotech, Lauper Instruments AG, Murten, CH).

Physical chemical characterization

TS and VS were determined in accordance with the Standard Methods (American Public Health Association APHA, 2016). Chemical oxygen demand (COD) and total nitrogen were determined, respectively, by Hach Lange LCK514 and LCK338 cuvette test (Hach Lange, Düsseldorf, DE).

Microscopic characterization

Biofilm characterization was observed by Scanning Electron Microscope (SEM) (Phenom, Thermo Scientific, Waltham, USA). RVC carriers with biofilm were fixed with 4% glutaraldehyde and 2% osmium tetroxide. After fixation, samples were sequentially dehydrated in ethanol and acetone series, dried by critical point dryer (EM CPD300, Leica, Wetzlar, DE) and gold coated to allow optimal observation under vacuum preserving the biological texture (Smart coater, Jeol, Peabody, USA).

NGS analysis

Samples were collected at specific time points: namely 0, 333, 669, 1341, 2014, 2354 h, from the reactors in the second experimental setting. Metabarcoding was conducted following the protocol described by (König et al., 2022b). DNA extraction was performed using the DNeasy PowerSoil kit (Qiagen), followed by PCR amplification with primers 1492R (5'-ACT TGC CTG TCG CTC TAT CTT CGG TTA CCT TGT TAC GAC TT-3') and 21F (5'-TTT CTG TTG GTG CTG ATA TTG CTC CGG TTG ATC CYG CCG G-3') for Archaea, and with primers 1492R and 27F (5'-TTT CTG TTG GTG CTG ATA TTG CAG AGT TTG ATC CTG GCT CAG-3') for Bacteria. Amplicons from the same sample were merged and assigned to the same barcode.

Library preparation was carried out using the SQK-PBK004 kit (Oxford Nanopore Technologies), and sequencing was performed on a MinION^™ Mk1B flow cell R9.4 (Oxford Nanopore Technologies) for 72 h.

16S rDNA data processing

The raw fast5 reads were base-called and demultiplexed using ONT Guppy basecalling software (v5.0.7) resulting in the production of fastq files. Subsequently, the quality of these files was assessed using LongQC [11]. Taxonomy assignment was carried out using the Dada2 R package (v1.18) in conjunction with the SILVA 138.1 database ("sh_general_release_dynamic_s_10.05.2021.fasta"). The "filterAndTrim" function was executed with specified parameters, including minLen = 1400, maxLen = 1600, multithread = TRUE, and verbose = TRUE. Following this, the "dada" function was employed with the parameters OME-GA_A = 1e-10 and DETECT_SINGLETONS = TRUE. Finally, the "assignTaxonomy" function was utilized with the argument tryRC = TRUE for taxonomy assignment. Sequences were deposited to NCBI Sequence Read Archive as Bio Project PRJNA1075654.

Data handling and statistical analyses

Physical chemical characterizations were performed in triplicates unless otherwise specified. Mean and standard deviation were calculated for each analysis. The cumulative curves of biogas production were fitted with nonlinear model using Gompertz modified equation (Lay et al., 1997) using GraphPad Prism (version 10.2 for Windows, GraphPad Software, La Jolla, California, USA, www.graphpad.com). The least square regression method was employed to derive the kinetic parameters, and the goodness of fit was quantified using R-square. This equation provided values for the ultimate biogas volume (Y_m), maximum biogas production rate (R_m), and lag time (l) as described in [19]. R software version 4.1.1 (R Core Team, 2021) was used for statistical analysis and visualization. The microeco package ([24, 25]), was used to analyze the microbial community dynamics during the tests. Alpha and beta diversity based on Shannon and Bray Curtis distance were calculated to characterize the communities. Relative abundance of the taxa over time and different experimental conditions were calculated and expressed using a correlation heatmap, the relationship between environmental parameters and taxa were assessed by redundancy analysis, following the analysis pipeline described in ([24, 25]).

Results and discussion

The use of conductive material to trigger a DIET effect to improve AD and methane production is widely reported in literature [1, 24, 25]. This is undoubtedly an interesting approach that need to be carefully analyzed before being applied at real scale. In fact, most of the papers published recently on this topic are limited to lab scale conditions [30]. Among the critical variables that could negatively impact on full-scale application of the DIET approach, are in fact volume reactors, retention time, feeding characteristics, use of electrogenic inoculum and indigenous microbial community in the digester.

At first the two pure cultures’ ability to produce methane syntrophically with CM is assessed in short-term experiments; then with real feed in presence of CM, the fate of bioaugmentation inoculum, process performance and microbial community dynamic were evaluated.

Inoculum ability to produce biogas and methane

The capability of Shewanella oneidensis and Methanosarcina barkerii in presence of RVC as CM to convert the acetate in biogas was investigated in batch mode by studying the kinetic parameters of the cumulative biogas production curves.

To start the experiment with a grown biofilm of Shewanella oneidensis, RVC was chosen as carrier (see material and methods section) with acetate as carbon source. The nonlinear fitting of the experimental data of the three tests is reported in Fig. 1.

Fig. 1

Non-linear fitting data for the three experimental setups with three replicates for each test, Gompertz mod. equation, and least square regression. R² > 0.81

The biogas cumulative production for the three tests showed different trends: the reactors with CM addition (SM RVCbio and SM RVC) showed a rapid increase in biogas production reaching the plateau around 100 h, while the reactor with just the syntrophic inoculum (SM) showed a less steep slope reaching the plateau at around 2 weeks (336 h). The exponential phase recorded for the presence of CM can be explained by a fast conversion of the carbonaceous compounds in the medium by the microorganisms in a syntrophic association. The slow increase in biogas production in the reactor with the microorganisms in suspension can be explained by the inoculum composition: the two taxa caused a change in the shape of the biogas production curve with an exponential phase, showing higher metabolic activity; the curve for the bioaugmented reactors with no CM shows a slower increase that however, reaches higher max cumulative production in the timeframe of the experiment (Fig. 2).

Fig. 2

Comparison of a max rate of biogas production (rm) and b max cumulative biogas production. One-way ANOVA, Tukey’s post hoc, one asterisk (*) identifies adjusted P values between 0.01 and 0.05, four asterisks (****) identify adjusted P < 0.0001

Within the 14 days selected as short-term timeframe, we confirmed the positive effect of CM on biogas and methane production kinetic. As the inoculum was added in sterilized conditions, the fact that biogas and methane were produced proved that the syntrophic cooperation between Shewanella oneidensis and Methanosarcina barkerii in methanogenesis did occur. Considering the reactions involved in the conversion of acetate to methane, the last phase (methanogenic) resulted for all the tests the limiting factor as proven by the experimental data fitting with mod.Gompertz equation compared to both logistic and first order [9]. CM presence does not affect which process phase is limiting which remains the methanogenesis.

Studying the biogas composition, the two biogas production curves with exponential slope (SM RVC and SM RVCbio) resulted in 45.8 and 54.4% in this phase. At plateau the best performing with 60.8% was SM RVC (Table 1).

Table 1 Methane percentages measured in the three experiments

The comparison of the kinetic data for biogas production showed that the test SM RVCbio had the highest rate (r_m) compared to both the SM and the SM RVC, considering then the maximum cumulative production (Y_m), the highest value was for the two taxa in suspension (SM).

Analyzing the biogas production rate (r_m) the three tests gave different results: test SM with the two microbial taxa in suspension gave the lowest value, higher was SM RVC with the conductive carrier and the highest value was recorded for SM RVCbio the test with microorganisms, conductive carrier and Shewanella sp. biofilm. The Y_m kinetic parameter comparison resulted in statistically significant different values with the SM reactor with highest value and then the two tests with RVC addition.

As the three tests were designed equal for the microorganisms’ content, considering also the Shewanella sp. cells attached to the RVC carrier, we can derive that the increased rate is due to the conductive carrier presence. In addition, the fact that the test with the added Shewanella sp. cells on RVC (SM RVCbio) gave an increased rate can support the occurrence of a DIET mechanism involved: syntrophic cells already attached on the surface, could exploit better the electric conductivity of the material. This finding is in agreement with literature as reported by [20, 30].

Therefore, considering the material characteristics of RVC and the experimental settings, the data showed that the addition of 8.78 g_RVC/L of RVC as CM with an available surface of 1.79 m² resulted in 86% higher maximum velocity comparing SMRVCbio (2.57 mL/g_VS*h) with SM (0.37 mL/g_VS*h).

Fate of syntrophic DIET-able inoculum in real conditions

The benefit of a bioaugmentation with DIET-able microorganisms in presence of indigenous microbial community and the long-term advantage of CM addition were addressed in the second set of experiments.

Biogas and methane production measured at each feed were analyzed and the kinetic parameters calculated by nonlinear fitting (Gompertz mod) are reported in Figs. 3, 4

Fig. 3

Rate of biogas production compared among feeds asterisks indicates significance (****P < 0.0001 and **P = 0.0032) one-way ANOVA

Fig. 4

Biogas and methane production measured for each feed

The maximum rate of biogas production (r_m) was significantly different for the two conditions at 1004 h and from the last two feeds, when the AD MRVCbio showed a faster biogas production as resulted by one-way ANOVA, Tukey’s post hoc, and adjusted P < 0.0001. Considering then the maximum cumulative biogas production (Y_m), at the beginning of the experiment, at the first feeding the AD MRVCbio had higher production and then the situation changed with the reference (reactor AD) producing statistically more cumulative biogas.

Another important factor for the energetic value of biogas is the methane content measured as percentage and then converted to volumes: for AD the lowest value was 137.2 mL CH₄/g VS, and for AD RVCbio was 164.5 mL CH₄/g VS both measured at 1341 h. For both experiments, the highest values (401.7 for AD and 357.1 mL CH₄/g VS for AD RVCbio) were recorded at the end of the monitoring (2354 h).

In the tests were added 8.78 g_RVC/L of CM corresponding to a surface 1.79 m² available for microbial attachment. In fact, by SEM observation, it was possible to visualize the RVC carrier fragment sampled from AD MRVCbio test, with microorganisms attached to the surface: putative Methanosarcina sp. cells are clearly recognizable due to the characteristic sarcinae morphology (see Fig. 5).

Fig. 5

SEM observation of biofilm grown on carriers after 2354 h. Intensity 10 kV, secondary electrons mode and 9900 × magnification, arrow indicates a putative Methanosarcina cells

Literature data on CM presence in 14 days tests in batch conditions reported the advantages for maximum methane production, increased velocity and reduced lag phase [15], in the present research, we show that with the addition of 1.77 (S/m)*m² RVC as carrier, after 100 days of process resulted in a faster methane production: 13% higher maximum velocity comparing AD MRVCbio (2.39 mL/gVS*h) with AD (2.08 mL/gVS*h). This advantage can be translated for process engineering applications in shorter retention time with possible smaller reactors volume.

Microbial community dynamic with RVC and DIET-able inoculum addition

In order to evaluate the fate of the syntrophic inoculum in presence of the complex community of the digestate, sequencing data were statistically evaluated.

The compositional complexity of the whole microbial community in the two experimental settings evaluated as differences in the alpha diversity by Wilcoxon rank sum tests, showed that the two communities had the same complexity; the two settings however did evolve over time significantly diverse communities as evidenced by beta diversity indexes and compared by Wilcoxon Rank Sum and Signed Rank Tests as shown in Fig. 6. Microbial communities after 669 h separated in two clusters showing that a change in composition became observable, maintaining however the same level of complexity (alpha diversity).

Fig. 6

Cluster distance of the two communities over time. In red the OTUs sampled at different times for the AD MRVCbio and in green the OTUs sampled from AD reactors

Most abundant OTUs

The taxa abundance was evaluated at the different taxonomic ranks and the order was chosen as the most informative level.

The abundance over time of the 20 taxa at order level is reported in Fig. 7. In the heatmap it is possible to observe that the different experimental conditions resulted in similar community trends: most dominant taxa are represented by archaeal microorganisms belonging to the order of Methanotrichales, Methanomicrobiales and Methanobacteriales and the bacterial taxa of the orders Eubacteriales, Marinilabiliales, Anaerolineales, Syntrophales and Burkholderiales. Over time the bacterial Marinilabiliales, Anaerolineales, Syntrophales, Burkholderiales and the archaeal Methanobacteriales become less abundant, while the archaeal taxa of the Methanosarcinales become important and the bacterial Thermolithobacterales, Thiotrichales and the Synergistales increase.

Fig. 7

Heatmap of the most abundant 20 taxa reported at order level

Comparing then the two experimental settings, the difference in relative abundance for the reference microbial community (AD) respect to the reactors with the inoculum and the RVC addition (AD MRVCbio) results to be in the archaeal community composition dynamic: in the reference (AD) is reported an increase in the relative abundance of Methanotrichales, Methanomicrobiales and Methanosarcinales and a decrease in Methanobacteriales, while for the AD MRVCbio community Methanotrichales, Methanomicrobiales decrease their abundance and the Methanosarcinales increase up to become the most abundant of the 20 orders reported.

Figure 8 reports the distance-based Redundancy Analysis (db-RDA) ordination plot that shows which process parameters as predictors can explain the taxonomical composition of the microbial community in the experimental settings. The contribution of the process parameters (process time and biogas volume), feeding characteristics (TS, VS, COD and total N) and biogas composition (CH₄, CO₂, H₂S) in the shaping of the taxa composition resulted significant for time and H₂S.

Fig. 8

dbRDA describing the correlations between environmental variables and taxa. Pearson correlation method, 999 permutations and Mantel test

For the AD MRVCbio test the microbial communities sampled at the end of the process (2014 and 2354 h) with Methanosarcinales playing an important role, are influenced by time and described by methane and CO₂.

Syntrophic inoculum fate

Analyzing the taxa abundance expressed in terms of percentage, we focused on order level as it allowed to study the fate of the inoculum added in the AD MRVCbio experiment.

Although Shewanella oneidensis was inoculated in the experiment AD MRVCbio, it was not found among the most relevant OTUs. Shewanella oneidensis belongs in fact to the order of Alteromonadales that was not among the most abundant identified taxa. The fact that it was not found abundant even in the sample after inoculation can be explained with over competition with the indigenous microorganisms or lack of expected cell diffusion from the initial biofilm into the digestate. Instead, Methanosarcina barkerii taxon was present both in the test with the added Methanosarcina barkerii in all the samples, but also in the non-inoculated reactors. Methanosarcina barkerii confirmed therefore to have an active role in the AD process and among the indigenous microorganisms of the digestate sampled from Chiasso plant.

The changes in the microbial community potential function were evaluated by focusing on the capabilities of metabolic transformation of the compounds that showed to have an impact in shaping the microbial community structure as per db-RDA analysis: therefore, process time, H₂S and biogas composition (CH₄ and CO₂) were considered.

Methanotrix sp. (previously known as Methanosaeta sp.) is a known acetoclastic methanogen [21] and the Methanosarcinales taxa are known to have high versatility in the methane production pathways: Methanosarcina barkerii has proved to have acetoclastic, hydrogenotrophic, and methylotrophic pathway enzymes [13]. Methanosarcinales increased with the addition of RVC proving to give high redundancy in the methanogenesis production that results in a microbial community more resilient and a higher stability of the process [34]. Methanosarcinales also produce methane through an interspecific electron transfer process, contributing to its relative abundance [22], so it stands to reason that a CM addition is influencing the Methanosarcina taxon abundance. Also, the fact that we observed a decrease in the Methanotrichales order and a corresponding increase in the Methanosarcinales order at 100 days contact with RVC could indicate a shift toward the hydrogenotrophic methanogenesis.

With regard to sulfur compounds metabolism, an increase over time in the abundance of taxa related to sulfur metabolism is observed for both reactors’ conditions, it seems that RVC presence could influence the H₂S production, but data are not yet conclusive and further research will address this topic.

The advantage in methane production rate quantified as 86% with RVC presence in short-term tests is reduced to 13% at 100 days process in semicontinuous mode. These results are in agreement with literature data considering batch mode of short-term test: [32] in an interesting review reported in fact that methane production rate increased between 79 and 300% in short-term laboratory-scale experiments with several material (Granulated Activated Carbon -GAC-, biochar, carbon cloths and graphene). The only value reported for continuous operation mode measured in 0.5L volume and acetate as substrate, resulted in 80% more methane production rate: the difference in the augmented rate is due to the different material used GAC instead of RVC, but mostly to the retention time considered 20 days compared to the 100 monitored in the present experiments.

It is worth mentioning that the RVC used in this research experiments was acquired from an aerospace company as it could guarantee the quality and purity of the material, as a consequence the material cost makes its application in the AD industry less favorable than other materials such as biochar or activated carbon. The RVC was however added as solid carrier at the beginning of the tests and the chosen solid structure made it easy to recover it from the reactor to be reused. Biochar on the other hand could release chemicals with potential inhibiting effects, making long-term assessment in presence of indigenous microorganisms difficult.

Conclusions

In short-term AD test CM addition (RVC) increased by 84% conversion rate of methane production. Also, Shewanella oneidensis as endoelectrogen and Methanosarcina barkerii as exoelectrogen in co-culture, confirmed to syntrophically convert organics to methane by electron transfer mechanism. In the 100 days experiment with real feed, the increased methane production velocity was confirmed, even if less than short-term test, at 13%. The high versatility pathway of Methanosarcinales is promoted by CM presence and results in a more stable microbial community able to cope with the variable feed composition present in a full-scale plant.