Investigating the Impact of Organic Loading Rates and Magnetic Nanoparticles on the Performance and Stability of Continuous Stirred Tank Reactors

Abstract

Research on energy demand is advancing, with the addition of nanomaterials in anaerobic digestion increasing stability, accelerating hydrolysis, and reducing microbial inhibition. However, further research is needed to determine the mechanisms, ideal dosages, and long-term impacts. This work used continuous stir tank reactors (CSTRs) to experimentally examine the biocompatibility of iron oxide nanoparticles (Fe₃O₄-NPs) at a concentration of 75 mg/L at various organic loading rates (OLRs) of 0.3, 0.8, and 1.3 gVS/L.d (CSTRs). The efficiency of the reactors was observed by considering various parameters, such as pH, soluble chemical oxygen demand (sCOD), TVFA formation and degradation, total solids (TS), and volatile solids (VS) removal, as well as methane (CH₄) generation. Hence, it was found that the reactor with added NPs (R1) yielded an optimum 725.9 mL/gVS of CH₄ and this was achieved at the lowest OLR of 0.3 gVS/Ld. However, another reactor (R2, without NPs), exhibited more stabilized results, ranging from 372.8 to 424.4 mL/gVS at 0.3 to 1.3 gVS/Ld of OLR, respectively. Therefore, in R1, the maximum removal of sCOD, TVFAs, and VS was achieved at 90%, 74%, and 93%, respectively, as compared to R2.

1. Introduction

The exploration of renewable energy sources and the harnessing of anaerobic digestion (AD) technology have become considerable points of interest in research in the last few decades [1]. Moreover, AD technology offers global benefits such as a reduction in greenhouse gases, utilization of waste (organic fraction) for energy recovery, and nutrient-rich digestion for agriculture. The addition of nanomaterials in the AD process has brought advancements in process stabilization, accelerated hydrolysis, and methanogenesis, as well as providing resistance against inhibition-causing microbes, etc. The AD of organic waste is preferred and termed as the most efficient method over the other methods available to treat it [2,3]. AD is preferred because it yields bioenergy (i.e., biogas/methane) [4], nutrient-rich digest for agricultural purposes, and has simple environment-friendly operating procedures [5].

In the past, numerous methods, such as pretreatment of the substrate with chemicals [6], hydrothermal and ionizing radiation [7], physical size reduction by any mechanical means [8], and biological means [9], were adopted to stabilize further the AD process for better disintegration of complex organic waste into biogas.

AD is a multi-biological-stage method [10] to treat organic waste at mostly mesophilic [11] and thermophilic temperature ranges [12]. However, factors like pH, temperature, organic loading rates (OLR), and volatile fatty acid (VFA) concentration are pivotal considerations in the AD process [10]. Moreover, the AD process can be carried out in batch or continuous mode. In continuous mode, the AD reactors work on the fill-and-draw method, where the OLR plays a significant role in process stabilization and calibration.

Ferguson et al. [1] revealed that AD process instability could also be linked to sudden changes in feedstock composition or OLR. Hence, the optimization of OLR concerning the waste composition to produce optimum methane or biogas in reactors remained the main objective of the various studies around the world. Furthermore, an imbalance of OLR in the reactors yields a higher formation of volatile fatty acids (VFAs) and an ammonia concentration that exceeds the permissible limits reported by [13].

In the last decade, the addition of nanoparticles to the AD process has been much researched by researchers around the world [2,14,15,16,17] who have studied the impact of NPs on biogas and CH₄ production in the AD process. Studies have reported that the addition of NPs in AD is presumed to be a biodegradation accelerator [16], CH₄ or biogas-production booster [18], and cited as the pH buffer [19]. NPs possess a large surface area, self-assembly and dispersibility, high reactivity, greater specificity, and good biocompatibility. However, their unique properties are due to their small size and the chemical reaction of the nanoparticles that increases reactivity. They increase anaerobic digestion by enhancing microbial and enzyme activities and electron transfer. Conductive NPs allow direct interspecies electron transfer (DIET), which accelerates methane generation. Additionally, metal-based nanoparticles provide vital micronutrients that improve microbial metabolism, whereas others support the stabilization of pH, buffer redox conditions, and absorb harmful chemicals such as ammonia and sulfide. These benefits result in a more efficient, steady, and robust digesting process [20,21,22]. Microbial fluctuation control is considered a significant parameter for the success of AD treatment.

The NP addition to the AD process is cited as the cardinal property of biological treatments. These nanomaterials enter the cell membrane and assist the route across biological barriers between the distribution and metabolism and maximize the immune systems’ uptake and adsorption of the microbes in an oxygen-free environment [16]. Meanwhile, the authors of [23] reported that CH₄ enhancements increased from 16.1% to 180% with the addition of Fe₃O₄-NPs. Wo et al. [24] cited that the addition of iron boosts electron transfer, the activity of bacteria, and methanogenesis, keeping digestion more efficient.

The objective of this study was to investigate the behavior of CSTRs in treating the organic fraction of municipal solid waste (OFMSW) in AD at different OLRs and supplement it with the optimized NP (Fe₃O₄) concentration. As a result, several operating conditions were designed to evaluate the influence of NPs in the continuous mode of AD, including CH₄ formation, bacterial responses, and the removal of other chemical substances such as total solids (TS), soluble chemical oxygen demand (sCOD), and total VFAs. The statistical influence of NPs in the AD process was investigated, with an emphasis on the impact of NPs on CH₄ production at various concentrations.

2. Materials and Methods

2.1. List of Materials and Instruments

Salts such as Fe (ii) and Fe (iii), ferrous chloride tetrahydrate (FeCl_2·4H₂O), ferric chloride hexahydrate (FeCl_3·6H₂O) were used as the precursor materials for the synthesis of nanomaterials, whereas sodium hydroxide (NaOH) was used as the pH controlling buffer and was obtained from Dae-Jung Chemicals & Materials Co., Ltd., Republic of Korea. Zeta potential instrument (Malvern Instruments Ltd., serial number: MAL-1,135,362) was employed for analysis of zeta potential, size distribution and composition of materials and atomic force microscopy (AFM).

2.2. Optimization Methods for AD Reactors

The optimized concentration of 75 mg/L of iron oxide (Fe₃O₄) was used in this study at mesophilic conditions (37 ± 1.5 °C) in CSTRs; a schematic diagram is shown in Figure 1. Stainless steel reactors with automated control panels were used. This concentration of NPs was optimized in our previous study using different concentrations and the optimum CH₄ generation was achieved at 75 mg/L [19]. This approach was adopted because several studies have reported the positive impact of NPs on bioenergy generation in the AD process. However, NPs may function as catalysts or promote the growth of microbes involved in the AD process. Additionally, NPs can accelerate the AD system, resulting in increased generation of biogas. This helps in breaking complicated organic molecules into simpler chemical compounds that bacteria can digest [25,26].

During the experimentation, the required temperature of the reactors was maintained by circulating water through a water jacket. A mixture of cooked and uncooked food waste (FW) was utilized as the substrate for this study, while the effluent from a buffalo dung biogas plant was added as the inoculum to initiate microbial growth. The characterization of FW and inoculum was carried out by conducting analytical tests, such as total solids (TS), VS, soluble chemical oxygen demand (sCOD), and total volatile fatty acids (TVFAs), according to the standard methods [27]. The characterization of both substrate and inoculum was reported in a previous study by [21]. The collected FW was shredded manually and then mechanically using a chopper (AG-3044 Deluxe Chopper). The results presented in this study are the average of two runs of each bioreactor (R1 and R2).

The collected inoculum and prepared substrate (FW) were stored at 4 °C before use, as [28] discussed. Furthermore, pH and temperature were monitored regularly and electrochemically measured with an Ingold pH electrode (Zhengzhou city, Henan, Taiwan) and thermometer (GH Zeal LTD, London, UK). Sodium hydroxide (NaOH, Daejung, Republic of Korea) was used as a neutralizer to raise the pH during the process. VFAs, sCOD, and microbial analysis were determined following the procedure cited by [29,30].

2.3. Synthesis of Fe₃O₄ NPs

The magnetite nanoparticles were prepared using the chemical precipitation method [31,32]. In the first step, the stock solution of FeCl₂. 4H₂O and FeCl₂. 6H₂O with 0.1 M, respectively, were added and mixed in 10 mL of ultrapure water. Then, the homogenized mixture was allowed at 85 Co for 15 min on the magnetic stirrer hot plate. Later, 0.5 M of NaOH concentration was added until the pH of the solution reached the range of 9–12. The mixture was continuously stirred during the process until the precipitate formed. The resulting precipitates were collected and washed three times with ultrapure water and ethanol to remove the impurities of synthesized nanoparticles. Finally, the prepared nanoparticles were dried in an oven at 150 °C for 4 h. The characterization of Fe₃O₄-NPs was carried out using UV–visible spectroscopy (Shimadzu, Kyoto City, Japan, UV-199i), transmission electron microscopy (TEM) (JEOL, Musashino, Tokyo, Japan, JEM 2100) and atomic force microscopy (SPM-8100FM, Shimadzu, Kyoto, Japan).

2.4. Reactors Working Mechanism

In this study, the reactors were loaded with three organic loading rates (OLRs), i.e., 0.3, 0.8, and 1.3 gVS/L.d, to monitor the behavior of the reactors with Fe₃O₄-NPs with a concentration of 75 mg/L (R1) and without NPs (R2). The volume of each reactor was 10 L, and the working volume was set at 7 L during the experiment. The volume of CH₄ generation was calculated daily based on the downward displacement of water. The acryl-based 15 L gas jars were prepared for the gas measurement. Both the reactors started loading from 0.3 to 1.8 kgVS/L.d, and the whole process was carried out for 120 days. The reactors were operated using the fill-and-draw method. Every day, diluted substrate was added, and digested slurry was removed from the reactors for the analysis of TS, VS, VFAs, sCOD, and pH measurements as described in [33].

2.5. Statistical Analysis

The experimental results were calibrated and measured by the two different statistical tools. Root mean square error (rMSE) and coefficient of determination (R²) were adopted to indicate the goodness of fit using a prediction model as reported by [34]. Karidio et al., 2023 [35] stated that the value of rMSE should be minimal and R² must be high for better model performance as indicators. The statistical analysis was performed depending on the overall average daily reading affected by the different OLRs using Microsoft Excel at the significance level of p < 0.05.

3. Results and Discussion

3.1. Characterization of Fe₃O₄ Nanoparticles

The formation of Fe₃O₄ Nanoparticles was confirmed for their size and morphology using various analytical techniques. TEM was used to confirm the size, crystallinity, and spherical morphology of the synthesized Fe₃O₄-NPs, as shown in Figure 2a. Most of the particles appeared spherical. Also, some particles were agglomerated due to the magnetic property of the nanoparticles. The size of the Fe₃O₄-NPs was observed in the range of 10–40 nm [36]. A histogram of particles to determine the size distribution of the nanoparticles is presented in Figure 2b. Most of the particles were observed in sizes ranging from 15 to 25 nm.

Zeta potential analysis was carried out for confirmation of the stability and surface charge of the synthesized Fe₃O₄-NPs, as presented in Figure 2c. However, the current study revealed that Fe₃O₄-NPs are more stable with a zeta potential value of −26 mV. The stability of the nanoparticles was measured at the optimal line between stable and unstable particles, as generally categorized as either higher than −26 mV or lower than −26 mV. The previous study reported that more stable nanoparticles were synthesized with the zeta potential value of −26 mV. As a result, the particle is more stable and has a strong negative surface charge [37]. AFM contour imaging in Figure 2b,c was also conducted for the size, geometry, size distribution, and morphology of the synthesized Fe₃O₄ NPs. The profiles of Fe₃O₄-NPs range in height from 10 to 30 nm. The measurement principles of AFM are customized to obtain data that is usually higher than the data of the x, y plane [38]. The acquired results revealed that the particles were spherical in shape and dispersed on the x and y planes. Most of the particles were agglomerated due to the magnetic properties of the nanoparticles.

3.2. Impact of OLR on Methane Generation

Methane contributes to biogas generation by 50–70% of total volume [16]. Furthermore, the OLR decides what volume of organic waste can be efficiently treated in reactors [39]. Moreover, the OLR streamlines the biological conversion capacity and, ultimately, biogas generation [40]. In the literature, various studies have used different OLRs to stabilize the biodigesters to generate maximum energy. Calzada et al. [39] stated that with a low OLR, the biodigesters operate inefficiently, whereas a high OLR is more susceptible to destabilization and ultimately leads to bioreactor failure. For example, the authors of [41] utilized OLRs of 3.3–4.0, 1.0–4.0, and 1.7–15 gVS/Ld, respectively. Whereas ref. [40] adopted the same strategy to increase the OLR gradually from 1.5 to 3.0 and 1.0 to 5.0 gTS/Ld, respectively. Wang et al. [42] concluded that treating FW in a CSTR at 1.0 to 2.5 gVS/L.d of OLR showed stable performance.

In this study, the same process was followed, and we started loading reactors with the lowest OLR of 0.3 and gradually increased it to 1.3 by adjusting 0.8 gVS/Ld as an intermediate concentration of substrate inside the reactors. During this variation in OLRs, the concentration or the dosage of NPs (Fe₃O₄) was kept constant for 120 days. The impact of the addition of NPs on CH₄ is given in Figure 3. Two reactors, R1 (with the supplementation of Fe₃O₄-NPs) and R2 (without NP loading), were run simultaneously to investigate the impact of NPs with different OLRs and the activities that took place in the conventional (R2) reactors.

Figure 3 shows that reactor R1 started with a greater CH₄ yield than R2; the initial yield of CH₄ from both reactors was 51.3 and 15.7 mL/gVS, respectively. The NPs stabilized the AD conditions inside the reactor and started producing more than three times the methane yield produced by the R2 reactor. Furthermore, R1 took under 20 days to reach its ultimate gas production of 725.9 mL/gVS.

Meanwhile, R2 produced a maximum CH₄ of 373.6 mL/gVS on the 39th day at 0.3 gVS/Ld. However, this yield (by R1) was much higher than the results reported by [43,44] for treating FW. These studies concluded the optimal yield of methane rates as 484.9 ±  38.5 mL CH₄/g VS, 371 mL/g of VS_added, 455 mL/gVS, and 396 mL/g VS_added, respectively. However, in this study, a mild destabilization in the reactors was observed when the organic loading was changed from 0.3 to 0.8 gVS/L.d after 40 days. The methane yield in R1 dropped to its lowest rate of 655 mL/gVS after achieving optimal production on the 60th day of the process. This imbalance in the performance of R1 might be caused by the increased toxicity to the methanogens, reduction in enzymatic activity, hampering direct interspecies electron transfer (DIET) between microorganisms, and inhibitory impact [25,26]. Likewise, the CH₄ yield rate became unstable and achieved 300 mL/gVS in the R2, and this fluctuation was detected from the 50th to 70th day of digestion.

With the decrease in methane production, the pH of the reactors dropped to a minimum of 5.60 in reactors R1 and 5.51 in R2. The pH was then adjusted by flushing the dropwise 2.0 M sodium hydroxide (NaOH) in the reactors to bring it up to the required range of 6.5 to 7.5, as recommended by [43,44] as successful and stable conditions for AD.

From Figure 4, it can be perceived that the production rate of methane in R1 was slowly going down and reached 645.4 mL/gVS.d from 725.9 mL/gVS, which was achieved at 0.3 gVS/Ld of OLR. Meanwhile, R2 showed stable CH₄ production. These results very much follow the supposition made by [45]. The optimum CH₄ yield in R2 was 424.4 mL/gVS.d, which was achieved on day 102 of digestion when the OLR was 1.3 gVS/L.d. However, R1 was found to be not so efficient in stabilizing the maximum yield rate, but it produced about 52% higher CH₄ yield than the maximum produced by R2. This suppression of methane yield in R1 could be because of the insufficient balancing of the microbe-to-NP dosing ratio for optimum conditions inside the reactor, which was achieved at 0.3 gVS/L.d. There might be another factor, such as the continuous activities inside the tank (R1) exhausting the NPs and ultimately impacting the chemical properties.

3.3. Impact of pH on CH₄ and TVFAs Generation

Anaerobic-digestion-process stability is directly associated with the total production of volatile acids (TVFAs) and stabilization produced during the hydrolysis step of the AD process [16]. The imbalance of pH inside reactors causes ammonia inhibition [46,47]. At the beginning of digestion, the concentration of TVFAs was maximum, and the pH started to go down, relatively.

From Figure 5a, it can be witnessed that the pH dropped and reached about 5.5 in R2 on the 17th day, while a mild drop in pH was also observed in R1. In R1, the destabilization of pH was not as high as it was in R2. Meanwhile, the concentration of TVFAs was maximum at this stage of the digestion, at 0.3 OLR. The TVFA concentrations in both reactors were recorded as 1822.6 and 2878.4 mg/L in R1 and R2, respectively (Figure 5b).

The pH drop was more rapidly recovered in R1 than in R2. Meanwhile, the CH₄ yield in both reactors reached the maximum generation. At this stage of stabilized generation after the 30th day, the generation of CH₄ in R1 and R2 was in the range of 700 to 725.9 mL/gVS and 361 to 373.3 mL/gVS, respectively. Hence, the pH was at this stage in the range of 6.6 to 7.6 in R1. The pH was also well adjusted in R2, which is required for well-stabilized conditions for any successful digestion, as suggested by [48] and further cited by [43,44]. From Figure 5b, it can be seen that the highest CH₄ (725.9 mL/gVS) in R1 was achieved at 0.3 OLR with maximum bioconversion of TVFAs, which were less than 500 mg/L.

While increasing the OLR from 0.3 to 0.8 gVS/Ld after stabilizing the methane yield and pH (around 7.5) with the maximum conversion of TVFAs, the accumulation of TVFAs was observed again. Meanwhile, in R2, the pH drop caused an increase in the concentration of TVFAs (from 1589.6 to 2122.4 mg/L) inside, and the pH drop subsequently dropped to less than 5.5. A further drop in pH in both reactors was stopped by flushing 2.0 M NaOH solution dropwise to assist with the required range. Thus, pH was again stabilized, and a homogeneous yield of CH₄ was obtained before switching to the 1.3 gVS/Ld OLR after the 80th day. At this OLR, the pH was maintained, and an improvement in CH₄ yield was achieved in the R2, from 373.6 to 424.8 mL/gVS, by increasing the OLR from 0.3 to 1.3, respectively. Moreover, R1 slowly reduced its maximum yield by increasing the OLR.

Table 1. The final digestate analysis of both biodigesters (with standard deviation, SD).

Parameters.	R1 (With F3O4·NPs)	R2 (Control)
TS (%)	25.13 ± 1.21	29.67 ± 1.67
VS (%)	21.91 ± 1.09	24.73 ± 1.25
VS/TS (%)	87.19 ± 4.45	83.35 ± 4.05
sCOD (mg/L)	996.59 ± 20.25	1749.7 ± 29.87
TVFAs (mg/L)	780.56 ± 16.19	990.5 ± 21.08
pH	6.23 ± 0.912	7.45 ± 0.512
Organic carbon (% from VS)	50.6 ± 2.78	48.35 ± 1.89
Total nitrogen (%)	3.06 ± 0.178	2.7 ± 0.65
C:N ratio	16.55:1	17.90:1

gives the final digestate analysis on the 120th day. The TVFA concentration in the reactors was well within the range reported by Nagao et al. [43]. His study suggested a concentration of TVFAs in the range of 500–600 mg/L for a well-stabilized AD process in a CSTR. This study resulted in a final concentration of the bioreactor of 500–800 mg/L, which is well in agreement with the results presented by [43].

3.4. CH₄ vs. Biogas Content

The CH₄ content in the biogas varies due to the substrate type, composition, and time of the digestion, as shown in Figure 6. The impact of NPs on the AD process to yield biogas versus the CH₄ portion was studied. For this purpose, to analyze the composition of biogas, the biogas samples were directly taken from the reactors (R1 and R2). The gas chromatography (GC) model (PACKARD model 437A) was used to analyze the samples.

The samples from each reactor were taken once a week. This analysis was carried out right from the first week to the 15th week of the digestion process. In the first week, GC analysis showed that the CH₄ content was about 56 and 47% in the R1 and R2 reactors, respectively. The content of methane in the R1 at this stage agreed with the results reported by [49]. Their study cited maximum CH₄ content in the whole search, ranging from 57 to 69%. Furthermore, ref. [50] utilized Fe to improve the digestibility of the process. He reported the CH₄ content as 60–65%.

3.5. Behavior of sCOD Removal and TVFAs

The performance of both reactors was further compared and analyzed by the removal rate of soluble chemical oxygen demand (sCOD) and TVFAs. From Figure 7, a higher rate for both was observed in R1, whereas R2 yielded slow degradation rates. A study conducted by [45] reported that mono digestion (only FW) could result in the final concentration of chemical oxygen demand at 571.5 ± 8.0 mg/gTS. This study found that the maximum sCOD removal rate reached 756.68 mg/L in R1, which is quite contrary to the result cited by [45] at 0.3 gVS/Ld of OLR. Additionally, R2 hardly maintained 1749.7 mg/L when the gas production was much more stable in the bioreactor. The influent concentration of sCOD in R1 and R2 was 9876.67 and 9768.32 mg/L, respectively. However, ref. [51] stated the concentration of sCOD was loaded to the digester as 14,500 mg/L.

Furthermore, the removal efficiency of both parameters during the well-stabilized period of digestion was 90 and 86.5% of sCOD in R1 and R2 at 0.3 and 1.3 gVS/Ld of OLR, respectively. Whereas, in a study conducted by [51], which discussed the AD of FW under thermophilic (50 °C) conditions, the maximum removal of sCOD achieved was 83%. This shows that the removal rate of sCOD is better than the results cited by [51].

Later on, the OLR was increased from 0.3 to 0.8 gVS/Ld in both reactors after 40 days of digestion, as shown in Figure 7. Unbalanced behavior was observed in the reactors for the removal of sCOD, and the bioconversion of TVFAs became lower. This impact was caused by the accumulation of TVFAs, which were again raised from 215.45 to 549.98 mg/L in R1, and a similarly higher concentration of TVFAs was analyzed in R2 as well. This accumulation caused a drop in pH from greater than 7 to less than 5 in both reactors. It was noted that when the concentration of TVFAs was increased inside the reactors, it hampered the removal of sCOD. The results at this stage were quite in agreement with the observation made by [51].

In addition to this, ref. [43] cited that at a low OLR, there are fewer chances of inhibition caused by the volatile fatty acids (VFAs). In these conditions, the AD of readily degradable organic compounds (like FW) provides a sufficient balance between the rates of hydrolysis and methanogenesis because methanogenic bacteria are more sensitive than hydrolytic and acidogenic bacteria to high concentrations of VFAs and the corresponding pH drop. Slowly, this condition became reversible in R1 because the feeding rate became higher as the OLR increased, which led to a higher accumulation of TVFAs. The rates of hydrolytic and acidogenic organisms became higher, which agrees with the results of [52].

3.6. TS, VS, TS/VS Disintegration

The disintegration of total solids (TS), VS and their removal ratios were rigorously analyzed during the entire 120 days of operation. Figure 8 shows the VS behavior of each OLR. The VS bio-removal received more attention because the methane yield and the bioreactor’s performance were measured in terms of it and the study of AD of FW at different OLRs. In one study, the removal of total volatile solids (TVS) was reported as 80.9 to 82.7%, when treating the FW using different sizes of the substrate [53].

From Figure 8, the optimum VS reduction was achieved in R1 when the OLR was 0.3 gVS/Ld. Whereas, the R2 yielded maximum removal at 1.3 gVS/Ld of OLR with 75.3% elimination of VS. Then R1 behaved a bit differently. The VS removal increased in R2 and reduced in R1. This effect might be caused by the reduced functionality of the NPs in R1, which failed to maintain their biochemical compatibility with the increasing OLR. In R2, the VS reduction was a bit fast at the OLR of 0.3 gVS/Ld, but at 0.8, the second OLR, the VS reduction became a little shaky as compared to the first OLR. The rise in pH again showed the stabilized behavior of the reactor, and gradually increased VS destruction was noted.

At the third OLR (1.3 gVS/Ld), initially, the removal of VS increased, but after the 100th day, it became almost steady. The results obtained from R2 (without NPs) followed the behavior of VS removal as reported by [43]. Moreover, in R1 (with NP addition) the reduction rate was quite contrary because it reduced with increasing OLR. It was reduced from 93% to around 75% at OLRs of 0.3 to 1.3 gVS/Ld, respectively.

Nevertheless, a slight disturbance was observed in the removal of VS during the middle OLR (0.8 gVS/Ld) due to increased substrate feeding and limited microbial concentration. This impact led to a reduction of more than 90% removal in R1, which was set at 0.3 gVS/Ld OLR and reached around 80% at 0.8 gVS/Ld, as described in Figure 8. Ref. [13] also reported that the OLRs, 0.8 gVS/Ld (which was the lowest OLR) yielded the maximum VS removal of 76% when they found only 63% VS reduction at 2.1 gVS/Ld OLR.

The estimation of hydraulic retention time (HRT) in CSTR-based AD of organic waste plays an important role. This HRT analysis leads to the design of the reactors needed to treat a specific amount of organic waste [40]. From Figure 9, it can be seen that the maximum VS reduction with optimum CH₄ yield was achieved at 0.3 gVS/Ld of OLR in R1. More than 90% of VS removal was achieved in 14 days of hydraulic retention time. Whereas R2 took 20 days to biodegrade the same loading rate as the substrate. Somewhere else in the literature, the mesophilic AD of municipal solid waste digestion gave 15 days of HRT when the reactors were loaded up to 15 gVS/Ld [54].

It was observed that both reactors behaved in opposite ways to each other. In R1, the HRT was increasing with the increase in the OLR along with the reduction in VS removal and reduced CH₄ yield. In contrast to this, in R2, the HRT gradually shortened and reached about 18 days to achieve maximum VS reduction with optimum CH₄ generation. In the literature, ref. [43] took more than 16 days to achieve maximum results. Furthermore, ref. [49], in their study, recommended the most suitable HRT for treating FW as 20 days in continuous AD. Besides these findings, in another research study, the methane yield was the highest in the reactor when an HRT of 12 d was used (223 LCH₄/kg sCOD_degraded) [51]. But in another study, the single-stage CSTRs are favored because they permit simpler designs that suffer less frequent technical failures and have a smaller investment cost [43].

3.7. Effect of Fe₃O_4-NPs on Microbial Communities

The dynamics of microbial communities during anaerobic digestion at different states of acclimation and inhibition are key to diagnosing ailments in this process [55]. The varying OLR in AD causes shakes in the microbial population communities and dynamics during treatment [56]. Thus, the study and the deep analysis of these microorganisms are the core of the digesters as AD is a biochemical process mediating the involvement of a variety of microbial groups [51]. However, most of them focused on the succession of dominant groups or the fluctuations in their relative abundance during the stable stage of the digester [57].

In this study, microbial analysis was carried out at two different stages of anaerobic digestion, such as at weeks 2 and 12 as shown in Figure 10. The analysis results revealed that during the week digestion, the reactors started adopting anaerobic conditions and stabilized. However, in week 12, the results during that the reactors have been stabilized and the impact of NPs on microbial growth is witnessed. Based on the study, shown in Figure 10, the distribution of four significant methane-producing microbes (Proteobacteria, Firmicutes, Bacteroidetes, and Actinobacteria) was totally different in R1 and R2 [58]. These four groups are of major importance in the microbial analysis of any AD process for methane generation. Furthermore, most of the members belonging to the Firmicutes phylum are syntrophic bacteria that can degrade various VFAs [59].

The average relative abundance of these Bacteroidetes, Firmicutes, and Proteobacteria was estimated as 33.85%, 31.36%, and 34.79% of bacteria, respectively [60]. Whereas, in another study, the microbial community in the effluent from digesters was dominated by representatives from the phyla Proteobacteria (77.8%) and Firmicutes (19.5%). In addition, the phyla Actinobacteria and Bacteroidetes were present at lower abundance, 1.5 and 0.6%, respectively [43,44]. Figure 10 shows the number of reads of these microbes in our study. It can be seen in Figure 11 that the three microbial communities (i.e., Bacteroidetes, Firmicutes, and Proteobacteria) significantly increased in R1, to which NPs were added, compared to R2. Their numbers of reads were about 10,000, 45,000, and 41,000, respectively. That is the way R1 yielded better performance in terms of CH₄ generation and VS reduction rates. Perhaps, this stabilized performance of R1 might be because of a proper DIET mechanism among the microbes, enhanced enzymatic activation, good resistance against the inhibitory compounds (such as sCOD and TVFAs) and boosted hydrolysis and metabolism [25,42]. Hence, it was observed that the addition of NPs showed biocompatibility during the 120 days of digestion, but their efficiency decreased as OLR increased.

3.8. Kinetic Modeling

Equation (1), as utilized by [43], was used to observe the kinetic behavior of CSTR. Equation (2), on the other hand, was used to estimate the rate of biodegradation (k). Mähnert and Linke [61] studied the kinetic behavior of the CSTR based on the mass balance equation of the substrate. A steady state was considered, and it was assumed that the feeding of the substrate was equal to the sum of removal and substrate disintegration.

$$\left.{\mathrm{M}}_{\mathrm{t}}={\mathrm{M}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\ast (1-{\mathrm{e}\mathrm{x}\mathrm{p}}^{-\mathrm{K}\mathrm{c}\ast \mathrm{t}}\right)$$

(1)

$$\mathrm{l}\mathrm{n}\left[{\displaystyle \frac{{\mathrm{V}\mathrm{S}}_{\mathrm{t}}}{{\mathrm{V}\mathrm{S}}_0}}\right]=-\mathrm{k}\mathrm{t}$$

(2)

where M_t represents the methane yield (in mL/gVS at STP) at time t (in days), M_max is the maximum methane yield (mL at STP), and k_C is the first-order conversion-rate constant (per day).

Table 2. Kinetic and statistical parameters of both reactors.

Bioreactor	Exp (Mt) (mL/gVS)	Model (Mt) (mL/gVS)	Kc (d−1)	R2	rMSE (%)
R1 (with Fe3O4 NPs)	725.901	724.611	0.057	0.833	0.317
R2 (control)	425.721	419.352	0.021	0.872	0.238

enlists the kinetic parameters and statistical checks.

The value of k_c was estimated using Equation (2) and plugged into Equation (1) to estimate the model-based methane yield (M_t). It can be seen from Table 2 that the model yielded the optimum methane yield in R1 as 724.611 mL/gVS. The observed (experimental) CH₄ yield was 725.901 mL/gVS with a higher rate of k_c 0.057 per day. Further model prediction suggested that, in R2, the model predicted the experimental data quite well, with the coefficient of determination (R²) of 0.872. The rate of biodegradation (k_c) was found to be 0.021 (d⁻¹), which is less than the rate observed in R1. In the literature, the k_c was also reported as 0.089 per day when the CSTR was run for about 250 days. Furthermore, Lin et al. [62] reported the k value as 0.07 (d⁻¹) during the 140 days in the CSTR. In contrast to these results, ref. [61] reported 0.032–0.0316 (d⁻¹).

Based on the above discussion, the first-order model predicted experimental data with a high goodness of fit of more than 80%, as reported by [63]. The kinetic parameters suggest that the addition of NPs enhanced the biodegradation rate, allowing us to achieve more yield with the same OLRs, which were used to compare in terms of R1 and R2. Adebayo et al. [40] suggested that the kinetic parameters such as M_t of M_max and k are the deciding factors in bioreactor design.

Ref. [61] reported R² in the range of 8 to 65%. However, our study showed a much better R² analysis, which was observed in more than 80% of the cases. Ref. [40] reported the R² at the mesophilic AD conditions as 0.860. Moreover, rMSE analysis showed very little deviation when the results were simulated with the model results. From Table 2, it can be seen that the rMSE in terms of both reactors was in the range of 0.238 to 0.317%. This deviation was way better than the rMSE results reported by [34]. In his study, the rMSE with a value of 0.59 was considered great and 2.01 was referred to as a very poor prediction during the univariate analysis.

4. Conclusions

The study evaluated the efficiency of Fe₃O₄-NPs as a supplement for anaerobic digestion (AD) in two concurrent CSTRs. The OLR ranged from 0.3 to 1.3, and the concentration of NPs demonstrated efficient biocompatibility. The R1 dramatically enhanced sCOD removal, as well as TVFA conversion and creation rates. The maximal TVFAs, sCOD, and VS elimination rates were 215.45 mg/L, 756.68 mg/L, and 92.2%, respectively, at 0.3 gVS/Ld of OLR. In R2, CH₄ generation rose from 372.8 to 424.4 mL/gVS at 0.3 to 1.3 gVS/Ld of OLR, respectively. The reduction in VS, sCOD, and TVFAs was further improved by enhancing the OLR in R2. In R2, the VS reduction was consistent at 1.3 gVS/Ld of OLR, reaching around 80% at its peak. However, pH variation was shown to be much more prominent in R2 than in R1, indicating the existence of NPs greatly reduces pH dips. So, in R1, the maximum removal of sCOD, TVFAs, and VS was achieved at 90%, 74%, and 93%, respectively, as compared to R2. The results support the use of green and biodegradable NPs for large-scale AD treatment.