Water lettuce (Pistia stratiotes L.) increases biogas effluent pollutant removal efficacy and proves a positive substrate for renewable energy production

Experiment 1: water lettuce phytoremediation performance

The WL was collected from the adjacent pond to the experiment site. The plants chosen varied from 4–6 cm in length, and each plant had from five to seven leaves at the initiation. We removed all the bad parts from the plants and rinsed them with tap water to eliminate any debris of other macrophytes. Then, the WL was placed into a plastic barrel filled with husbandry anaerobic digester effluent (HADE) for 1 week, allowing it to adapt to its environmental conditions.

The experiment was designed in a batch system using styrofoam boxes (60 cm long × 45 cm wide × 41 cm high). Each box contained 30 L of HADE, which was not added to during the experiment. The WL was placed into boxes for different treatments by the plants arranged in 50%, 25%, and 12.5% of the box’s surface area. In parallel, a control treatment (0%) was set up without WL. Each treatment was replicated in five boxes. Accordingly, a total of 20 boxes were prepared. All the experimental boxes were randomly arranged on a scaffold under natural conditions. A plastic film was set up to cover the experiment boxes if it rained. The experiment ended in 30 days. The pollution elimination percentage was calculated as shown in Eq. (1):

(1) $$\mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}(\mathrm{\%})={\displaystyle \frac{{\mathrm{C}}_{\mathrm{i}}-{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{C}}_{\mathrm{i}}}\times 100}$$in which C_i is the initial concentration of pollutants and C_e is the final concentration of pollutants remaining in the effluent.

The HADE was collected from the existing PE biogas plant at the experiment site and installed in the boxes without any dilution. The HADE was examined before the initial experiment and after the phytoremediation tests at 6, 12, 18, 24, and 30 days. The water samples were collected as mixed samples from each treatment to achieve a pooled sample. The physicochemical parameters of pH, TSS, BOD₅, COD, total Kjeldahl nitrogen (TKN), and total coliforms were chosen as required by the QCVN 62-MT:2016/BTNMT National Technical Regulation on Livestock Effluent. The collected HADE samples were analyzed according to the Standard Methods for the Examination of Water and Wastewater (APHA, 1998).

The growing ability of the WL was recorded before and after the phytoremediation process every 5 days for the 30 days of the experiment. The plants’ fresh weight was determined using a digital balance to weigh all the plants within each box. The root length and leaf length were measured by randomly choosing three plants from each box and recording the average value of the three maximum values of the root and leaf (Fig. 2).

Experiment 2: anaerobic digestion performance

WL from the nearby pond was collected and fed into the existing PE biogas plant as Supplemental Material. After each collection time, the WL was dried within a day in shady conditions to avoid nutrient loss, and it was then fed into the biogas digester. The quantity of WL co-fed every 2 days into the HDPE biogas plant was 10 kg wet weight within the first 10 days, 40 kg wet weight in the next 30 days, and 24 kg wet weight in the last 20 days. At the start, the volatile solids (VS) content of the WL was determined by drying a sample of the biomass at 105 °C for 24 h and then heating it at 550 °C for 2 h. The VS value was then calculated based on the weight of the dried sample.

The daily produced biogas was recorded onsite by a VIGADO G1.6 gas meter (PDM 095-2008; Vietnam). For the gas components, biogas samples were collected in an aluminum bag after feeding the WL every 4 days for 60 days. The gas samples were transported to the biogas laboratory at Can Tho University and analyzed using a portable biogas analyzer (Biogas Pro 5000; Landtec, Dexter, MI, USA) for the biogas components, including CH₄, CO₂, and other compositions (Fig. 3).

Pollution removal

Figure 6A shows the phytoremediation performance of the HADE using WL every 6 days for 30 days. The pH levels slightly decreased during the phytoremediation process for all treatments. Although the treatment with 12.5% surface cover showed the largest change in pH values between days 12 and 18, the difference in pH across all treatments was negligible.

The concentration of contaminants (TSS, BOD₅, COD, TKN, and TP) rapidly reduced in the first 6 days of running the experiment (Figs. 6B–6F). There was a slight decrease for the rest of the period. The higher coverages of WL displayed better treatment efficacy of HADE. All the WL treatments demonstrated a higher potential for pollutant elimination than the control treatment (0%) (Fig. 6).

Figure 7 shows the pollutant treatment efficiency of WL in the effluent from the digester. In the first 6 days, there was a sharp decrease in the treatment efficiencies of the WL, but then a slight decrease between days 6 and 30 was shown. The TSS removal efficiency in the first 6 days of TSS was highest at 80.47–94.53% (Fig. 7A), followed by TP (68.36–82.86%, Fig. 7E), BOD₅ (44.19–59.30%, Fig. 7B), COD (44.3–59.73%, Fig. 7C), and TKN (38.25–46.75%, Fig. 7D). After 30 days, the reduction efficiency of TSS and TP did not continuously increase much compared to the first 6 days, while the treatment efficiencies of BOD₅, COD, and TKN rose much higher than during the first 6 days (Fig. 7).

Recent studies have confirmed high treatment efficiency on BOD₅ (83.3–92.8%) and COD (79.2–85.9%) when the phytotreatment of polluted river water, palm oil mill wastewater, and sewage wastewater using macrophytes was applied (Shahid et al., 2019; Wei, 2019; Schwantes et al., 2019). However, other research has reported lower removal efficiencies of phytotreatment. Jyotsna, Bhasin & Punit (2015) confirmed removal efficiencies of 44.8% of TSS, 62.4% of BOD₅, and 63.8% of COD for pulp and paper mill effluent after 30 days by lesser duckweed. Tang et al. (2009) recorded reduction efficiencies of 33.2% of COD and 21.8% of TKN for wetland constructed to treat eutrophic river water. However, the organic concentrations in the pulp and paper mill effluent and eutrophic river water were much higher than in this study. Treatment effectiveness can vary depending on various factors, such as the initial pollutant levels, the duration of treatment, and the specific conditions of the treatment environment.

In our study, the removal efficiency of COD was slightly lower than that of BOD_5, which is similar to the study by Hamzah, Yusof & Alimon (2016) of phytoremediation testing on palm oil mill final discharge wastewater (84.7% of BOD₅ and 22.3% of COD). In fact, recalcitrant compounds, such as high humic acid in pig dung (PD), are known to cause low treatment efficiency of COD of the phytoremediation (Norulaini, Ahmad Zuhairi & Muhamad Omar, 2001; Zhang et al., 2012).

High removal efficiencies of over 89% of TP were recorded for all the WL treatment coverages after 30 days. The utility of phosphorus in aquatic plants is essential in various physicochemical and biological processes. In addition, values from 61.8% to 65.0% were the reduction efficiencies of TKN after 30 days of phytoremediation. In a similar approach using swine wastewater, Sudiarto, Renggaman & Choi (2019) documented that TKN and TP removals within 21 days were 71.4% and 32.2%, respectively. Similarly, Parwin & Paul (2019) reported the highest removal efficiencies of 94.4% of TKN and 98.1% of TP using water hyacinth to treat kitchen wastewater within 4 weeks. In contrast, the lowest reduction efficiencies of BOD₅, COD, and TKN using water caltrop to treat municipal wastewater have been recorded as 18.3%, 14.2%, and 27.2%, respectively (Kumar & Chopra, 2018).

Water lettuce biomass

Previous findings disclosed that the biomass of WL increased up to the first 10 days and then diminished as the basal leaves decayed (Fonkou et al., 2002) or due to their maximum uptake limit of nutrients (Hamzah, Yusof & Alimon, 2016). Our study showed that the biomass gradually increased through the experiment period (Fig. 8). The difference compared to the previous report was due to the required space for the growth of WL. Based on the produced WL, the doubling time for the wet-weight biomass from the treatments of 50%, 25%, and 12.5% surface coverage were estimated at 12.32, 12.11, and 12.28 days, respectively (Table 2). The rapid increase in WL’s biomass indicates its potential for eliminating pollutants and providing valuable material for biogas production.

Figure 9A shows that the root lengths of the WL increased over time. Root lengths were recorded from 6.66 to 8.08 cm (an increase of 21.3%), 6.52 to 8.06 cm (an increase of 23.6%), and 6.16 to 8.04 cm (an increase of 30.5%) after 30 cultivated days for the treatments of 50%, 25%, and 12.5% surface coverage, respectively. The maximum leaf length of growth recorded in the treatment of a surface cover of 12.5% WL show that with fewer bodies (Fig. 9B). The nutrients were absorbed and formed up to the root system instead of the leaves. During the first 15 days, the length of the WL’s leaves increased in all treatments but then decreased as a result of the decay of the lower leaves and the emergence of new leaves. Consequently, a high amount of biomass was recorded as the WL continuously grew in the experimental containers.

Figure 10 depicts the Pearson correlation of the pollutant parameters in the HADE among the treatments, as indicated by the color and size of the circles. The magnitude of the correlation is reflected by these visual cues. A strong positive relationship was found between pollutant concentrations, with r values ranging from 0.8 to 1.0. The BOD₅ and COD displayed the highest correlation. The Pearson correlation revealed adequate interdependence between the pollutant composition and the prospect of WL to effectively eliminate pollutant levels in the HADE. Similarly, leaf length and root length showed a moderate correlation (r = 0.53), while root length and fresh biomass exhibited a high correlation (r = 0.85) (Fig. 11). No relationship was found between leaf length and fresh biomass. As such, the root length played a crucial role in increasing the WL’s biomass. Macrophyte roots are vital for reducing contaminants in the aboveground part of their bodies. Moreover, macrophytes also secrete biopolymer from their roots, which assist in flocculation (Sharma, Singh & Manchanda, 2015). Non-settling and colloidal particles are also removed, at least partially, by bacterial growth, which results in the removal of some colloidal solids and the microbial decay of other organic pollutants.

Biogas composition

Figure 13 shows the change in biogas composition over time. The biogas composition recorded on the first day was 55.3% CH₄ and 30.5% CO₂. During the experiment, the CH₄ content varied from 54.1% to 59.9% and the CO₂ content varied from 21.9% to 31.7%. Nam et al. (2015) reported 47.7–56.6% and 46.5–54.4% methane content for onsite experiments of co-digestion of rice straw and water hyacinth, respectively, with PD using PE biogas plants in the VMD. Another study on co-digestion of cow dung and WL found CH₄ content from 50.6% to 54.8% after 2 weeks of fermentation (Phuong et al., 2015), while Cong et al. (2022) discovered that solely digesting WL produced the highest concentration of methane by 62.2% on day 35. This study noted a high methane content from the start of the experiment, as the HDPE plant had been in operation for 2 years and was being fed with PD. The biogas could be applied to energy consumption due to its high methane content, particularly for household cooking in the VMD.

Feasibility of farm-scale biogas production and AD effluent removal systems

Biogas production can be considered a renewable energy source in most localities. Unlike other renewable energy sources, such as solar and wind power, biogas energy sources are used directly by households. Accordingly, biogas plants are more appropriate for decentralized energy production areas in the VMD. It is notable that bio-gasification from AD depends on several factors, such as carbon/nitrogen (C/N), hemicellulose content, pH, and the buffering capacity of the substrate (Güngören Madenoğlu et al., 2019). In farm-scale experimentation, WL showed great biogas production potential for co-digestion with PD. CH₄ concentration was achieved by more than 50% after 2 weeks, implying conformity for household cooking and heating (Ngan et al., 2020).

Under these circumstances, WL source availability is the main challenge in expanding the system for households. In the VMD, many household farms have sufficient land area to install biogas, as well as a pond for growing WL combined with aquacultural activities. WL can be grown on a pond with a nutrient source provided by the livestock biogas plant. The contaminant-eliminating productivity of WL was reliable throughout the current study. Nutrient removal by WL involves not only eutrophication reduction of surrounding water bodies, but also significant biomass production for renewable energy generation. WL biomass can be used to feed the biogas plant directly or applied as simple bio-pretreatment technologies (soaking in biogas effluent or anoxic mud for a 5-day period) to enhance the biogas yield, as suggested by Nam et al. (2021) and Nam, Van Cong & Van Thao (2023). The co-digestion of WL and PD could solve common issues concerning insufficient animal waste sources for biogas production, because many farms in various localities typically have a smaller standing stock of animals (permanently or temporarily) (Nam et al., 2017, 2021). Our study encourages the co-digestion of WL with PD to enhance renewable energy production. Notably, the use of renewable energy produced from AD allows for energy independence in the context of high energy consumption demand. Moreover, the application of household biogas plants to produce low-cost energy increases household economic efficacy owing to saving electrical use, liquefied petroleum gas (LPG) and firewood, while reducing greenhouse gas emissions from agricultural activities (Jan, Truc & Nam, 2018).