3.1. Overall reactor performance
Reactor performance in terms of pH and CH4 production rate largely varied depending on the conditions (Figs. 2a and 2b). The pH value in an anaerobic digestion reactor is used as a simple indicator of process stability, and the appropriate pH range is 6.5–8.2 [19]. In this experiment, the pH values ranged from 6.8 to 7.5 for both the biochar- and sponge-operated reactors throughout the whole experimental period (Fig. 2a). Similarly, the CH4 production in both carrier-filled reactors was generally stable between 1.0 and 1.3 L L− 1 day− 1 (Fig. 2b). In contrast, the pH and CH4 production rate in the control condition suddenly dropped around day 40 and continued to decrease until the end of the experiment. This decrease in pH was probably due to the accumulation of undegraded intermediates.
The DOC concentration in the effluent was very stable at around 1000–1200 mg L− 1 in the biochar- and sponge-filled conditions, leading to high TOC removal efficiencies above 80% (Fig. 2c). In the control condition, the DOC concentration increased, leading to a decrease in TOC removal efficiency from day 33 onward. VFAs are the main contributors to acidification in anaerobic digestion. When the VFA production rate exceeds its decomposition rate, the pH decreases and process instability or failure can occur [19, 20]. In the biochar- and sponge-filled conditions, the VFAs did not accumulate and were stable between 300 and 600 mg L− 1 until day 49 (Fig. 2d). Meanwhile, in the control condition, the accumulation of VFAs began after day 30 and continued to increase exponentially until day 44 from which they became stabilized afterward.
Throughout the experimental period, the carrier-filled conditions showed higher CH4 production, organic carbon removal efficiency, and pH stability than the control condition. The main reason for this is the development of a biofilm on the carriers’ surface, which contains a microbial community with a high population density. It was reported elsewhere [21]that a biofilm holds 10 times more microorganisms than suspended sludge and, thus, the formation of the biofilm enables rapid degradation of VFAs by the attached microbial community. The biofilm allows for a longer average residence time, which provides beneficial acclimation conditions for the microorganisms, which could improve the pH stability and resistance to less advantageous environmental conditions such as acidic environments [11]. Therefore, the performance of the reactors containing the carriers did not decrease even when the acidic WHJ of pH 4.02 was being supplied. In contrast, in the control condition, all microorganisms were present in the suspended sludge and the density of microbes in the reactor may have decreased day by day due to substrate supply and sludge discharge. In addition, the C/N ratio of the substrate was 38, which is higher than the sustainable range of 20–30 for anaerobic digestion [22]. Substrates with high C/N ratios inhibit microbial growth because there is insufficient nitrogen for microbial cells to properly function, leading to VFA accumulation and reduced pH [23, 24]. Although SBR is characterized by relatively low bacterial efflux, the substrate in this experiment limited the growth of the microorganisms, resulting in acidification of the control reactor without a carrier.
The carbon mass balance calculated from the results on days 35–49 showed the difference in the performance of each reactor (Fig. 3). In the biochar- and sponge-filled conditions, biogas accounted for nearly 90% of the total carbon and the substrate was well converted into gas. In contrast, the control condition showed a high value of the “others” component (36%), clearly indicating inefficient substrate conversion to biogas. These results indicate that biochar and sponge were appropriate carriers for stable anaerobic digestion of WHJ.
Table 1 summarizes the different studies on the anaerobic digestion of water hyacinths or WHJ. Compared to those in other studies, the substrate used in this study had the lowest pH value. Another difference is that the C/N ratio of the substrate in a previous study was 23 [6], which is suitable for anaerobic digestion, whereas the C/N ratio of the substrate in this study was 38. Normally, such low pH and high C/N ratio of substrates can inhibit anaerobic digestion reactions and reduce reactor performance [19, 23]. Despite this, this experiment showed the highest CH4 production rate of 413.8 mL g COD− 1 and a relatively high organic removal efficiency of 90% (TOC) compared to other experimental studies. The high treatment performance even with substrates being unfavorable to anaerobic digestion suggests that the use of the coffee husk–derived biochar carrier may provide superior operating conditions in the treatment of other WHJ.
Table 1
Comparison with previous studies
Substrate | Reactor type | Substrate characteristics | OLR (g-COD L− 1 d− 1) | CH4 production rate | COD (TOC) removal (%) | References |
pH | COD (mg L− 1) | mL g-COD− 1 | mL g-VS− 1 |
WH | CSTR | n.d. | n.d. | 0.73 (g-VS) | n.d. | 240 | n.d. | [25] |
WH | Two-stage | 5.8 | n.d. | 3.75 | 146.4 | n.d. | 73 | [26] |
Pretreated WH | 3.75 | 340.8 | 82 | [26] |
WHJ | UASB | 7.0 − 7.5 | 15630 − 23500 | 8.85 | 150 | n.d. | 86 | [6] |
WHJ | MIC | 17.93 | 210 | n.d. | 82 | [6] |
WHJ | Batch | 6.8 | 5960 | n.d. | n.d. | 635 | n.d. | [27] |
WHJ | SBR (biochar) | 4.0 | 29617 | 2.96 | 413.8 | 471 | 90 (TOC) | This study |
*Water hyacinth: WH, water hyacinth juice: WHJ, chemical oxygen demand: COD, hydraulic retention time: HRT, organic loading rate: OLR, and volatile solid: VS. |
*Reference: [25]Hanisak et al., 1980, [26]Barua and Kalamdhad, 2019, [6]Liu et al., 2020, [27]Hudakorn and Sritrakul, 2020. |
*n.d. : Not determined. |
*UASB: up-flow anaerobic sludge blanket, MIC: modified internal circulation, CSTR: continuous stirred tank reactor, SBR: sequencing batch reactor. |
3.2. SEM observation
The results of the SEM observation of each carrier before experimental use are shown in Fig. 4. The surface of the biochar carrier was covered with numerous latticelike irregularities and the size of the pores was about 20 µm (Figs. 4a and 4b). The sponge carrier had very large voids (about 200–600 µm) and a smooth fiber surface (Figs. 4c and 4d). SEM observation of the carriers on day 49 showed biofilm formation on both carriers (Fig. 5). On the surface of the biochar, aggregation of microorganisms was observed, mainly rod-shaped microorganisms in the pores (Figs. 5a and 5b). In particular, the pores of the biochar held microorganisms. Similarly, Ma et al. [28] reported that the pores of biochar acted as a shelter for microorganisms. Microorganisms and biochar are easily bound by van der Waals forces due to their hydrophobic surfaces [29]. In the sponge carriers, clumps of sludge were held in their voids and the filamentous microorganisms were agglomerated (Figs. 5c and 5d). The tight attachment of the microorganisms to the sponge carriers was attributed to the hydrophilic groups and cationic surfactant active groups of polyurethane [30]. Although the microbial aggregation process of the biochar and sponge carriers may have been different, both successfully formed biofilms, indicating high treatment performance.
3.3. Microbial community
Fifteen bacterial phyla were detected to be more than 1% in the seed sludge with Bacteroidetes (25.8%), Proteobacteria (13.2%), Cloacimonetes (12.8%), and Firmicutes (12.2%) predominating (Fig. 6a). Bacteroidetes and Firmicutes were predominant in all the reactors on day 49 (58.0–83.4%). Bacteroidetes and Firmicutes are bacterial phyla commonly found in a typical anaerobic digestion reactor and are responsible for the degradation and hydrolysis of complex macromolecules by releasing extracellular enzymes [31]. The relative abundance of both phyla was significantly higher in the biochar- and sponge-filled conditions (58.8% and 47.2%, respectively). The pores of biochar have been reported to act as a shelter for microorganisms and enrich these species [32]. Bacteroidetes and Firmicutes which increased by the addition of the carriers may have contributed to the rapid degradation of SS in the WHJ. In addition, Firmicutes are capable of degrading acetate and the low presence of Firmicutes leads to the accumulation of acetate due to reduced consumption [33]. This is consistent with the lower relative abundance of Firmicutes in the control reactor compared to the other reactors and with the higher resultant concentration of VFAs.
In total, 12 genera of archaea were detected to be more than 1% in the substrate, seed sludge, and reactors (Fig. 6b). Methanosaeta and Methanolinea were dominant in the seed sludge (23.5% and 38.2%, respectively). These two genera were also dominant in all the reactors, where their relative abundance was the highest in the control condition (82.0%) and moderate in the biochar- and sponge-filled conditions (53.0% and 51.5%, respectively). Biofilm carriers provide a habitat for microorganisms and retain them due to their porosity or large voids [34, 35]. In anaerobic digestion, archaeal diversity is directly related to reactor stability, and a decrease in diversity has been observed in loaded environments for microorganisms [36]. Differences in the relative abundance of Methanosaeta and Methanolinea occupying each reactor indicate that the addition of the carriers improves the archaeal diversity and performance stability. Unclassified Woesearchaeia (class level) and Candidatus Diapherotrites archaeon ADurb.Bin253, a type of Woesearchaeia, were also detected in the WHJ and reactors (Fig. 6b). Woesearchaeia is one of the classes in the phylum Nanoarchaeaeota. Nanoarchaeaeota have extremely small genomes and cell sizes and are generally known to live as parasites or symbionts with other archaea [37]. Woesearchaeia is thought to be a heterotroph that degrades organic matter to produce acetate and hydrogen and thought to live in symbiosis with methanogens and is more commonly detected in soil than in water [38–40]. Woesearchaeia was the dominant archaeon among the WHJ archaea possibly because the water hyacinths in this experiment were rooted in a swampy area and had soil attached to them. Candidatus Diapherotrites archaeon ADurb.Bin253 detected in the seed sludge and reactors has never been reported to be detected in anaerobic digestion. Candidatus Diapherotrites archaeon ADurb.Bin253 was present in low relative abundance (1.1%) in the seed sludge, but its abundance significantly increased in all the reactors during the experiment (10.5–28.6%). Nanoarchaeaeota were reported to increase significantly when fed substrates had high C/N ratios (> 22.0) rather than low C/N ratios (< 16.7) [41]. It is possible that the high C/N ratio (38.1) and the supply of swamp-derived WHJ in the reactors created a suitable environment for Candidatus Diapherotrites archaeon ADurb.Bin253 to grow.