Enhancement of the start-up and performance of an upflow anaerobic sludge blanket (UASB) reactor using electrochemically-enriched biofilm

Abstract

A novel approach was developed to accelerate the start-up of a 20-L UASB reactor under mesophilic conditions. Two runs were conducted, where the first run (Run I) was inoculated with anaerobic sludge, and the second run (Run II) was inoculated with the same sludge supplemented with enriched electro-active biofilms collected from the working and counter electrodes of anodic and cathodic bio-electrochemical systems (BESs). Reactors’ performance and microbial dynamics were monitored over 41 days. Methane production in Run II exceeded 200 mL-CH₄/g-COD within 10 days, compared to 29 days in Run I. Run II achieved 80% removal of soluble COD after 13 days as compared to 23 days in Run I. Sludge washout in Run II stabilized after 3 days, achieving 70% VSS removal, whereas Run I required 17 days. Greater extracellular polymeric substance (EPS) values and higher protein-to-polysaccharide ratios in Run II may indicate accelerated granules formation mediated by EPS. 16S rRNA gene sequencing analysis results revealed shared genera between both runs but different relative abundances. Methanothrix dominated in Run I, while other archaeal genera, mainly Methanosarcina and Methanobacterium increased in abundance in the Run II. The Enterobacteriaceae family was prevalent in both reactors, with three genera, Citrobacter, Klebsiella, and Enterobacter distinctly dominating at different time points, suggesting potential links with the initial seed sludge or enriched biofilm consortia. The addition of electrochemically grown biofilm in Run II likely enhanced the microbial diversity, contributed to the rapid development of granular syntrophic communities, and improved reactor performance.

Graphical Abstract

Supplementation of an upflow anaerobic sludge blanket reactor (UASB) reactor with biofilms enriched in bioelectrochemical systems (BES) polarized at different electrode potentials reduced the start-up time. According to the results, the greater diversity of the BES microbial communities as well as higher protein-to-polysaccharide ratio in Run II likely contributed to the improved performance.

Introduction

There is a worldwide interest in high-rate anaerobic digestion systems for the treatment of high-strength organic wastewater. Anaerobic digestion is an attractive process for wastewater treatment because of its ability to convert organic waste to innocuous material with potential for energy recovery as methane [1]. Among the high-rate anaerobic digestion systems, upflow anaerobic sludge blanket (UASB) reactors are the most widely implemented with more than 3000 plants operating worldwide [2,3]. They are known to be robust high-rate systems in terms of performance, sludge stabilization, land requirement, and methane yield [4–8]. Several studies reported on UASB reactors achieving COD removal rates up to 98% and methane yields up to 330 mL-CH₄/g-COD for various types of waste streams including domestic, synthetic, and industrial wastewaters [5,9–11]. The key feature of the UASB reactors is the formation of immobilized aggregated biomass known as granules which are retained at the bottom of the reactor with the aid of a gas-liquid-solid (GLS) separator system [11–13].

The successful performance of UASB reactors is dependent on the formation of these well-structured dense granules that are comprised of an inert core covered by a consortium of syntrophic microbes that breakdown the organic waste [14,15]. The granules have better settling properties than normal sludge which enables vigorous granule-liquid interaction and high hydraulic loading rates. In granules, inter-species mass transfer limitations are reduced between syntrophic groups. However, a major drawback of the UASB technology is the long start-up time which can take up to 8 months accompanied by sludge washout, low methane yields, and inefficient COD removal rates [6,16]. UASB start-up is hindered mainly by the slow growth of methanogenic and hydrolytic microbes that form the granules required for treatment of complex wastewaters [12,17].

The formation of the granules is reported to be carried out by acetoclastic methanogens that act as scaffolding which immobilizes syntrophic bacteria through EPS secretion [12,13,18,19]. Many studies reported on the use of amendments such as cations, natural polymers, and synthetic polymers to accelerate the formation of granules by binding syntrophic anaerobic communities in UASB reactors. The addition of multivalent cations such as Ca²⁺, Fe²⁺, and Al³⁺ has shown positive effects on granulation though this effect seemed to be dependent on the wastewater COD concentration [20,21]. Better COD removal has also been observed in the presence of polymers, while other reports have suggested better removal efficiency without polymer addition [22–24]. It is worth noting that polymers can bind microbial communities in an indiscriminate manner.

Other studies focused on the addition of enriched pure strains to the UASB reactors to improve granulation and overall performance. Li et al. (2015) reported on the enrichment of a UASB inoculum with a Methanosaeta strain. This resulted in the improvement of the UASB performance and granulation by immobilizing synergistic bacteria and methanogens, consequently facilitating volatile fatty acids (VFAs) degradation [18]. Keyser et al. demonstrated that the enrichment of a UASB inoculum with the bacterial strain Enterobacter sakazakii significantly enhanced the COD removal of winery wastewater influent in a shorter start-up time [25]. However, enrichment using a single strain may not be practically applicable for the treatment of complex substrates.

Several studies have focused on the integration of bioelectrochemical systems into UASB reactors to improve digestion performance [26–30]. Some anaerobic communities were reported capable of extracellular electron transfer via direct interspecies electron transfer (DIET) through cell-to-cell or cell-to-mineral interaction [31–34]. This mechanism was attributed to the exhibited electroactivity of some anaerobic communities, such as those belonging to the genera Geobacter and Clostridium, as well as the Anaerolineaceae family [35–40]. Furthermore, methanogens following the acetoclastic pathway (Methanosaeta and Methanosarcina) [41] or the hydrogenotrophic pathway (Methanobacterium) [42] have been shown to exist simultaneously in anaerobic digesters, several of which are capable of direct electron transfer [33,43].

Interactions between electroactive communities and electrodes have been reported to be similar to interactions observed in anaerobic digester granules. The integration of bioelectrochemical systems into UASB reactors resulted in improved COD removal or methane yield due to the development of biofilms on the electrodes in addition to influencing the suspended biomass in the reactor [44–46]. Other studies focused on the supplementation of UASB sludge with conductive materials such as granular activated carbon (GAC) as a mediator to DIET for electroactive bacteria and methanogens which resulted in those communities’ proliferation, subsequently leading to enhanced performance [17,47]. Operational conditions, such as poised voltage, support material, substrates, and inoculum type were reported to influence the development of the microbial community [48,49].

No studies to date have reported on the selective enrichment of biofilms on electrodes for harvest and direct use in UASB sludge as an inoculum to accelerate the start-up and improve the performance of UASB reactors. This could mitigate the challenges of using other methods such as the addition of pure strains, polymers, support materials, or applied external voltage. In this study, the performance of a UASB reactor inoculated with biofilms enriched in anaerobic bioelectrochemical systems (BESs) was compared to that of a control UASB. The harvested biofilms were grown in anodic and cathodic BESs at selected constant potentials to enrich bacterial and methanogenic electroactive communities using the same substrate as the one fed to the UASB reactors. Reactor performances were assessed based on start-up time, COD removal, methane yield, and sludge settleability. Microbial community characterization in was conducted using high-throughput sequencing of the 16S rRNA gene.

Materials and Methods

UASB configuration and operation.

A 20-L lab-scale UASB reactor was used over two separate runs. The reactor assembly consisted of a reactor equipped with a gas-liquid-solid (GLS) separator system (Armfield Ltd, Hampshire, England). The UASB reactor was seeded with anaerobic sludge obtained from the anaerobic digester of the Bkassine WWTP (Saida, South Lebanon) and operated at a temperature of 35°C. Run I served as control, where the UASB was inoculated with only anaerobic sludge, while Run II was inoculated with the same anaerobic sludge with the addition of enriched electro-active biofilms obtained from the working and counter electrodes of the anodic and cathodic BESs. During both runs, the reactors treated synthetic feed with a COD of 2000 mg/L. The organic content of the feed consisted of the following per 1 g-COD: 186 mg milk powder, 633 mg sucrose, and 50.6 mg citric acid. The macronutrients and trace metal composition is shown in the Supplementary Information section. The UASB reactors were operated at a hydraulic retention time (HRT) of 48 hours and an organic loading rate (OLR) of 1 g-COD/L.d during both runs.

UASB effluent quality testing.

The reactor effluents were monitored regularly for chemical oxygen demand (COD), volatile fatty acids (VFAs), total suspended solids (TSS), volatile suspended solids (VSS), and pH using standard methods.

UASB bed and profile analysis.

To monitor the stabilization of the UASB reactor, regular sampling of the UASB bed sludge was conducted to measure total and soluble COD, pH, TSS/VSS, Total VFAs, and the sludge volume index (SVI).

UASB biogas collection and analysis.

Biogas produced in the UASB was collected and analyzed using gas chromatography (GC) (7890B Agilent Technologies, USA) to determine CH₄ and CO₂ composition. The methane yield rate was calculated from the methane production normalized to the influent COD mass.

Microscopic examination.

Optic microscopy was performed on the UASB bed sludge to determine the size of granules. Images of the bed biomass dispersed in distilled water were taken using a high-quality camera and microscopic lens (LNKOO 30X Zoom Bright LED Stereo Microscope Magnifier).

UASB Microbial Sampling.

The samples for microbial community analysis were collected from the bed of the UASB reactors at various time intervals and from the raw seed sludge biomass (Sludge C and Sludge E). The DNA was then extracted for high-throughput sequencing of the 16S rRNA gene to investigate the total microbial community.

EPS extraction.

Extracellular Polymeric Substances (EPS) analysis was conducted for UASB granules and raw sludge inoculum to quantify the content of proteins and carbohydrates. The phenol-sulfuric acid method with glucose as the standard was utilized to quantify the polysaccharide content (Total Carbohydrate Assay Kit, Sigma-Aldrich), while the total protein Pierce BCA protein assay kit (Thermo Scientific) was employed to determine the protein content of the extract. Results were averaged for duplicate samples from each sludge and normalized against the total extracted DNA concentration obtained from those same samples.

Detailed descriptions of the materials and methods are provided in the supplementary information of this paper.

Results

Biogas production and gas composition.

The lab-scale UASB reactor experiment was conducted in two consecutive runs. The first run served as control using anaerobic sludge as the inoculum (Run I). The second run was inoculated with the same anaerobic sludge in addition to enriched biofilms at the anodic and cathodic BES working and counter electrodes (Run II). The start-up was considered complete once stable methane production, sludge washout cessation, and stable COD removal (greater than 80%) were observed. After approximately 5 days of continuous operation, Run I and Run II started producing biogas containing CO₂ and CH₄. Figure S1 shows the methane composition of the produced biogas with Run II reaching a stable methane composition of around 70% after 15 days, while Run I required more than 34 days to exceed the 60% composition.

The daily volumetric methane production in Run I and Run II is shown in Figure S2, normalized over the effective volume of the UASB reactor [11]. Daily methane production in Run II was higher as compared to Run I, with rates greater than 100 mL/L.d obtained in 6 days while Run I required 28 days to reach similar results. The methane yield with respect to the amount of added COD was computed to assess methanogenic efficiency (Figure 1). Run II achieved stable yields greater than 200 mL-CH₄/g-COD_added in about 10 days of operation whereas Run I needed 29 days to achieve this threshold. Several studies reported similar methane yield results for UASB reactors treating complex wastewaters [6,11]. A higher methane yield in a UASB could be dependent on the efficient enzymatic activity by the different syntrophic partners (hydrolytic, acidogenic, acetogenic, and methanogenic) that make up the reactor’s immobilized granules. Together, these results demonstrate that UASBs amended with the enriched electroactive biofilms (Run II) adapt faster to the operational conditions through enhanced microbial activity compared to traditional anaerobic sludge inoculum (Run I).

Figure 1 -.

Comparison of the temporal variation in total COD (tCOD) removal, soluble COD (sCOD) removal and methane yield in both UASB runs. Symbols: Run I-tCOD (○); Run II-tCOD (●); Run I-sCOD (□); Run II-sCOD (■); Run I-methane yield (△); Run II-methane yield (▲). Methane yield was calculated as the daily methane production per influent COD.

COD removal and sludge washout assessment.

Effluent quality and treatment efficiency were evaluated in both runs to monitor enhancement in the treatment process. Data in Figure 1 demonstrates an improvement in the COD removal in both total and soluble forms, in addition to the total VFA removal between the two runs (Figure S3). Start-up was completed after nearly 19 days in Run II, compared to Run I which required 31 days to exceed 80% total COD removal efficiency. Because total COD incorporates both soluble and particulate material, the soluble COD removal was also assessed. Run II achieved 80% removal of soluble COD after 13 days as compared to Run I (23 days), which indicates efficient removal of the organic content in the influent wastewater at a faster rate in Run II. Comparable results to Run II COD removal were achievable for UASB processes treating wastewater containing complex organic components [4,13]. Sludge washout was another critical parameter evaluated in both UASB runs. Cessation of sludge washout is indicative of functional microbial immobilization in the form of granules within the UASB reactor. Sludge washout was not detected in the effluent of UASB Run II after approximately 3 days of operation (Figure S4, greater than 70% VSS removal), indicating the development of a well-adapted and settleable sludge whereas Run I required 17 days to reach VSS removal rates higher than 70%, with heavy washout observed within the first 2 weeks.

Granule sludge formation and settleability assessment.

To assess the differences between the sludge developed in the UASB bed of Runs I and II, several tests were performed to characterize the developed biomass (Table 1). During both runs, the pH remained within favorable methanogenic ranges over 6 weeks of operations. To measure biomass growth, the VSS of the bed was measured and showed an increase reaching 28.22 g/L in Run I (week 6), whereas Run II had a lower VSS concentration of 25.67 g/L. VSS/TSS ratios also increased in the two runs (0.75 and 0.87 in Run I and II by week 6, respectively) which shows the development of more stable biomass.

Table 1 -.

Different characteristics of the UASB bed biomass during Run I and Run II

	Run I				Run II
	Week 1	Week 2	Week 3	Week 6	Week 1	Week 2	Week 3	Week 6
pH	8.44	7.15	6.82	6.84	8.44	6.75	7.21	6.96
VSS (g/L)	7.11 ± 192	29.33 ± 2.03	30.33 ±0.33	28.22 ± 0.51	8.00 ± 0.00	23.16 ± 1.18	22.67 ± 0.33	25.67 ± 0.00
VSS/TSS	0.58	0.69	0.85	0.75	0.71	0.85	0.87	0.87
tVFA (mg-HAc/L)	835	637	305	1025	835	500	868	1092

The tVFA data presented in Table 1 suggests that there were differences in substrate conversion dynamics between the two runs. At the end of week 1, both runs exhibited high initial tVFA concentrations (835 mg/L as acetic acid). Similar overall trends were also seen in week 2. However, tVFA concentrations started to increase in week 3 for Run II but not for Run I. This may indicate higher hydrolytic and acidogenic activity of the enriched electroactive biofilms with enhanced breakdown of organic matter into VFAs (and subsequently to methane). Further, the increase in tVFA in Run II in weeks 3 and 6 was not associated with a decrease in pH (likely because of the system buffer), and methane production was not negatively impacted (Figure S1).

The sludge volume index (SVI) was used to assess the settleability; and therefore, the degree of granulation in the bed sludge. Results in Figure 2 show the variation in the SVI between the runs over three-time intervals. Low SVI values are indicative of a granular sludge with good settleability while higher values represent more flocculant biomass [50]. Despite having relatively similar SVI during week 1 (61 and 66 mL/g-TSS for Run I and Run II, respectively), the values dropped sharply by week 3 which highlights the development of more consolidated sludge mass. By week 6, the SVI had dropped to approximately 33 mL/g-TSS for Run I (which corresponds to the flocculant sludge range of 20 to 40 mL/g-TSS), while Run II had an SVI of 17.8 mL/g-TSS which falls within the range of granular sludge (less than 20 mL/g-TSS) [15,51,52]. These results show that despite having lower VSS than the control run, the enriched UASB successfully developed granular biomass with high settleability coupled with efficient reactor performance and enhanced start-up.

Figure 2 -.

Comparison of the sludge volume index (SVI) of the bed sludge in both UASB runs at various points during their operations: Run I (red), Run II (blue).

Archaeal community structure and dynamics.

The predominant Operational Taxonomic Unit (OTU) at the genus level in the archaeal community was affiliated with Methanothrix (formerly Methanosaeta) for both reactors. Methanothrix is unique among methanogenic archaea in its ability to produce methane exclusively from acetate [53]. Methanothrix soehngenii was the species to which OTUs were most closely related. M. soehngenii was previously reported during the start-up of a mesophilic UASB treating maize alcohol wastewater under long-term low-strength operation, which favored its dominance due to its ability to compete for acetate with its long filaments [54,55].

Other methane-producing archaea such as Methanosarcina, Methanobacterium, and Methanospirillum were also detected in the control UASB while being more abundant in the enriched UASB (Figure 3). In Run II, the abundance of Methanothrix decreased as compared to Run I, while Methanosarcina increased after enrichment (31.5% on day 1) and remained relatively high at days 15 and 36 (20% and 16%, respectively).

Figure 3 -. Heatmap of the most abundant archaeal genera (> 2% relative abundance in at least one sample).

Key: AWE: working electrode anode, CWE: working electrode cathode, ACE: counter electrode anode, CCE: counter electrode cathode. Run I: UC19, and UC41 (UC stands for UASB Control) collected on days 19 and 41. Run II: UE1, UE15, and UE36 (UE stands for UASB Enriched) collected on days 1, 15, and 36 after enrichment. Sludge C and E: raw seed sludge fed to the control and enriched reactors, respectively.

High-throughput sequencing indicated that the abundance of hydrogenotrophic methanogens, namely Methanospirillum and Methanocorpusculum, decreased even though they were abundant in the seed sludge used (21.3% and 9.1% of the total archaeal community, respectively) while Methanobacterium increased during Run I. The hydrogenotrophic methanogen Methanobacterium increased as well from 3.2% in the seed sludge of the enriched reactor (Run II) to 11.6% after inoculation of the enriched biofilms, possibly due to the high relative abundance of Methanobacterium in the electrode biofilms of BES-A and BES-C. Further, Methanobacterium was present in higher abundance in the enriched biomass compared to the non-enriched biomass (4.8% and 1.1%, respectively). Methanospirillum thrived in the enriched reactor (Run II) compared to the control reactor (Run I) (reaching 3.5% on day 15). Both Methanobacterium and Methanospirillum, are responsible for the conversion of the intermediate products of digestion (such as acetate, hydrogen, and carbon dioxide) into methane [56,57].

Bacterial community structure and dynamics.

As shown in Figure 4 Figure 5, the dominant bacteria belonged to three major phyla: Proteobacteria, Firmicutes, and Bacteroidetes. Other minor phyla were present as well in the bacterial consortia.

Figure 4 -. Analysis of the total microbial diversity of BES and UASB samples at the phylum, genus, and species levels (> 1.5% relative abundance in at least one sample).

Key: AWE: working electrode anode, CWE: working electrode cathode, ACE: counter electrode anode, CCE: counter electrode cathode. Run I: UC19, and UC41 (UC stands for UASB Control) collected on days 19 and 41. Run II: UE1, UE15, and UE36 (UE stands for UASB Enriched) collected on days 1, 15, and 36 after enrichment. Sludge C and E: raw seed sludge fed to the control and enriched reactors, respectively.

Figure 5 -.

The relative abundance (%) of the most predominant bacterial species in BES and UASB samples.

The phylum Proteobacteria was the most prominent in both UASB runs, with Enterobacteriaceae as the most abundant family. However, within this family, three genera Citrobacter, Klebsiella, and Enterobacter were distinctively dominant in both reactors. This could be linked to the presence of these genera in the initial seed sludge fed to each reactor or in the enriched biofilm. Citrobacter and Klebsiella were not detected in the sludge inoculated in the UASB reactor of Run II, but they grew in abundance immediately after enrichment with biofilms grown on BES electrodes. Citrobacter freundii, specifically, thrived in Run II, with a relative abundance of 34.4% on day 15 and 49.2% on day 36. This contrasts with Run I where C. freundii did not exceed a relative abundance of 1.1%. Citrobacter, particularly C. freundii, has been identified as an exoelectrogenic microorganism in microbial fuel cells. Feeding mode (batch feeding in BESs compared to continuous feeding in UASB) and seed sludge to feed ratio could have influenced their behavior and proliferation, and affected competition with other microorganisms present in both systems [58–60]. Operational and environmental conditions in anaerobic wastewater treatment systems have been reported to exert selective pressure on the community, therefore, influencing the population structure, diversity, and heterogeneity [61–64].

Klebsiella reached 32.4% relative abundance on day 19 in Run I compared to 6.8% on day 15 in Run II. Likewise, it had a relative abundance of 23.6% and 3.6% on days 41 and 36 of Runs I and II, respectively. Klebsiella strains were previously observed in a UASB reactor with glucose as substrate [26]. Enterobacter doubled in abundance between days 19 and 41 during Run I (5.5% and 12.8%, respectively), whereas it was mildly enriched during Run II (2.4% on day 15 and 1.4% on day 36). Species of the genus Enterobacter have been previously found to reduce the start-up time and increase the COD removal in the granular sludge of a UASB reactor treating brewery wastewater [25,65].

The detected members of the Enterobacteriaceae family in the present study were facultative anaerobes. These microorganisms are fermentative and employ analogous metabolic pathways, resulting in the production of acetate, ethanol, and formate [66–68]. These three genera showed remarkable combined abundances accounting for 39.2% at day 19 and 37% at day 41 in Run I, likely helping to maintain a stable fermentation rate. It was suggested that Enterobacter and Klebsiella contributed to the consumption of the residual oxygen in an agitated granular sludge bed bioreactor, promoting the growth of the anaerobic Clostridium species that are sensitive to oxygen [69–71]. These findings align with those of the current study, where an increase in the relative abundance of both Enterobacteriaceae members in both runs co-occurred with an increase of that of Clostridium sensu stricto.

The phyla Firmicutes and Bacteroidetes contain fermentative bacteria that exhibit a strong hydrolytic capacity, which is particularly evident during the acidogenic phase [72]. Although the seed inoculum in both runs were rich with unclassified Bacteroidetes (18.2% and 36.8% for sludge of Run I and II, respectively), a sharp decreasing trend was observed by the end of both runs (0.3%). In the phylum Bacteroidetes, Macellibacteroides was the most present genus and is known to produce various VFAs from carbohydrates (lactate, acetate, butyrate, and iso-butyrate) [73]. A sharp increase in Macellibacteroides from day 19 to day 41 in the control reactor was observed (from 1.1% to 15.8%). This genus’ abundance was very low in the enriched UASB (1.3-2.3%).

OTUs belonging to the Lactobacillaceae family were the predominant subgroup among Firmicutes. Two lactic acid-producing bacteria, Latilactobacillus and Lactobacillus, increased in abundance compared to the inoculum of the control UASB and were enriched at day 19, but subsequently decreased at day 41. Several studies reported the negative effect of the development of Lactobacillales populations in anaerobic systems on substrate conversion rates, which can lead to reactor failure [74–76]. Previous studies have reported the hindrance of hydrogen production by Clostridium strains, which was attributed to the secretion of bacteriocins by lactic acid-producing bacteria [74,77]. An increase in the abundance of Clostridium sensu stricto on day 41 of the control run was concomitant with a decrease in lactic acid-producing bacteria, the latter of which were not detected in the enriched reactor.

The genus Enterococcus was only abundant in the sludge used in Run II. Both genera decreased throughout the enrichment experiment, reaching similar abundances at day 15 (5.5%) and 36 (0.5% and 0.3% for Enterococcus lemanii and Enterococcus mundtii, respectively). The genus Trichococcus was present in significant abundances in seed sludges (Run I and Run II, around 10% relative abundance). A drastic decrease was seen for Trichococcus in the control run (from 9.8% in the seed to 1% on day 19). Although the inoculum contained Trichococcus, it was noted that the dominance of Trichococcus was gradually replaced by Latilactobacillus, both belonging to the same order. The Trichococcus OTU had 100% sequence similarity with Trichococcus flocculiformis, which is a filamentous bacterium first isolated from bulking sludge [78]. However, the introduction of T. flocculiformis to syntrophic co-cultures of Syntrophomonas wolfei and Methanospirillum hungatei resulted in a notable enhancement in volumetric methane production, which was linked to the tendency of T. flocculiformis to aggregate with syntrophic partners [79]. The presence of both Trichococcus (5.5% relative abundance) and Methanospirillum (3.5% relative abundance) at day 15 in the enriched sludge may have enhanced the settleability and methane production rates in the UASB reactor of Run II. A study on sewage sludge-derived hydrochar using glucose as substrate also showed higher methane production that corresponded with Trichococcus and Methanothrix as dominant genera, suggesting a potential role in DIET [80]. The high relative abundances of Methanothrix and Trichococcus during Run II of this work supports these previous observations.

Several members of Chloroflexi are known acetogens, providing acetate to other microorganisms through carbohydrate fermentation [81,82]. This phylum was found in both UASBs (5.4% and 2% highest abundance in Run I and II, respectively). Within Chloroflexi, the Anaerolineaceae family has been commonly identified in UASB systems, specifically [83].

Diversity analyses of microbial populations.

To gain a more comprehensive understanding of the microbial community structure, α-diversity was measured using various indices such as the Sobs, Chao, Shannon, and Simpson (measured as Inverse Simpson) indices and are represented in Table 2.

Table 2 -.

Richness and diversity estimation of microbial communities in samples obtained during BES enrichment and UASB Run I and Run II.

Sample ID	Chao	Sobs	Good’s coverage (%)	InvSimpson	Shannon
AWE	11740	3366	98	25.7	4.3
ACE	11739	3223	98	13.6	3.7
CWE	11105	2916	98	15.6	3.7
CCE	10019	2745	98	11.7	3.6
Sludge C	9160	3574	97	18.1	4.2
UC19	9846	2800	97	9.0	3.8
UC41	9496	3151	97	9.0	3.4
Sludge E	13762	3287	97	6.5	3.0
UE1	21770	5933	96	25.7	4.9
UE15	13277	3017	96	7.0	3.5
UE36	16514	2816	98	3.6	2.4

The Sobs index measures the observed richness of OTUs. The InvSimpson and Shannon indices provide measures of the diversity and evenness of OTUs [84,85], with higher values indicating higher diversity and evenness. Both the AWE and ACE samples had very similar species richness (Chao index values of 11740 and 11739, respectively). The AWE sample had higher diversity and evenness of microbial species compared to the ACE sample, as indicated by the InvSimpson and Shannon indices (25.7 and 4.3 for AWE and 13.6 and 3.7 for ACE, respectively).

When comparing the AWE, CWE, and CCE samples, the AWE microbial communities exhibited higher species richness, diversity, and evenness as compared to the CWE and CCE samples. When comparing samples from both before and after enrichment from Run II (Sludge E and UE1), results indicate that UE1 had higher richness than Sludge E. The Sobs value for Sludge E was 3287, which was lower than that of UE1 (5933). The latter, therefore, had higher species diversity compared to the Sludge E (InvSimpson 25.7 and 4.9 and Shannon 6.5 and 3, respectively). Sample UE1 had the highest InvSimpson and Shannon indices and, therefore, the highest diversity and evenness of the microbial community among all samples. It also had the highest estimated richness with a Chao index of 21770 [86]. These findings indicate that the enrichment with BES biofilms caused a notable rise in species diversity during the start-up phase of UASB reactors. Species richness, diversity, and evenness dropped gradually after the enrichment, as evidenced by progressively decreasing values of related indices (Sobs, InvSimpson, and Shannon) in samples UE15 and UE36.

Sludge C (from the control run) exhibited the greatest number of observed species, as indicated by its Sobs index of 3574. Similarly, this sample had the highest diversity, species richness, and evenness with relatively high InvSimpson and Shannon indices (18.1 and 4.2, respectively). The microbial richness in the control reactor biomass communities declined notably between the start-up of Run I and day 19 (from 3574 to 2800). This decrease, observed in both runs, may be due to the shift from continuous stirred tank reactor (CSTR), where the inoculum sludge was initially developed in, to UASB, which might have resulted in the inhibition of certain species and, concurrently, the proliferation of other species that thrived in the system. The adaptation of the microbial communities to the UASB likely led to a recovery of microbial diversity as indicated by the Sobs and InvSimpson indices on day 41. Overall, changes in the calculated diversity indices and richness values of the samples were in line with the observed variations in the relative abundances of different microbial groups.

EPS content and composition in the raw and UASB sludge.

The EPS content and composition in the seed inoculum and UASB sludges are presented in Table 3. Results showed that the enriched UASB Run II samples had higher amounts of EPS with higher concentrations of both PN and PS compared to the seed Sludge C and control UASB Run I samples.

Table 3 -.

Content composition of EPS in the inoculum sludge and UASB granular sludge.

EPS composition (μg/μg DNA)
	Control UASB Run I			Enriched UASB Run II
Content	Sludge C	UC19	UC57	Sludge E	UE 15	UE 32
PN	15.1	56.8	41.5	1	495	633.9
PS	20.9	103.3	89.3	2.4	51.5	790.5
Total EPS	35.9	160.1	130.8	3.4	546.5	1424.4
PN/PS	0.7	0.6	0.5	0.4	9.6	0.8

Key: PN: protein, PS: polysaccharide, PN/PS: protein to polysaccharide ratio

Polysaccharides were the dominant component of EPS during Run I causing a decrease in the PN/PS ratio. Conversely, on day 15 of Run II, the PN/PS ratio increased (9.6) compared to the seed Sludge E (0.4), indicating a shift in the composition of EPS towards more protein-like substances. The increase in protein content in these tightly bound extracted EPS was reported to reflect the development of stable granular sludge as a transition from flocs to granules [87,88]. Additionally, a higher PN/PS ratio was previously linked to higher hydrophobicity, which is favorable for EPS stickiness and sludge granulation [89,90]. At day 32 of Run II, a substantial increase in both PN (633.86 μg/μg DNA) and PS (790.53 μg/μg DNA) content was observed. Interestingly, proteins have been considered as redox mediators that shuttle electrons to acceptors or from donors by extracellular electron transfer (EET), probably playing a significant role in enhancing methanogenesis [91–95].

Microscopic analysis of UASB granule morphology.

Imaging analysis was performed on the granular sludge in the UASB bed using a microscope lens (Figure S5). The size of the granules ranged between approximately 0.5 mm and 2 mm. Previous studies have reported variable sizes of sludge granules operated in UASBs operating under different conditions, with sizes ranging from 0.2 mm to 5 mm [4,6,96]. Variability in granular sizes could be attributed to environmental and operational conditions rather than performance reasons. These conditions are directly correlated to the hydrodynamics within the UASB system and are affected by physical parameters such as upflow velocities and shear stress caused by influent flow, which can help determine agglomerate sizes to suit the applied conditions [19,97].

Correlation between the microbial community and reactors’ performance.

Spearman correlation analysis revealed strong associations (borderline significance due to the small sample size) between key microbial taxa and reactor performance parameters on days 19 and 41 of Run I, and on days 15 and 36 of Run II (ρ ≥ ±0.95, p = 0.051). Notably, both Citrobacter freundii and Citrobacter europaeus exhibited strong positive correlations with VSS removal, which aligns with their dominance in Run II (44.8–64.2% relative abundance on days 15–36, Figure 4) and might have contributed to rapid sludge stabilization as evidenced by the high VSS removal achieved (≥95%, Figure S4). In contrast, their low abundance in Run I (<1.1%, Figure 4) corresponded with delayed sludge stabilization (Figure S4). Both Citrobacter species were also associated with protein production in the EPS matrix, which could play a key role in the development of stable granular sludge. Additionally, Citrobacter spp. have been demonstrated to facilitate extracellular electron transfer through membrane-associated proteins (possibly c-type cytochromes) [98]. This suggests that Citrobacter spp. may have contributed to sludge stabilization by enhancing electron transfer in Run II. Conversely, Klebsiella pneumoniae and Enterobacter mori exhibited strong negative correlations with VSS removal, which is consistent with their low abundance in Run II compared to Run I (17 days to stabilize).

Latilactobacillus graminis and Lactobacillus delbrueckii, which were abundant in Run I but absent in Run II (Figure S4), correlated negatively with methane yield. The low methane yields observed on days 19 and 41 in Run I (Figure 1) illustrate their potential negative impact. High abundances of lactic acid bacteria (LAB) such as these have been previously shown to suppress hydrogen production and contribute to reactor failure [99,100]. It should be noted that some other studies on UASB systems have shown a positive association between LAB, hydrogen-producing bacteria, and acetoclastic methanogens, which resulted in higher methane production [101,102]. Thus, the impact of LAB on UASB performance still requires further investigation. Unclassified Anaerolineaceae also had negative associations with COD removal and methane production, as indicated by its higher abundance in Run I and its reduced abundance in Run II. Conversely, methane content was strongly associated with Macellibacteroides fermentans, which increased in Run I (from 1.1% on day 19 to 15.8% on day 41, Figure 4). This increase correlated with a rise in methane content in biogas (from 48% to 77% by day 41, Figure S1).

Discussion

The goal of this research was to reduce the start-up time of UASB reactors through the addition of electroactive biofilm enriched in BES systems. All indicators suggested an improved start-up and performance in Run II as compared to Run I. Daily methane production in Run II was higher than Run I. Methane composition of the produced biogas in Run II reached a stable level in a shorter time frame than in Run I. Stable methane yield with respect to the amount of added COD was achieved in Run II (200 mL/g-COD_added) in 19 days less (around 65% time reduction) compared to Run I. Moreover, sludge ceased to be detected in the effluent of Run II in a much shorter time frame than in Run I, indicating the development of well-adapted and settleable sludge. The high SVI values in Run I sludge indicated poor settleability, therefore, a flocculant sludge, whereas the sludge in Run II was characterized as granular, having low SVI values and good settleability. A difference in the physical structure of both sludges was observed, with granular sludge being more compact and stable than flocculant sludge.

Even though the microbial communities shared similar genera in both runs, their relative abundance was significantly different (p-values < 0.05). A substantial change in the archaeal community structure was observed in Run II as compared to Run I. Methanothrix comprised the majority of methanogens in the control UASB sludge, but their proportion decreased in the enriched reactor with the increase of the relative abundance of various genera including Methanosarcina, Methanobacterium, Methanospirillum, Methanomassiliicoccus, Methanoculleus, and Methanoregula. The greater diversity of methanogens may have pointed out the need for metabolically versatile methanogens to obtain high methane production rates through hydrogenotrophic, acetoclastic, and methylotrophic pathways as well as to speed up the sludge granulation process [13,18,103].

Similarly, the investigation of the bacterial taxa showed a significant difference between the control and enriched runs, particularly within the phylum Proteobacteria. Notably, members of the genus Citrobacter thrived in the biofilm-enriched reactor, becoming the most abundant among all microorganisms, while having a lower abundance in the control run despite being present in the seed sludge. Past studies reported on electroactive bacterial enrichment in coupled bioelectrochemical-UASB reactors and their impact on improving the reactor’s performance [104,105].

Contrary to what was anticipated, Clostridium sensu stricto, which was expected to be abundant after biofilm enrichment, was relatively low during Run II. They increased throughout both runs, possibly due to their ability to form spores in order to survive and recover [106,107]. It is worth mentioning that Enterobacter and Klebsiella may have contributed to this rebound [70,71].

The EPS values were greater in Run II compared to those in Run I, which was likely due to the addition of the EPS-rich electrochemically grown biofilms in Run II [108]. The higher EPS values and PN/PS ratios may have accelerated the formation of EPS-mediated granules in Run II. EPS form a matrix that enmeshes the microbial cells, thus mediating interaction [109,110]. Polymer additives have been reported to accelerate the aggregation of sludge [23,111]. However, a very few studies focused on the bacterial production of EPS to accelerate the granulation of sludge in UASB reactors [112]. Moreover, some studies have reported on the role of EPS in microbial EET within electroactive biofilms. The composition of EPS, particularly higher protein content, is advantageous for driving EET in anaerobic granular sludge [113–116] and anodic biofilms [117]. This positive impact is due to certain functional carboxylic groups present in hydrophobic proteins that contribute to the negative charge of the sludge [118]. Proper supplementation of protein in the influent substrate can provide the needed components for synthesizing EPS, successful granular formation, and enhancement of microbial interaction [12,111].

The shaping of anaerobic microbiomes depends largely on feedstock type (livestock manure, food wastes, municipal solid wastes, industrial wastewater, etc.) and composition in terms of carbohydrates and proteins [119,120]. Furthermore, selecting the appropriate inoculum in anaerobic digesters has been shown to be essential in affecting the microbial structure and enhancing process performance and robustness [121–124]. The greater community diversity in the BES also may have accelerated the faster development of the syntrophic communities in Run II. A complex feed, such as the one used in our study, is reported to require a more diverse microbial community to ensure rapid granulation by key methanogens and syntrophic organisms [3,103,111]. A high microbial diversity in anaerobic digesters ensures the process’ resilience and efficacy [125]. Thus, the fast development of a diverse community in the BES and its use as an inoculum may have aided in the accelerated start-up of Run II, possibly through supplementation with symbiotic microorganisms that facilitated substrate consumption and intragranular substrate diffusion between the different species [3,126]. However, some of the exoelectroactive microorganisms identified from the BES systems were not present or did not thrive in Run II. Various factors, such as operational conditions and metabolic versatility in the sludge, could have played a role in the adaptation of these microbes to different environments [127,128].

Through the results of this work, it appeared that the growth of a diverse syntrophic electrochemically enriched biofilm in the BES and its use as an inoculant directly aided in the accelerated start-up of the UASB reactor in Run II. Similarities in the microbial community composition include the presence of genera such as, Klebsiella, Clostridium, Enterobacter, and Methanothrix, which were consistently observed in both runs, however a higher microbial diversity was seen in Run I compared to Run II where an increased abundance of specific taxa like Citrobacter spp. was noted. The communities in Run II were able to form viable granules faster. The EPS content and composition in the enriched run may have positively influenced aggregation and led to increased efficiency of anaerobic treatment and biogas production.

The biofilm-enriched sludge used in Run II reduced the formation time of granular syntrophic microbial communities from the flocculent sludge compared to Run I, which had direct implications on digestion performance. This was likely aided by the microbial community structure’s significant divergence between Run I and Run II. In Run I, facultative anaerobes (Klebsiella and Enterobacter) and lactic acid bacteria (Lactobacillales) dominated during the first 15 days of the experiment (Figure 4), possibly contributing to decreased digestion efficiency as evidenced by the relatively low methane yields (Figure 1). However, the increase in the abundance of the fermentative-acetogenic bacterium Macellibacteroides fermentans correlated with an increased methane yield by day 41 (Figure 1) [129]. Conversely, biofilm enrichment in Run II promoted the selection of taxa such as Citrobacter spp., which likely facilitated granulation through secretion of protein-rich EPS. It is noteworthy that secreted proteins can also support potential extracellular electron transfer and function as electron shuttles [59,60,98]. Further, a probable syntrophic relationship between Trichococcus and Methanospirillum may have improved both sludge settleability and methane yield [79]. These findings demonstrate that the enrichment of biofilms influenced microbial structure within the UASB reactor by proliferation of functional syntrophic microbes. The resulting synergistic consortia likely supported the reduced startup time and enhanced performance through improved granule stabilization and substrate degradation.

Conclusion

This study’s results indicated that the start-up of a UASB reactor could be accelerated via addition of enriched electro-active biofilms collected from the working and counter electrodes of anodic and cathodic bio-electrochemical reactors. The enriched UASB reactor (Run II) reached 80% COD removal after 13 days compared to 23 days in the control reactor (Run I). Moreover, sludge washout was not detected in the effluent of Run II after 3 days of operation, whereas Run I required 17 days to reach 70% VSS removal rates. The daily methane production in Run II exceeded that of Run I, achieving production rates of 200 mL-CH4/g-COD within just 10 days, compared to 29 days in Run I, indicating a 65%-time reduction to achieve this value. Similarly, the methane content of the produced biogas in Run II stabilized at approximately 70% after 15 days, while Run I required over 34 days to reach 60% methane content.

The enriched biofilm UASB (Run II) in our study had greater EPS values, PN/PS ratios, and greater microbial diversity. In Run I, a longer time was needed for a flocculent sludge consortium to form the syntrophic granular community for efficient digestion of the complex feed. Conversely, the supplementation of the inoculant with EPS-rich biofilm from the electrodes had accelerated the formation of viable granules and therefore decreased start-up time. Even though the microbial communities shared similar genera in both runs, viable granules formed faster in the enriched run. The method developed in this study could be used to mitigate slow UASB start-up which is hindered by the slow growth of syntrophic methanogenic and hydrolytic microbial communities that form the granules, especially when treating complex wastewaters. Future studies may focus on delineating specific signatures of microbial activity associated with various species and digestion pathways, along with optimizing the protocol for varied operational conditions and feed streams.

Highlights:

• The addition of electrochemically-grown biofilm accelerated UASB reactor start-up.

• Enriched biofilm enhanced granule formation and anaerobic treatment efficiency.

• COD removal, microbial diversity, and biogas production improved with biofilm enrichment.

• Enterobacteriaceae family was the most abundant in both reactors.

Data Availability

The sequences obtained in this study were deposited in the NCBI Sequence Read Archive under the BioProject accession number PRJNA924133 https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA924133