The experiment utilized a fixed-bed reactor made of synthetic glass (10 cm in outer diameter, 18 cm in height, and an effective volume of 1 L). The reactor was operated under anaerobic conditions with a one-time feed setup.

To regulate the operating temperature, the bioreactor was placed inside a biochemical constant-temperature chamber (Model MIR 254, Sanyo, Japan) maintained at 20°C.

Gas volume was measured using the drainage method, and regular samples were taken from the reactor outlet to determine the concentrations of volatile fatty acids (VFAs), chemical oxygen demand (COD), pH, and methane.

In the reactor, cylindrical carbon fiber fabric was added as a biofilm carrier (outer diameter of 3 cm and a height of 5 cm, supplied by the Japanese Carbon Company in Tokyo, Japan). The height of the carbon fiber fabric and the liquid level were kept consistent. Each cylindrical carbon fiber carrier occupied 10% of the effective volume of the reactor. Three gradients were set: 0, 10, and 20% carbon fiber carriers, labeled cv0, cv10, and cv20, respectively.

The reaction substrate was an artificially synthesized glucose wastewater composed of water, syrup (homogeneity: 70%, solid content: 45%), and industrial feed (mainly protein, fat, and crude fiber). The COD: N:P ratio was maintained at 300–500 5:1 (Zhang et al., 2012).

The initial inoculum was obtained from laboratory-cultured activated sludge that had been long-term acclimated at 20°C. It was passed through a 1 mm sieve to remove larger solids. The resulting inoculum had a total solid (TS) content of 48.82% and a volatile solid (VS) content of 4.56%, sludge inoculum was controlled at 250 g for all treatment groups.

The biological augmentation strain was laboratory-purified L2, which was isolated and purified from the sludge of wetlands in Yanbian Prefecture, Jilin Province, China (latitude 43°45′22″ N, longitude 128°44′10″ E). This strain is Gram-negative, rod-shaped, strictly anaerobic, and belongs to the Methanomicrobium, which can produce methane via both H₂ and CO₂ metabolism and organic acid metabolism (Liu and Whitman, 2008).

Three different microbial liquid concentration gradients were established, namely 10, 20, and 30%, corresponding to 30 mL, 60 mL, and 90 mL added microbial liquid, respectively. The total volume of microbial liquid added to the culture medium was controlled at 90 mL. The control group used a sterile culture medium without microbial liquid. Finally, artificial glucose wastewater was used to finalize the volume to 600 mL. Each experiment was repeated three times.

VFAs were determined using high-performance liquid chromatography (LC-MS2020, Japan), COD was measured using a water quality monitoring instrument (Lovi-bond 99731COD, Germany), pH was determined using a compact pH meter from Horiba, and CH₄ levels were measured using a biogas analyzer (Model ADG, Landtec, United States).

Sludge samples were taken from the reactor at predetermined time points, sample DNA was extracted using the MN NucleoSpin 96 Soil DNA extraction kit. Specific amplification of the bacterial 16S V3 + V4 variable region was carried out using the primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). For archaea, the 16S V3 + V4 variable region was specifically amplified using the primers Arch349F (5′-GYGCASCAGKCGMGAAW-3′) and Arch806R (5′-GGACTACVSGGGTATCTAAT-3′). PCR products were quantitated by gel electrophoresis (ImageJ software), and samples were mixed at a 1:1 ratio based on their mass. DNA was then purified using the OMEGA DNA nucleic acid purification kit. After purification, a library was constructed for high-throughput sequencing, which was provided by Beijing BioMighty Biotech Co., Ltd.

A bioinformatics analysis was performed using the BioMighty cloud platform. Beta diversity distances were calculated using QIIME (version 1.9.1) and alpha diversity analysis was performed using Mothur v. 1.30. The igraph, Hmisc, pheatmap, and vegan packages were used to organize and initially map the data in the R 4.2.1 environment.

Figure 1 compares the performance of low-temperature anaerobic digestion in fixed-bed reactors at different microbial liquid concentrations (control, 10, 20, and 30%) and effective volumes of carbon fiber carriers (0, 10, and 20%). Over a 28 days period, the daily methane production in all treatment groups was higher than that in the control group, and methane production in each treatment group reached its peak on day 21.

As the microbial liquid concentration increased, daily methane production also increased. After the addition of carbon fiber carriers, the daily methane production in the low-concentration microbial liquid treatment group reached a higher level, especially in the first 21 days of the experimental period. In anaerobic digestion processes, low temperatures generally limit bacterial metabolism and reaction kinetics (Zhou et al., 2018); however, the low energy consumption advantage of low-temperature anaerobic digestion makes it valuable for research related to anaerobic digestion in cold regions (Martí-Herrero et al., 2022). Previous studies have shown that the addition of a reasonable amount of exogenous microorganisms through bioaugmentation can effectively optimize the efficiency of anaerobic digestion systems (Martin-Ryals et al., 2015; Li et al., 2022; Linsong et al., 2022). Our results demonstrate that carbon fiber carriers can to some extent reduce the cost of bioaugmentation, implying that carbon fiber carriers can lower the concentration requirements for microbial liquid in fixed reactors while achieving high methane production.

In addition, the cumulative methane production data indicate that bioaugmentation with Methanomicrobiaceae effectively enhances the total cumulative methane production in the system and that the cumulative methane production is directly proportional to the microbial liquid concentration. In the treatment group with a 20% carrier volume, the methane production at the highest microbial liquid concentration was relatively reduced; however, in the treatment group with a 10% carrier volume, 30% microbial liquid resulted in the highest methane and total gas production, suggesting that the excessive addition of carbon fiber carriers inhibits the gas production efficiency of the system.

Furthermore, as shown in Figure 2, the COD removal efficiency in all treatment groups was higher than that in the control group at all time points. The COD removal efficiency was directly proportional to the microbial liquid concentration. In the treatment group with a 10% carbon fiber carrier volume and 30% microbial liquid, the maximum COD removal efficiency was 93.48%. In the treatment group with a 20% carbon fiber carrier volume, the COD removal efficiency remained relatively consistent during the latter stages of the experiment. Regarding the anaerobic fermentation process, the pH value in all treatment groups fluctuated between 7.0 and 8.0, which is within the suitable pH range for anaerobic digestion (Zhang et al., 2010). Additionally, the pH value in the treatment groups was lower than that in the control group, indicating that bioaugmentation microorganisms and carrier addition effectively improved the environmental conditions in the reaction system.

As shown in Figure 3, the organic acid concentration in each treatment group was significantly higher than that in the control group at the beginning of the reaction. As digestion progressed, the concentrations of formic acid and acetic acid significantly decreased, indicating that the reaction system effectively inhibited the generation of organic acids and maintained them at a relatively stable level.

Figure 4A illustrates the top 10 phyla in terms of abundance in each treatment group. At 0% carrier volume, the abundance of Acidobacteria, Proteobacteria, and Actinobacteria increased with increasing microbial concentration. By contrast, the abundance of Firmicutes, Bacteroidetes, and Caldiserica decreased with increasing microbial concentration. The addition of carbon fiber carriers reduced the abundance of Acidobacteria and Firmicutes while significantly increasing the abundance of Proteobacteria.

Figure 4B displays the archaeal community structure in each treatment group. At cv0, the abundance of Candidatus_Udaeobacter increased with the concentration of microbial agents, while the abundance of Bathyarchaeia decreased. Previous reports suggest that Candidatus_Udaeobacter contains highly affinitive hydrogenases that can metabolize H₂ to provide energy for the respiratory chain under nutrient-limited conditions (Greening et al., 2015; Willms et al., 2020). On the other hand, Bathyarchaeia are known for their autotrophic acetogenic metabolism capabilities (Xie et al., 2022; Hou et al., 2023). This indicates that following the addition of Methanomicrobiaceae, which produce methane through H₂ and CO₂ metabolism pathways, the microbial community in the reaction system gradually shifts toward these pathways, while the proportion of microorganisms utilizing organic acids to produce methane decreases. These findings are by the organic acid measurements, further demonstrating that bioaugmentation with Methanomicrobiaceae effectively reduces the biological production of organic acids and enhances the efficiency of methane production via H₂ and CO₂ pathways. At cv20, the abundance of Bathyarchaeia was relatively higher in comparison with that at cv10, suggesting that an excessive amount of fiber carriers can alter the community composition and reduce methane production.

LEFse was used to assess the phylogenetic tree and LDA scores of bacteria and archaea in different carbon fiber carrier ratio treatments, as shown in Figure 5. The significantly different bacteria in the treatment without carbon fiber carriers included Acidobacteria and Clostridiaceae_1, which have the highest LDA scores. At 10% effective carrier volume, the significantly different bacteria were Synergistetes and Bacteroidetes, with Bacteroidetes having the highest LDA score. At 20% effective carrier volume, the different bacteria were Chloroflexota and Proteobacteria, with Proteobacteria having the highest LDA score.

The phylogenetic tree of archaea in different treatments is displayed in Figure 5. In all treatments, Methanobacterium-related species consistently exhibited a significantly dominant position. This phenomenon was more pronounced at 10% effective carrier volume; while at 20% effective carrier volume, Bathyarchaeia was present in more marked abundance. This further validates the earlier results showing differences in acetate and methane production.

FAPROTAX functional gene prediction can annotate and predict microbial community functions based on the taxonomic classification of 16S sequences. In Figure 6, the horizontal axis represents the samples, and the vertical axis represents the functional groups of the top 10 dominant bacteria in the microbial community. Among these groups, The abundance of hydrogenotrophic_methanogenesis increased to different degrees in all treatment groups, implying that the addition of Methanomicrobium strains significantly improved the methanogenic functioning of the anaerobic digestive system, with 30% exogenous microorganisms and a 10% volume of carbon fiber carrier had the highest functional abundance of hydrogenotrophic_methanogenesis, thus suggesting that the specific biofortification with carbon fiber carrier can maintain stable methanogenic efficiency at low temperatures.