To evaluate microbial biomass’s performance, herein, acetolactic methanogenic activity, an assay called specific methanogenic activity (SMA) [29] was applied. A batch experiment was performed for the initial inoculum and experimented biomass samples collected at the end of the experiment. SMA assay is an indirect measurement using to evaluate a microbe’s performance in utilizing a specific substrate. The unit of SMA was expressed as g of substrate calculated in terms of COD per g of biomass calculated as VS per day (g COD/g VS.d). In practice, anaerobic batch digestion was carried out in 125 ml serum vials with a liquid working volume of 80–100 ml. These vials seeded with tested inoculum and feed with a specific substrate to target a group of microbes; acetate for MPB and glucose for non-MPB, at 0.1–0.5%. About 0.1 to 0.3 of food to microbes (F/X) was applied. In each vial, the nutrient solution in the formula of Smolder et al. [30] was added after adjusted pH around 6.8–7.2. The vials were then flushed with a mixed gas of nitrogen and carbon dioxide (70:30), capped with butyl rubber stoppers and aluminium crimps. All tested bottles were mixed by hand twice daily and incubated at a controlled temperature (37±1.0°C) incubator. The produced biogas and its composition were regularly monitored. Each tested condition was done in triplicate, and the controlled substrate and inoculum vials without substrate addition were also operated.

The number of eukaryotic bacteria (EuB) and archaea (Arc) were enumerated using the most probable number (MPN) technique with five-fold serial dilution [28]. About 3–5 grams of samples were diluted in a phosphate buffer solution with one-nine fraction in the tube in practice. It was then mixed by vortex and sonication. The dilution (10⁻⁵–10⁻¹⁴) of the stock sample were prepared, and the sample was then transferred into five tubes containing the selected culture media specifically different from enumerating EuB and Arc. The culture media was prepared accordingly to Zhang and Noike’s formula [31] with some modifications. For EuB, to easily monitor the operational change of reaction, phenol red was added as a pH indicating sign. The pH was reduced when acids were produced from the substrate digestion of microbes. The colour turned significantly from primarily red to yellow. While acetate at a concentration of 20 mM was added to the medium of Arc cultivation. The tested tube was incubated and analyzed for 7–15 d, depending on the microbial group. To verify the positive condition, the tested tube that depicted a development growth sign by changing of colour or reducing of acetate, which verified by GC for EuB and Arc, respectively, was used as the criterion. The undeveloped tube was defined as a negative. An MPN table [28] was used to determine the number of EuB and MPB.

To identify the existence of methanogenic Arc in WAS, a 16S rDNA library was performed. The experiment divided into four steps, including 1) DNA extraction, 2) PCR amplification and clone library, 3) PCR-DGGE classification, and 4) DNA sequencing and analysis. The operation of each step was summarized. Step 1 DNA extraction: DNA extraction was based on Zhou et al. [32] using SDS with some modifications. Two replicated samples at 0.5 g (wet weight) of homogenized sludge was mixed with 700 mL of DNA extraction buffer (100 mM Tris-HCl, pH 8.0, 100 mM sodium EDTA, pH 8.0, 100 mM sodium phosphate, pH 8.0, 1.5 M NaCl, 1% CTAB) and 50 μL of lysozyme (10 mg/mL) in 2 mL of microcentrifuge tube with horizontal shaking at 200 rpm, 37°C for 30 min. After shaking, 100 mL of 20% SDS was added, the sample was then incubated at 70°C for 30 min. After the centrifugation, the supernatant was collected and transferred into a new 2.0 mL microcentrifuge tube. The supernatants were combined and mixed with an equal volume of chloroform-isoamyl alcohol (24:1). The aqueous phase was recovered by centrifugation and precipitated by isopropanol at room temperature for 4 h. The pellet of genomic DNA was obtained by centrifugation at 16,000 × g for 20 min at 4°C. It was washed with 70% of ethanol and re-suspended in sterile deionized water. Finally, genomic DNA was purified further using an extraction kit for Gel/PCR DNA fragments (Geneaid, Taiwan) as specified by the manufacturer. Step 2 PCR amplification and clone library: The nearly complete 16S rDNA of Arc was amplified using universal primers A20F [33] and U1492R [34]. The PCR conditions were (i) 5 min at 95°C, (ii) 30 cycles of 30 s at 95°C, 30 s at 55°C, and 2 min at 72°C, and (iii) 7 min at 72°C. Each amplicon of 16S rDNA was ligated to pGEM-T easy vector and transformed into E.coli DH5α. Step 3 PCR-DGGE: PCR-DGGE was selected to provide an initial indication of genetic diversity. Practically, 16S rDNA clones of Arc libraries were extracted for plasmid using the alkaline lysis method. The specific primer for specific variable region V3–4 of Arc (344GC-F/522R) [35] was applied to increase the number of DNA fragments. PCR products were then classed by Denaturing Gradient Gel Electrophoresis (DGGE) [36]. About 6.5% of Acrylamide gel (40–55% Denaturant gradients) in 0.5X TAE at 200V was operated for 5 h (C.B.S. DGGE model 2000 system, CA). Then, SYBR® Gold was applied for gel-colouring to visualize the DNA band’s position under UV light conditions. Clones with the same band position on DGGE gel were grouped into so-called operation taxonomic units (OTUs). Step 4 DNA sequencing: The clone in each OTUs was then submitting to analysis for DNA sequencing, and the obtained results were compared to the database using a BLAST search in NCBI [37].

The morphological observation was conducted to visualize the microbes’ characteristic in WAS, focusing on methanogens. The initial WAS sample was compared with the experimented biomass samples of AHR1 for the morphology visualizing by scanning electron microscopy (SEM). The initial WAS and biomass samples of the suspended and attached bed of AHR1 were collected. All samples were prepared and submitted for SEM analyses (JEOL/Oxford SEM model JSM-IT300) at the Scientific and Technological Research Equipment Center of Chulalongkorn University.

The operational performance of AHRs inoculated differently by WAS and co-culture of WAS with the other inoculums is summarized in Fig. 1–3 and Table 3. The different start-up times to accomplish about 4.0 kg COD/m³.d were promoted in each reactor, taking about 1.5–3.3 months. In Fig. 1, an organic loading rate applied to each AHRs is showed. It was observed that the organic loading rate of AHR1-WAS and AHR2-WAS-PM was slow. It took almost 70–98 d or 2.5–3.3 months to achieve 4.0 kg COD/m³.d compared to 1.5–1.7 month of AHR2, AHR3, and AHR5. The unbalance of acid production and utilization capacity was a crucial indicator in this time-consumed factor. In general, during the start-up period, the balance of acid production by non-methanogenic bacteria and acids utilization by methanogens is the key. In wastewater treatment, methanogen is typically limited due to the growth limit rate and affected the digester’s overall capacity. This capacity compares the ability of microbial biomass per specific mass to grade incoming substrate, which varies not only by feeding but also by remaining or excess organic matter in biomass. Considering HRT, it corresponded well with the loading rate. The HRT was reduced relative to OLR increase from about 13 d to 4 d subsequently. The rapid reduction in HRT or an increase in OLR in AHR1 and AHR2 harmful to its stability. This reduction caused the VFAs to surge production and accumulation towards this initial period. A longer time was needed to maintain the previous loads and stability. This period was an adaptation required for microbial acclimatization and increase population number to combat with the remaining substrate originally contained in WAS.

It was stated previously that WAS and manures contained high organic content. This remaining organic was an obstacle to feed a new substrate. After digestion was prolonged for 1–2 cycles, the delay in organic loading capacity was observed. AHR1-WAS performed a slightly lower organic loading rate. This unfavourable condition associated directly with an accumulation of VFA and a dropping of pH, particularly on the 18th–25th of operation. The pH of AHR1 was reduced continuously from the initial values of 7.5 to 6.5 within 20 d. Thus, the OLR of AHR1-WAS was intermittently reduced to maintain its stability. After a week of adaptation, the reactor could recover its previous stability and load. Similar to AHR1, AHR2-WAS-PM was instability promoted. A production of VFAs was rapidly and raised an accumulation of acidic intermediates initially for 1.9 to 2.5 g/L in day 10th to day 20th. This rapid production of volatile fatty acids promoted acidic conditions that may be caused by the higher acid production activity of PM depicted as GUR for 31.1 g COD/g VS.d. Although the OLR of AHR1 and AHR2 was lower than the other at the same period, these reactors still demonstrated a satisfying performance in COD removal efficiency. A similar performance was found for other reactors. This finding was the same as previous works of Benabdallah et al. [23] and Kobayashi et al. [41]. They reported the importance of inoculum’s quality for the start-up of anaerobic digester. Fig. 1 shows the performance of AHRs indicating by organic loading rate with time.

The reduction of organic constituents in wastewater directly affects by anaerobic microbial performance in the digester [7]. Results showed that inoculum performance was correlated with the digester’s performance, especially at the initial stage of start-up. The different removal efficiencies were found significantly (p < 0.05). At OLR of 0.5–1.0 kg COD/m³.d, similar COD removal efficiency (COD_rem) were observed in almost all AHRs: 86.4–95.4% and 92.2–96.4% for TCOD and SCOD, respectively. The performance corresponded to the remaining organic substance inside each inoculum itself, particularly WAS. AHR1 and AHR2 depicted a slightly low rate of removal performance. These removal efficiencies were increased subsequently even when the OLR increased. At OLR of 4.0 kg COD/m³.d, the similar performance of COD removal efficiencies of 89.4–95.4% and 92.2–96.4% for TCOD and SCOD, respectively, were observed in all reactors. The effluent concentrations were about 1.0–2.5 g TCOD/L or 0.8–1.7 g SCOD/L. These results depicted that microbial performance in each AHR was appropriately developed with time and organic loaded. The previous results also stated similarly, the importance of seed inoculum development of UASB [19] and an anaerobic fixed film reactor [16] were related to time and inlet feed. It was worth noting that the effluent concentration of AHR1 was high initially by the remaining substrate of WAS. This substrate also promoted organic acids accumulation, which needed a longer time to acclimatize and utilize these intermediates, as stated previously. In Fig. 2, the COD concentration in the effluent of each AHR is shown.

In comparison, the increase of OLR was more rapid in AHR3-WAS-CD and AHR5-WAS-AN. The benefit of the MPB populations was positive in these digesters. The unfavourable condition promoted by the non-methanogenic population’s activity found contrastingly for co-inoculation with PM (AHR2-WAS-PM). The promotion of VFAs and their accumulation occurred. The balance of acid production and utilization was then vital to the stabilization of excess VFAs. The importance of the methanogenic population in seed inoculum on stabilizing the variation during start-up anaerobic digester had been studied previously. The development of microbes that could combat a high concentration of VFAs, mainly HAc, was also stated [38]. A large microbial community, particularly the methanogenic diversity in cow manure, had been reported. While the opposite conclusion was stated by Maroun and El-Fadel [41]. They noted that the anaerobic digester’s start-up was possible without external seed inoculation instead of itself microbial composition. The microbial self-inoculation was also proposed by others elsewhere for sludge digesters [42]. However, the difficulty in maintaining the operating and controlling stability for correct digestion was observed.

The gases production was differed from reactor to reactor depending on the inoculation, operational performance, and time of each reactor. However, at the end of the start-up period, about 6.3–9.6 L/d of biogas was produced in each AHR. At an initial of 0.5–1.0 kg COD/m³.d, the biogas production rate was low, ranged between 1.1–2.5 L/d for all reactors, but the difference in their methane content was observed. This variation was depending on organic load and inoculation differently. The produced biogas contained different and various compositions of methane for 21.9–38.0%. Then, these daily gas productions increased with OLR increase. At OLR of 4.0 kg COD/m³.d, the maximum gas production of about 6.3–9.6 L/d was promoted. The methanogenic performance indicated as methane composition in produced biogas was increased relative to digestion time. A proper adaptation of microbes was observed significantly. Fig. 3 shows the biogas production and methane content of produced biogas in each digestion tank.

The methane productivity of each AHR differently varied by OLR and time. About 0.08–0.15 L-CH₄/g COD_rem was obtained at the initial stage. Then, these yields were increased and were stable at 0.21–0.32 L-CH₄/g COD_rem. The maximum methane productivity was orderly as AHR5, AHR3, AHR2, AHR4, and AHR1. A lower increase in gas production yield was observed in AHR1, corresponded with other performance indicators, i.e. COD_rem and loading rate. A similar finding was also found in the early stage of the start-up in many works. A specific substrate regulated this to the methanogenic population. It was noticeable that the co-inoculation of WAS and other inoculums favoured a stable and productive performance of methane production seen by methane content in the produced biogas, particularly AHR3, which inoculated by WAS and CD. The acceptable quality in the excellent background history of starter seed resulted in the rapid development of its methanogenic performance. The higher methane content was found for co-inoculation rather than mono sludge inoculation, particularly AHR4 and AHR5. This result was similar to Oz et al. [43], who concluded that the increase of inoculated seed’s methanogenic population was empirical and indicated methane production activities during the start-up operation. This production was well agreed with famous characterization confirmed as an intensive concentration and profitable production of MPB in ruminal manures. Moreover, the different loading capacity and time to complete the OLR of 4.0 kg COD/m³.d were promoted. It was found that almost two months was faster when inoculating with WAS and CD than WAS only. While the other manure depicted a similar performance and achieved about 25–35 d more quickly than AHR1. However, the long term operation performance and stability of each digester should be carefully considered. The ultimate loading capacity of each reactor was related closely to its initial start-up stage.