Ex-situ biogas upgrading in thermophilic trickle bed reactors packed with micro-porous packing materials

Two thermophilic trickle bed reactors (TBRs) were packed with different packing densities with polyurethane foam (PUF) and their performance under different retention times were evaluated during ex-situ biogas upgrading process. The results showed that the TBR more tightly packed i.e. containing more layers of PUF achieved higher H 2 utilization efficiency ( > 99%) and thus, higher methane content ( > 95%) in the output gas. The tightly packed micro-porous PUF enhanced biofilm immobilization, gas-liquid mass transfer and bio-methanation efficiency. Moreover, applying a continuous high-rate nutrient trickling could lead to liquid overflow resulting in formation of non-homogenous biofilm and severe deduction of biomethanation efficiency. High-throughput 16S rRNA gene sequencing revealed that the liquid media were predominated by hydro-genotrophic methanogens. Moreover, members of Peptococcaceae family and uncultured members of Clostridia class were identified as the most abundant species in the biofilm. The proliferation of hydrogenotrophic methanogens together with syntrophic bacteria showed that H 2 addition resulted in altering the microbial community in biogas upgrading process.

• Biological biogas upgrading process was evaluated in trickle bed reactors (TBRs).Two thermophilic trickle bed reactors (TBRs) were packed with different packing densities with polyurethane foam (PUF) and their performance under different retention times were evaluated during ex-situ biogas upgrading process.The results showed that the TBR more tightly packed i.e. containing more layers of PUF achieved higher H 2 utilization efficiency (>99%) and thus, higher methane content (>95%) in the output gas.
The tightly packed micro-porous PUF enhanced biofilm immobilization, gas-liquid mass transfer and biomethanation efficiency.Moreover, applying a continuous high-rate nutrient trickling could lead to liquid overflow resulting in formation of non-homogenous biofilm and severe deduction of biomethanation efficiency.High-throughput 16S rRNA gene sequencing revealed that the liquid media were predominated by hydrogenotrophic methanogens.Moreover, members of Peptococcaceae family and uncultured members of Clostridia class were identified as the most abundant species in the biofilm.The proliferation of hydrogenotrophic methanogens together with syntrophic bacteria showed that H 2 addition resulted in altering the microbial community in biogas upgrading process.

Introduction
Developing biogas upgrading processes and full utilization of waste streams is crucial to achieve a CO 2 -neutral Europe by 2050 (Commission, 2020).The upgraded biogas with CH 4 content above 95% could be utilized as vehicle fuel or injected into the natural gas network (Deng and Hägg, 2010).To increase the CH 4 content of biogas, CO 2 removal methods, i.e. absorption, adsorption, membrane filtration, and cryogenic separation, are conventionally used (Sun et al., 2016).These methods are robust, reliable, and capable to treat high gas loadings.However, these methods are challenged by high energy consumption and operational costs.Biological methanation of CO 2 coupled with external H 2 is an attractive approach that could compete with conventional CO 2 removal methods specially at low H 2 prices.Biological upgrading of biogasalso called second generation biogas upgradingis capturing the CO 2 contained in the biogas and converts it into CH 4 and thus recycling it, rather than plain CO 2 removal (Villadsen et al., 2019).This approach has also environmental benefits CO 2 capturing and recycling technology.Therefore, biological biogas upgrading offer huge potential in respect to feasibility, technological easiness, and potential.To increase economic advantage compared to other upgrading methods and maintain an efficient energy process, the external H 2 must be provided through sustainable sources such as water electrolysis powered by surplus electricity produced from solar/wind power during the production peak periods (Porté et al., 2019).
Developing of ex-situ biogas upgrading methods is a promising approach to optimize the overall process through using dedicated external reactors (Alfaro et al., 2018).Results from previous researches confirmed superior biomethanation efficiency of ex-situ process at thermophilic temperature over mesophilic condition (Bassani et al., 2017;Rachbauer et al., 2016;Strübing et al., 2017).Despite the low H 2 solubility and low gas-liquid mass transfer, the thermophilic temperature resulted in higher biomethanation due to increased microbial growth rates at thermophilic over mesophilic temperature demanding more H 2 (Angelidaki et al., 2018;Sposob et al., 2021).The ex-situ biogas upgrading relies on the biological activity of hydrogenotrophic methanogens whose abundance can be increased by applying bioaugmentation with pure cultures or endogenous enrichment of hydrogenotrophic communities.The application of mixed adapted cultures is more beneficial compared to pure cultures in practical applications because the H 2 adapted microbial communities are more robust and do not require sterile conditions, which would add extra costs to the process.Moreover, utilization of microbial acclimatized consortium is more efficient in H 2 conversion resulting in higher CH 4 yields compared to pure cultures (Angelidaki et al., 2018).
The main bottleneck of ex-situ biogas upgrading technology is the limited gas-liquid mass transfer of hydrogen, which strongly affects the hydrogen availability for methanogens (Strübing et al., 2017).Different reactor configurations, gas recirculation flows, installed gas diffusion devices, packing materials and stirring intensity are strongly affecting gas-liquid mass transfer and contact area between microorganisms, gas and liquid phase (Porté et al., 2019;Szuhaj et al., 2016).Many researches have been conducted to maximize H 2 gas-liquid mass transfer (Alfaro et al., 2018;Bassani et al., 2016;Díaz et al., 2015).Using trickle bed reactors (TBR), a column consisting of packing materials with high specific surface area on which biofilm is developed, has been indicated to efficiently support biomethanation (Jensen et al., 2021).The mixture of biogas and H 2 is forced through the packed bed either co-current or counter-current with nutrients containing liquid which is trickling over the packing materials.The biofilm attached on the surface of packing materials is composed of a specific arrangement of immobilized cells within a matrix of extracellular polymeric substances.This organization results in symbiotic behaviors that optimize microbial co-existence (Garrett et al., 2008).Rachbauer et al. (2016) used a TBR with an immobilized biofilm on polypropylene packing rings and achieved 96% CO 2 conversion.Strübing et al. (2017) used an anaerobic pilot-scale TBR filled with two different types of packing materials (RFK 25 L type carrier and Hel-X bio carrier HXF12KLL) and placed in polypropylene net bags and reached 15.4 m 3 CH 4 /(m 3 trickle bed .day)with methane concentration above 98%.In order to increase H 2 gas-liquid mass transfer, Jensen et al. (2021) used TBR packed with six different packing materials.They showed that clay materials facilitated H 2 overall mass transfer coefficient resulted in 99% H 2 conversion, underlying its potentials for industrial upgrading.Sieborg et al. (2020) employed a TBR packed with polyurethane foam (PUF) to increase H 2 mass transfer in ex-site upgrading process.They achieved stable conversion of H 2 at gas retention time as low as 1.32 h under thermophilic condition via hydrogenotrophic methanogenesis.Ashraf et al. (2020) used the same TBR to increase biogas upgrading efficiency.They reported that the injected CO 2 /H 2 ratio plays an important role to manipulate pH level to keep the reactor stability.Tsapekos et al. (2021) performed ex-situ upgrading using a pilot-scale TBR filled with PUF.Digested municipal biowaste was trickling as nutrient source of microbial cultures and the outlet biogas reached more than 98% purity with respect to CH 4 .In these studies, the reactors filled with packing materials, but the possible influence of packing density in a trickle bed reactor for ex-situ biogas upgrading is still unexplored.Therefore, it is important to elucidate the effect of packing materials for designing, upscaling and optimization of upgrading reactors where gases are among the reactants.Moreover, the microbial community in TBRs and their importance for efficient operation of the process remains uncharacterized.Since the efficiency of the biomethanation process is strongly dependent on a balanced microbial consortium, analysis of microbial community in biogas upgrading systems will provide essential information for process optimization.
The aim of the present research was to improve the process performance and biomethanation efficiency of TBR, by achieving better dispersion of CO 2 /H 2 in the liquid phase and subsequently, improve gas to liquid mass transfer.In this context, PUF as a micro-porous packing material with different packing densities was placed inside the TBR and the methanation performance during ex-situ upgrading process was evaluated.Furthermore, the microbial composition of biofilm and liquid media during different GRTs and especially after homo-acetogenesis incidents in TBR was explored via high throughput 16S rRNA amplicon sequencing.In addition, the trickling strategy was modified to overcome low biomethanation efficiency of TBRs created by high and non-optimized liquid trickling rate.Biochemical and microbial parameters were monitored to understand the effect of packing density and trickling strategy on gas-liquid contact and overall process performance.

Inoculum and nutrient media
Enriched hydrogenotrophic inoculum was collected from the effluent of an efficient biogas upgrading reactor (Bassani et al., 2015) and directly used to shorten the lag phase and consequently start-up period.Digestate from a manure-based biogas reactor was used to provide the essential nutrients for the microbial community.The digestate was incubated in a thermophilic incubator for at least three months to ensure total degasification and avoid any residual biogas production in TBRs.Furthermore, the digestate was sieved through a 2 mm net to avoid clogging problems during trickling.Liquid samples were taken from the liquid sump for biochemical analysis twice per week while the same volume of digestate was added to ensure sufficient nutrients supply.The characteristics of nutrient media and hydrogenotrophic inoculum used in the TBRs are presented in Table S1 (E-supplementary file1).

Reactor's setup and operation
Two gastight trickle bed reactors of 800 mL working volume were filled with different density of packing materials and operated at thermophilic conditions (54 ± 1 • C) and atmospheric pressure for 120 days.

P. Ghofrani-Isfahani et al.
The two reactors (R 1 and R 2 ) were filled with polyurethane foam (PUF) in loose (9 PUF layers) and dense (12 PUF layers) structures, respectively.The TBRs configurations are illustrated in Fig. 1.A hot water recirculation system was used to thermally insulate the reactors from outside.The liquid medium was provided from the vessels to the top of the TBRs by peristaltic pumps and trickled to the packing materials through a distributer with seven ports (2 mm diameter for each port).A synthetic gas mixture composed of 23% CH 4 , 15% CO 2 and 62% H 2 to resemble a mixture of biogas (~60% CH 4 and 40% CO 2 ) and H 2 was continuously provided from the bottom through a tube (counter-current direction).The gas mixture was constantly supplied to the reactors utilizing peristaltic pumps without any gas recirculation in the system.Indeed single-pass plug flow operation was established in TBRs.A water trap was used prior to gas production measurement to eliminate H 2 O from upgraded biogas.The experiment was performed in two periods in which the gas retention time (GRT) of TBRs was reduced by increasing the gas supply rate.The upgrading performance of both TBRs was evaluated in two different GRTs denoted as period I (GRT = 4 h) and period II (GRT = 2 h).GRT change was applied to TBRs immediately after achieving steady-state conditions in the first period.
The methane production rate in TBRs (CH 4produced ) was determined according to Eq. ( 3): where, CH 4out (L.L r − 1 .day− 1 ) is the CH 4 flow rate in the effluent biogas and CH 4in represents the CH 4 flow rate in injected gas to the reactors.

Microbial sampling and analysis 2.4.1. Microbial sampling
Triplicate samples were collected from the biofilm formed in both TBRs on the surface of packing materials and liquid media of the most promising TBR.Two liquid samples were taken from the reactors under steady state conditions in different GRTs of 4 and 2 h (nominated as PI and PII-2 in Fig. 6, respectively).In order to investigate the microbial community when the biomethanation efficiency decreased significantly, (period II, day 109), an extra liquid sample was collected during period II (nominated as PII-1 in Fig. 6).Two samples were taken from biofilm formed on packing materials in GRTs of 4 and 2 h (nominated as Biofilm-PI and Biofilm-PII, respectively).It should be noted that the gray and yellow arrows in Fig. 2 show the time of liquid and biofilm sampling, respectively for microbial community analysis.

Microbial analysis
High-throughput 16S rRNA gene amplicon sequencing was used to reveal microbial community composition.A mixture of Phenol: Chloroform:Isoamyl-Alcohol with ratio of 25:24:1 was applied to clean the samples and to improve DNA purification.The genomic DNA was extracted in triplicates using DNeasy PowerSoil® (QIAGEN GmbH, Hilden, Germany).The quality and quantity of extracted DNA were assessed using μDrop plate (ThermoFisher Scientific).
The polymerase chain reaction (PCR) amplification was performed on the V3-V4 region of 16S rRNA gene using universal primers of 341F/ 806R.The Illumina MiSeq sequencing platform was applied for next generation sequencing utilizing the 500 cycles chemistry.The raw DNA sequencing data were analyzed using the QIIME2 software (V.2021.4,QIIME; Quantitative Insights Into Microbial Ecology) which is developed based on a plugin architecture (Bolyen et al., 2019).The detailed procedure followed was previously described (Ghofrani-Isfahani et al., 2021).The raw reads were deposited in Sequence Read Archive (SRA) database (https://www.ncbi.nlm.nih.gov/sra) with accession number PRJNA772534.BLASTn search against NCBI 16S rRNA (bacteria and archaea) database was applied as an additional taxonomic alignment.Taxonomic alignment of the most abundant species, was considered based on the identity thresholds reported by Yarza et al. (2014).Heatmaps were done using Multiexperiment Viewer software (MeV 4.9.0) to present the relative abundance of the most relevant amplicon sequence variants (ASVs).The STAMP software was used for statistical analysis to evaluate the dissimilarity among the liquid media and biofilm samples and identify significant relative abundant changes.The current study is focused on the most abundant ASVs i.e. relative abundance higher than 2%, which existed in at least one sample.

Process performance and efficiency of trickle bed reactors
TBRs performance at steady-state conditions during the two operational periods are presented in Table 1.The TBRs performance was evaluated in terms of output gas composition (CH 4 %, CO 2 %, and H 2 %), methane production rate (L CH 4 L r − 1 .day− 1 ), ηH 2 and ηCO 2 .Inoculation with an enriched hydrogenotrophic culture shortened the start-up time leading to rapid establishment of biomethane in both TBRs.The biogas composition and CH 4 production rate for R 1 and R 2 during whole operation time are shown in Figs. 2 and 3, respectively.
The CH 4 concentration reached 96% (Fig. 2) in the output gas one day after start-up confirming the rapid adaptation of microbial community in both TBRs.Both reactors demonstrated high biomethanation efficiency during period I (PI) with average CH 4 content of 96 and 97% in R 1 and R 2 , respectively (Table 1 and Fig. 2).Therefore, both reactors could achieve the biomethane content fulfilled with natural gas standards, i.e. > 95% in period I (Fig. 2).The high biomethanation efficiency of TBRs during period I indicated the development of sufficient biofilm consisting specialized microorganisms for biogas upgrading.The average CH 4 production rate in R 1 and R 2 was 0.88 and 0.83 L CH 4 L r − 1 .day − 1 during period I (Fig. 3) which were 98% and 92% of the theoretical value, respectively.This indicated that hydrogenotrophic methanogenesis was the dominant metabolic process in the TBRs.The H 2 , CO 2 utilization efficiencies, pH measurements and VFAs concentration for R 1 and R 2 during whole operation time are shown in Figs. 3 and 4, respectively.The TBRs had high H 2 and CO 2 utilization efficiencies of 99% and 93% on average, respectively during period I. Particularly, R 2 showed higher methane content (97%), H 2 utilization and CO 2     conversion efficiencies (100% and 93%, respectively) compared to R 1 .It is hypothesized that the tightly packed micro-porous material (dense structure) in R 2 , allowed the formation of a stable microbial community and enabled better distribution of the injected gas mixture.Indeed, the higher density of the PUF in R 2 provided higher contact area for gasliquid mass transfer and longer retention time for gas bubbles which led to higher mass transfer and boosted biological conversion of H 2 and CO 2 to CH 4 compared to R 1 with lower density of packing materials.
Total VFA concentrations and pH values were relatively constant i.e. 0.2 g/L and 8.1 on average, respectively during period I in TBRs.The pH range is in agreement with previous researches which reported that biomethanation process can be significantly decreased after pH of 8.5 (Ashraf et al., 2020;Bassani et al., 2015).As shown in Fig. 4, acetate was most dominant individual VFAs in both TBRs in whole experimental period.Decreasing of the gas retention time at day 60 from 4 to 2 h by doubling the input gas flow rate resulted in increase of CH 4 production rate in both TBRs due to the higher CO 2 an H 2 supply to the reactors.Moreover, overall mass transfer driving force, i.e.H 2 concentration gradient between gas and liquid phase, is increased by increasing gas flow rate which could also result in higher CH 4 production (Ghofrani-Isfahani et al., 2021).From day 60 to day 93, the CH 4 production rate was stable and high with average values of 1.58 and 1.73 L CH 4 L r − 1 .day− 1 for R 1 and R 2 , respectively (Fig. 3).The GRT decrease affected negatively the biomethanation efficiency of R 1 .More specifically, the CH 4 content in the output gas decreased to 87% on average while remained relatively constant for R 2 (96%).Immediately after GRT decrease, the pH values slightly decreased in R 1 due to the lower CO 2 removal efficiency (87%) from the liquid phase, during this period.From day 94, CH 4 production rate in both TBRs started to decrease significantly down to 0.36 (day 113) and 0.62 (day 107) L CH 4 L r − 1 .day − 1 for R 1 and R 2 , respectively.Accordingly, CH 4 content for both reactors decreased down to 46% (day 106) and 52% (day 107) for R 1 and R 2 , respectively.
Total VFAs and especially, acetate concentration, started to accumulate reaching up a concentration of 1.05 (for R 1 ) and 2.17 g/L (for R 2 ) at day 106.This finding is indicating utilization of supplied H 2 and CO 2 for acetate production (i.e.homo-acetogenesis) instead of methane production (i.e.hydrogenotrophic methanogenesis).VFAs accumulation led to that the pH was reduced to 7.1 and 6.8 in R 1 and R 2 , respectively.This is in accordance with previous studies reporting acetate accumulation decreased the bioconversion efficiency of H 2 and CO 2 to CH 4 in thermophilic upgrading reactors (Porté et al., 2019;Rachbauer et al., 2017;Tsapekos et al., 2021).A possible reason for acetate accumulation and subsequently lower methane purity could be attributed to high trickling rate and liquid overflow on the top part of the TBRs leading to nutrients deficiency on the lower part of the TBRs.Therefore, in order to verify this hypothesis, the liquid trickling rate in both TBRs was decreased from 5.4 to 0.84 L. L r − 1 .day− 1 at day 110.Not only recirculation rate but also trickling strategy was changed from continuous to once per day.Therefore, the declined biomethanation efficiency was attributed to the high liquid trickling rate and structural feature of PUF as a packing material, which is a thick filling material resulting in insufficient contact between microbes with gases and nutrients.Especially, in the bottom part of packing bed materials, from where the feeding gases are supplied to the TBRs, the bad contact of microbes to Fig. 5. Monitored pH values and concentration of Total VFAs, acetate and propionate in R 1 (loose structure) and R 2 (dense structure).
P. Ghofrani-Isfahani et al. substrates (gasses) and nutrients (liquid) led to a formation of non-homogenous microbial biofilm on the TBRs bed.The insufficient biofilm formation was further validated by biofilm collection for microbial analysis.A thicker biofilm was formed on the top part of the TBRs where the nutrients were trickled compared to the bottom part.On the other hand, liquid overflow may lead to wash-out of the microbial community and subsequently lower biomethanation efficiency.
As shown in Fig. 5, immediately after modifying the trickling rate and strategy (continuous to discrete trickling, i.e. once per day), the acetate concentration started to decrease in both TBRs and subsequently CH 4 content and production rate were recovered to the earlier levels in period II (day 60-93).The average CH 4 concentration and production rate were 90% and 1.3 L CH 4 L r − 1 .day− 1 , respectively for R 1 and 95% and 1.7 L CH 4 L r − 1 .day− 1 , respectively for R 2 from day 115-120.As can be seen from Fig. 4, H 2 utilization efficiencies restored to 96 and 100% (average value) during this period for R 1 and R 2 , respectively.It could be concluded that the quick response of the TBRs to the changes in nutrient recirculation strategy confirmed the key role of efficient trickling and nutrients distribution and subsequently higher accessibility of microbes in the biofilm to the nutrients in both TBRs.
Similarly, previous studies reported high recirculation rates of liquid can damage the structure and functionality of the hydrogenotrophic methanogens (Muñoz et al., 2015) or lead to wash-out of biomass from the reactor in high rate anaerobic reactor systems (Baransi-Karkaby et al., 2020).Ullrich and Lemmer (2019) investigated the influence of liquid flow modulation of a TBR on biological hydrogen methanation.The results showed that as pause intervals without trickling became longer, the methane content increased from 89% to up to 97% at circulation intervals of 2 and 1440 min, respectively.Therefore, periodic circulation of liquid media in TBR led to improved gas-liquid mass transfer resulting in high biomethanation performance which is consistent with obtained results in this study.In addition, the periodic liquid circulation can also reduce the energy consumption of the process.Atta et al. (2014) also stated that periodic liquid flow modulation in Fig. 6.Heat map of relative abundance (>2%) for most abundant microorganisms in the inoculum and populating R 2 in the liquid media (PI, PII-1 and PII-2) and in the biofilm (Biofilm-PI and Biofilm-PII) of the reactors during the steady state condition.The color scale is shown on the top of heat map, which varies from low abundance (black) to high abundance (red).(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)P. Ghofrani-Isfahani et al. a TBR would considerably enhance the mass transfer rate and, thus the solubility of the gaseous phase within the liquid.Moreover, it can minimize uneven liquid distribution in the TBR, which strongly affects reactor performance.
In general, the better performance of R 2 (packed with dense structure) compared to R 1 can be explained by its higher density leading to efficient distribution of the injected gas mixture and higher mass transfer to the biofilm which resulted in creation of immobilized biofilm on the surface of packing materials and higher biomethanation efficiency.Dupnock and Deshusses (2017) showed that direct contact between the gas and the biofilm in a biotrickling filter was promoted by utilizing PUF as a high surface packing material and low trickling rate of the liquid.
Based on the obtained results in the current study and in accordance with previous researches in biofilm containing systems (Alfaro et al., 2018;Annachhatre, 1996;Kougias et al., 2020;Lee et al., 2012;Tsapekos et al., 2021), utilizing an efficient packing material with increased surface area for microbial activity is a key factor to have immobilized and well-structured biofilm and subsequently efficient performance of biogas upgrading process.
Overall, it can be concluded that applying a tightly packed microporous material (i.e.PUF) could be an alternative packing material for the fixed bed of trickle bed reactors ensuring efficient gas distribution and efficient biogas upgrading process in terms of methane production, conversion efficiencies and output gas quality.Furthermore, proper liquid trickling is also an important factor which significantly enhance gas-liquid mass transfer and biomethanation efficiency and ensure nutrients availability for the microorganisms.

Microbial community analysis in biogas upgrading trickle bed reactors
The microbial analysis was focused on R 2 (packed with dense structure) due to its superior biomethanation efficiency compared to R 1 .The taxonomic alignment of the most abundant ASVs found in microbial community are presented in E-supplementary file2.The relative abundance of the most abundant ASVs detected in liquid media and biofilm samples are presented using heatmap configuration in Fig. 6.
As can be seen from Fig. 6, the most abundant microorganisms in the inoculum belonged to Pseudomonas sp..The relative abundance of Pseudomonas sp.ASV1 was 5.29%.Additionally, other members within Pseudomonas genus were detected (Pseudomonas sp.ASV2-7 and Pseudomonas formosensis).Although the inoculum was rich in Pseudomonas, the relative abundance of these acetate utilizers decreased significantly in all samples taken from TBRs including both liquid and biofilm samples.Another ASV which was found to be dominant only in the inoculum was assigned to species level, Bacillus thermocloacae (99%, sequence similarity), with relative abundance of 2.34%.
Methanobacterium formicicum ASV1 (99%, sequence similarity) was the predominant microorganism in the liquid samples, which is considered as a hydrogenotrophic methanogen and is able to create syntrophic relationship with acetate oxidizing bacteria (Treu et al., 2018).The relative abundance of M. formicicum ASV1 was 3.34% in inoculum and was increased to 7.84% and 7.68% during PI and PII-2 as a consequence of high upgrading efficiency during these periods.Apart from M. formicicum ASV1, other members within Methanobacterium genus were detected (i.e.M. formicicum ASV2-7) at relative abundance of 4.20 ± 1.92% and 4.11 ± 2.15% in the liquid samples during PI and PII-2, respectively.This is in accordance with previous researches reporting the enrichment of the liquid phase with M. formicicum species (Methanobacterium genus) in various ex-situ biogas upgrading configurations using up-flow reactors (Bassani et al., 2017;Ghofrani-Isfahani et al., 2021;Kougias et al., 2020), TBRs (Porté et al., 2019;Tsapekos et al., 2021) and bubble columns (Alfaro et al., 2018;Kougias et al., 2017).As it is shown in Fig. 6, the relative abundance of M. formicicum ASV1 decreased to 4.97% for the liquid sample taken during PII-1 confirming the lower biomethanation efficiency during this period.
The predominant microorganism in the biofilm samples was Peptococcaceae sp.ASV1 which belongs to Eubacteriales order of Clostridia class, known to be involved in syntrophic acetate oxidation (SAO) metabolism (reversed Wood− Ljungdahl Pathway) to convert acetate into CO 2 and H 2 by SAO bacteria (Kougias et al., 2020).In accordance, a recently published paper exploring PUF as a packing material for biogas upgrading found that thick biofilm was mainly composed of Clostridia from Clostridiaceae family (Tsapekos et al., 2021).According to the BLASTn search, Peptococcaceae sp.ASV1 was related to Desulfofundulus kuznetsovii DSM 6115 with 91% sequence similarity.The relative abundance of Peptococcaceae sp.ASV1 was 7.13 and 8.84% in the biofilm samples collected in period I (Biofilm-PI) and II (Biofilm-PII), respectively.The remarkable abundancy of Peptococcaceae species in the biofilm can be probably attributed to their high potential to generate biofilm and possible syntrophic relationship with hydrogenotrophic methanogens.Considering VFAs concentrations (Fig. 5), acetate accumulation detected from day 95 could probably be justified by high abundance of these syntrophic acetate oxidizing bacteria.The high abundance of other members of Peptococcaceae family in the biofilm (i.e.Peptococcaceae sp.ASV2-10) with average relative abundance of 4.54 ± 2.19% and 4.79 ± 2.49% during period I and II (Biofilm-PI and Biofilm-PII), respectively suggested the possible important role of these microbes in biofilm formation in ex-situ biogas upgrading process.
Another abundant ASV in both liquid samples and biofilm was Firmicutes sp.ASV1, whose relative abundance was significantly enhanced by GRT decrease in the liquid samples by 4.65 and 4.60 folds for PII-1 and PII-2, respectively compared to period I (PI).Interestingly the proliferation of Firmicutes sp.ASV1 was rich in the biofilm with relative abundance of 3.51 and 3.05% during period I (Biofilm-PI) and II (Biofilm-PII), respectively, potentially indicating the ability of this microbe to form biofilm.This ASV was assigned to a newly discovered order MBA08, which belongs to Clostridia class.Müller et al. (2016) suggested that order MBA08 may represent SAO bacteria.Therefore, the proliferation of these acetate utilizers even in PII-1 (lowest biomethanation efficiency) can be attributed to the high acetate concentration observed during this period.The BLASTn search against NCBI database showed 89% similarity to two different species i.e.Thermoflavimicrobium dichotomicum and Aminipila butyrica which prohibited certain taxonomy alignment of Firmicutes sp.ASV1.
Other members of MBA08 order (i.e.Firmicutes sp.ASV2-4) were detected in biofilm and liquid samples.The dominance of SAO bacteria together with hydrogenotrophic methanogens confirmed their syntrophic relations which is consistent with previous researches in thermophilic ex-situ biogas upgrading process (Corbellini et al., 2018;Kougias et al., 2017Kougias et al., , 2020)).Indeed, the produced CO 2 and H 2 by SAO bacteria were utilized by the hydrogenotrophic methanogens to generate CH 4 .The enrichment of biofilm by Firmicutes sp.compared to liquid samples could be related to their higher affinity to carbon and H 2 source in the biofilm or their possible syntrophic relation with biofilm forming bacteria.High abundance of Firmicutes sp. both in the liquid phase and biofilm suggested their possible versatile metabolism.
The Ureibacillus sp.ASV1 was detected in rather high relative abundance (3.83%) in the liquid sample collected in period I (PI) with fold change of 2.97 compared to inoculum.The active presence of bacterium with Ureibacillus was previously reported in thermophilic composting of anaerobic sludge (Nakasaki et al., 2009).Based on the P. Ghofrani-Isfahani et al.BLASTn search, this ASV was affiliated to two microbial species containing Ureibacillus genus (Ureibacillus suwonensis and Ureibacillus thermophilus) with 98.5% similarity.Additional members of genus Ureibacillus (i.e.Ureibacillus sp.ASV2-6) were detected with average relative abundance of 2.99 ± 0.74% in the liquid samples during period I.The relative abundance of these ASVs decreased significantly to ≤0.1% in the liquid phase during period II and Ureibacillus sp.ASVs were not detected in biofilm samples.Jønson et al. (2020) and Ashraf et al. (2020) also reported the presence of Ureibacillus genus in biotrickling filter used for biogas upgrading process.
Alicyclobacillaceae sp.ASV1 and ASV2 were observed in high relative abundance of 3.63% and 4.17% in liquid samples obtained during PII-1, respectively.This ASV was assigned to an uncultured bacterium of the order OPB54 belonging to Clostridia class within the phylum Firmicutes.The highest relative abundancy for OPB54 was detected in the liquid sample collected in PII-1 when lower methanation efficiency occurred which is in accordance with highest levels of acetate concentrations.Therefore, Alicyclobacillaceae sp.ASV1 may be involved in acetate metabolism, but the ecological role of this group of bacteria in thermophilic AD is still uncertain (Huber et al., 2021).Furthermore, the relatively high abundance of ASVs assigned to Alicyclobacillaceae family in the liquid samples during PI (1.52% on average) and PII-2 (2.09% on average) revealed the ability of these species to create from syntrophies with hydrogenotrophic methanogens in biomethanation process.In general, the existence of diverse species of SAO bacteria and hydrogenotrophic methanogens in the microbial community of TBRs during ex-situ biogas upgrading confirmed their potential syntrophic relationships as previously were reported (Angelidaki et al., 2018;Kougias et al., 2020).

Conclusions
The present work showed that a packing material characterized by a thick and dense structure can improve hydrogen gas-liquid mass transfer and biomethanation efficiency.Trickle bed reactor (TBR) packed with more layers of polyurethane foam (PUF), i.e. dense structure, achieved higher methane content and H 2 utilization efficiency equal to >95% and >99%, respectively.The results revealed the crucial role of packing material and its properties on biogas upgrading efficiency.Proper trickling strategy was also shown to be an important factor in TBRs as it increases the resistance of reactor against liquid overflow and uneven gas-liquid distribution along the bed.In addition, efficient trickling strategy was able to recover the TBRs performance after homoacetogenesis incidents.The microbial analysis revealed that the most abundant microbes observed in the biofilm belonged to Peptococcaceae family.Moreover, the biofilm was also dominated by a newly discovered order MBA08 assigned to syntrophic acetate oxidizers.This argument demonstrates that anaerobic digestion microbiome is strongly stimulated during hydrogen assisted methanogenesis.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

•
Polyurethane foam (PUF) with different densities were packed inside the TBRs.• Dense PUF substantially enhanced biofilm immobilization and H 2 mass transfer rate.• Biofilm was predominated by syntrophic acetate oxidizing bacteria.

Fig. 1 .
Fig. 1.Schematic diagram of the trickle bed reactors; R 1 and R 2 were filled with polyurethane foam (PUF) in different packing densities i.e. loose and dense structures, respectively.

Fig. 2 .
Fig. 2. Biogas composition (CH 4 , H 2 and CO 2 concentration) in the effluent gas during the whole experiment for R 1 (loose structure) and R 2 (dense structure); arrow and circle symbols indicate DNA extraction for R 2 from liquid media and biofilm, respectively.

Table 1
TBR performances under steady state condition.R 1 and R 2 packed with micro-porous packing materials in loose and dense structures, respectively.
a Period II-A (days 60-93) when the biomethanation efficiency was high and stable at GRT of 2 h.b Period II-B (days 115-120); when the TBRs recovered from homo-acetogenesis (low efficiency) by modifying the trickling rate and strategy at GRT of 2 h.P.Ghofrani-Isfahani et al.