Waste-derived substrates in vertical-flow constructed wetlands for an efficient removal of high-concentration heavy metals

Contaminations caused by heavy metals (HMs) in aquatic environments have become a grave ecological concern that poses significant threats to the environment and human health (Cheng et al. 2002; Mustapha et al. 2018). HMs can pose significant risks to living organisms due to their bio-accumulating nature, affecting organisms through direct consumption and biomagnification in trophic systems (Yin et al. 2016; Salama et al. 2019; Ali et al. 2020; Zhao et al. 2020; Chen et al. 2021; Yu et al. 2021).

Numerous techniques have been developed for the treatment of wastewater containing HMs, including membrane separation (Xiang et al. 2022), chemical precipitation (Kurniawan et al. 2006), and electrolysis (Wu et al. 2023), which offers highly efficient removal of HMs (Lu et al. 2021). However, operation and maintenance of these technologies are costly (Zhang et al. 2019; Ang et al. 2023) and raise doubt on their economic viability for impoverished and developing countries (Mustafa 2013; Liu et al. 2020).

Constructed wetlands (CWs) are viable, nature-based, and cost-effective wastewater treatment options that have been used for more than 50 years (Brix 1994; Vymazal 2014; Stefanakis 2018) and demonstrated their effectiveness in removing HMs from wastewater (Song et al. 2011). Additionally, CWs can deal with different environmental factors and loading contaminants (Nair 2008; Yadav et al. 2012). A CW system mainly comprises substrate, vegetation, and microorganisms, and their cooperative interplay between physical, chemical, and biological processes efficiently reduces a wide range of organic and inorganic pollutants from wastewater (Yu et al. 2021). The removal mechanism of HMs in CWs is intricate and involves processes such as substrate adsorption, precipitation, co-precipitation, formation of insoluble salts through complexation (e.g., metal sulfides, carbonates), plant accumulation, and microbial metabolism (Lesage et al. 2007; Song et al. 2011; Vymazal & Březinová 2016; Zhao et al. 2024), and proven effective for the treatment of HMs from domestic wastewater (Sun et al. 1998; Singh & Srivastava 2016), industrial effluents (Maine et al. 2006; Khan et al. 2009; Sukumaran 2013), landfill leachate (Madera-Parra et al. 2015; Bakhshoodeh et al. 2016; Saeed et al. 2021a), and acid mine drainage (Sheoran & Sheoran 2006; Chen et al. 2021; Nguyen et al. 2022). However, most of those studies incorporated low concentrations of HMs, necessitating a study to determine CW's effectiveness in high-concentration HM treatment. According to studies, vertical-flow constructed wetlands (VFCWs) are suitable for a variety of wastewater treatment, including HMs (Vymazal 2014; Dan et al. 2017; Liu et al. 2020; Ghezali et al. 2022), and require less land than other CWs, making them more practical in developing countries and urban settings.

Substrate plays a significant role in the elimination of HMs in CWs. It determines the effectiveness of CWs by filtering and adsorbing contaminants, promoting vegetation growth, and providing habitats for microorganisms (Wu et al. 2015; Bavandpour et al. 2018; Zhang et al. 2020; Chen et al. 2021; Nguyen et al. 2022). Gravel and sand are conventional and most utilized substrate media in CWs (Sun et al. 2012; Herrera-Melián et al. 2014; Bavandpour et al. 2018). Alternative substrates, such as WDMs have exhibited improved HM removal in CWs due to their unique physicochemical properties, promoting more efficient chemical and microbial removal mechanisms (Batty & Younger 2004; Bavandpour et al. 2018; Abedi & Mojiri 2019; Chen et al. 2021, 2024; Lizama-Allende et al. 2021; Saeed et al. 2021a; Li et al. 2023). Moreover, their easy accessibility and affordability make them even more appropriate for utilization as CW substrates, thereby contributing to effective solid waste management.

This study used waste-derived materials, such as brick clinker (locally known as Jhama), eggshell, and date palm fiber (DPF) as filter medium. Among them is over-burnt brick (Jhama), which is not useful for construction and is considered waste material. However, it has a coarse surface texture and is porous compared to standard bricks. Owing to their surface properties, Saeed et al. (2021b) have employed them as a substrate in CW and reported their effectiveness in reducing chemical and biochemical contaminants in municipal wastewater. The second organic waste used in the study is eggshell, which has the potential to adsorb HMs due to its surface properties and chemical composition (Daraei et al. 2014; Özcan et al. 2018; Makuchowska-Fryc 2019). However, to the best of our knowledge, it has not yet been applied to CW systems to remove HMs. Likewise, eggshells and DPF have not yet been used as a component of the filter medium in CWs, but it has been utilized as a low-cost biosorbent in the elimination of wastewater pollutants, including dyes, organic compounds, and HMs (Ahmad et al. 2012; Amin et al. 2017; Al Arni et al. 2023). Therefore, we hypothesize that these substrates might reduce HMs in CW systems and be used as a filter medium with the combination of appropriate wetland macrophytes. Selection of suitable plant species is also crucial for enhancing the HM removal efficiency of CWs, as they possess photo-stabilization, remediation, and uptake capabilities (Liu et al. 2007; Yadav et al. 2012; Mustapha et al. 2018; Rahman et al. 2022). In this study, two macrophytes, Canna indica and Hymenocallis littoralis (commonly known as spider lily), were employed as wetland plants due to their fast growth and resilience.

The main objectives of this study are (a) HM removal capacity of those waste-derived (eggshell, DPF, and Jhama) substrate medium in VFCWs; (b) efficiency of the VFCWs in reducing HMs from highly contaminated influent loadings; (c) HM accumulation capacity of C. indica and H. littoralis and compare their effectiveness as CW plant species. The study's findings will help evaluate the suitability of CWs for treating highly concentrated HMs in wastewater using easily accessible WDMs.

CW1, CW2, and CW3 were filled with gravel as control systems for comparing the waste-derived filter media with a gravel-based medium. At the same time, CW4, CW5, and CW6 were packed with Jhama brick, eggshell, and DPF as experimental systems, respectively. The Jhama brick and gravel sizes were 5–20 mm and 10–30 mm, respectively. Additionally, a 5 cm layer of large stones (20–40 mm) was placed at the top and bottom of each vertical column to facilitate a smooth flow of influents and effluents. Two plant species, H. littoralis (commonly known as white spider lily) and C. indica (commonly known as canna lily or Indian shot), were utilized in this investigation. We introduced five plant specimens into each planted system. C. indica was placed in the Jhama (CW4), eggshell (CW5), and one gravel-filled (CW2) system, respectively. Conversely, H. littoralis was planted in the DPF (CW6) and one gravel-filled system (CW3). One gravel-filled chamber (CW1) remained unplanted. This CW setup was adopted to evaluate the effectiveness of both WDMs and plant species simultaneously. As CW4, CW5, and CW6 are filled with different WDMs and introduced different plant species, we implemented two gravel-based control systems (CW2, CW3) that included both plant species but no WDMs, which allowed us to assess the effectiveness of both WDMs and plants simultaneously. Additionally, the gravel-filled unplanted (CW1) system was considered both substrate and plant control at the same time. However, this study also examined the accumulation of HMs in different WDMs, control media (gravel), and plants before and after the study to get empirical validation.

Synthetically prepared wastewater containing eight HMs (As, Cr, Cd, Pb, Fe, Zn, Cu, Ni) was manually introduced into the systems from a storage tank in a cycle sequencing batch mode. Synthetic wastewater formula and chemical ingredients are given in Supplementary Information (Tables S1 and S2) (Saeed & Sun 2011). Each CW system receives 5 L of synthetic wastewater as influents in every cycle. Following a retention time of 3 days, samples were collected from effluent discharge points of CWs. The total experimental duration was 90 days. During the first 30 days of operation, all CW systems were exposed to a HM feed to establish stability without effluent sampling. Afterward, experimental analyses of influent and effluent samples were started for 60 days. Water samples from the six CWs were tested for pH, DO, temperature, and HMs during that period. Sampling and analysis were conducted four times, on the 15th, 30th, 45th, and 60th days. The pH, DO, and temperature measurements were conducted on-site using an HQ 40d multi-meter (HACH, Germany). For HM analyses, water samples were digested with 0.5 mL concentrated HNO₃ per 10 mL of the sample at 105 °C on a hotplate for 2 h (APHA 2012). Plant and substrate samples were collected before and after the 3-month operation period. To digest the substrate, concentrated HNO₃ and H₂SO₄ were added to a 500 mg sample at a ratio of 1:2. The mixture was then heated to 150 °C for 6 h or until the solution became clear (APHA 2012). After harvesting the plants from different CWs, the roots, stems, and leaves were separated and dried at 75–80 °C for 48 h. After grinding, a 300 g sample was subjected to digestion using a 3:1 mixture of HCl and HNO₃ for 2 h on a hotplate at 105 °C temperature (Nguyen et al. 2022). The HM content was determined using the Inductive Coupled Plasma optical emission spectrometer (PRODIGY 7, ICP-OES, USA). SEM-EDS analysis of substrate materials was conducted using a scanning electron microscope (SEM) (ZEISS Gemini SEM 500, UK) to determine the surface characteristics and elemental composition.

An analytical method was employed to measure precision and bias by employing reagent blanks and sample replication. The investigation revealed that both accuracy and bias were less than 10%. Before use, sampling bottles and laboratory glassware were soaked in a 2% solution of HNO₃ for 24 h, followed by cleansing and rinsing with tap and distilled water and eventually oven-drying. Careful selection of standards and instrument optimization were conducted to maintain accuracy in the calibration curves of ICP-OES.

The mass balance of HMs in each CW system was demonstrated by considering the total input of HMs (mg) into the system through influents over the duration of the study, accumulation of HMs in various components of the CW system, such as substrates (Equation (4)), plants (Equation (5)), and the quantity of HMs eliminated through effluents. The overall calculation of mass balance is incorporated in Supplementary Information (Table S4).

One-way ANOVA was employed to ascertain the statistically significant disparity in treatment performances across six CWs. A paired t-test was conducted to find the statistically significant variation between influent and effluent metal concentration. All the descriptive analyses and visualizations were performed using GraphPad Prism 8 and MS Excel 16. The Pearson correlation coefficient was performed in R Studio to comprehend the relationship between environmental parameters and HM removal. The configuration figure of CWs was created utilizing Adobe Illustrator 21.

Parameters such as pH, dissolved oxygen (DO), and temperature in the influent and effluent have been monitored to determine the interaction of water's physiochemical parameters in CW systems. The result shows the pH values of the influent were observed between 7.82 and 8.48, with a mean value of 8.0 ± 0.3 (Supplementary Information, Figure S2). A significant difference in pH is observed between the influents and effluents (p < 0.01). The pH of the effluents from the planted systems ranged between 6.86 and 7.70, while the unplanted CW1 exhibits a pH range of 7.46–8.23. The DO of the influents ranged from 3.52 to 8.16 mg/L. DO values of the effluents showed a continuous decrease during the experimental period, where systems CW1, CW2, CW3, CW4, CW5, and CW6 ranged from 2.39 to 6.79, 2.18 to 8.08, 2.44 to 7.2, 1.79 to 6.42, 0.28 to 3.99, and 1.34 to 6.87 mg/L, respectively. The water temperature during the experimental period ranged between 30 and 34 °C, and no significant difference was observed between the influents and effluents (p > 0.05). The ambient temperature recorded during the experimental period was very high, ranging from 35 to 40 °C, contributing to a high evaporation rate, particularly in the planted systems. The recorded evaporation from systems CW1, CW2, CW3, CW4, CW5, and CW6 ranged from 125 to 400; 340 to 1,100; 320 to 500; 240 to 1,000; 588 to 1,700; and 100 to 430 mL/day, respectively.

The mean concentration of HMs in influent and effluent and the corresponding removal efficiencies are displayed in Table 1,. Among the gravel-filled control CWs, unplanted CW1 exhibited a removal efficiency ranging from 80.5 to 84.3%, while planted CW2 and CW3 demonstrated removal efficiencies of 76.1–85% and 79–88.6%, respectively, for all analyzed HMs. In the control group, unplanted CW1 showed slightly higher removal rates than planted CW2, possibly due to lower evapotranspiration, which concentrated HMs in planted systems (CW2). CW2 experienced an average evapotranspiration rate of 740 ± 393 mL/day due to the presence of plants, whereas unplanted CW1 showed 271 ± 109 mL/day average transpiration (Table 1). Conversely, among the experimental CWs, Jhama, eggshell, and DPF-filled CW4, CW5, and CW6 showed efficiency rates between 82.7 and 93.4%, 91.7 and 97.4%, and 92.6 and 98.7% respectively. A statistically significant variation in HM removal efficiency was observed among different CW systems (ANOVA, p < 0.05). According to Supplementary Information (Figure S1), a decreasing trend in HM reduction has been witnessed with the increase of influent HM concentration, particularly in the gravel-filled CWs. However, it is noteworthy that high level of HM concentration had minimal impact on CW4, CW5, and CW6 (Figure 2).

Additionally, the HM removal efficiency in eggshell-filled CW6 has increased over time despite an increase in influent concentration (Figure 3). A consistently noteworthy As removal rate (95.5–98.2%) is observed in the DPF-filled CW6. Furthermore, the CW5 and CW6 exhibit a removal rate of Cr (91.1–100%) and (95.9–99.9%), respectively, as well as a removal rate of Pb (92–100%) and (93.4–99.6%), respectively. Among the eight studied HMs, Fe exhibits the lowest removal rate (68.6–96.4%) across all CWs, including CW5 and CW6. The CW6 demonstrates the highest efficiency in removing nickel (98–99.3%), while the CW5 achieves a removal rate of (92.8–98.3%) for Ni. Conversely, inorganic media-based CWs, such as those filled with gravel and Jhama, are less effective in removing nickel (70.3–93.5%). A similar trend is observed in the removal of copper, with CW5 and CW6 exhibiting a removal rate of (94.7–99.1%), CW4 exhibiting a rate of (86.2–93.4%), and gravel-filled CWs exhibiting a rate of (73.3–94.4%). Overall, CWs filled with organic substrates (such as eggshells and DPF) demonstrate greater efficiency in removing metals than those filled with inorganic substrates. Organic substrates frequently contribute to the removal of metals through the processes of adsorption and absorption, while organic compounds such as organic matter, lignin, and chitin interact with metals and facilitate the process of sorption (Yadav et al. 2012; Batool & Saleh 2020; Saeed et al. 2021a). Furthermore, Jhama brick was more effective among inorganic substrates than gravel-filled CWs. The superior elimination of HMs in CW4 filled with Jhama can be elucidated by its porous surface structure compared to gravels. The increased specific surface area and augmented number of adsorption sites in Jhama presumably resulted in a heightened removal of HMs compared to gravel (Sekomo et al. 2012; Zhang et al. 2020).

The overall growth of C. indica and H. littoralis is recorded higher in DPF and eggshell-based CWs than the inorganic substrate-filled ones. The concentration of HMs in various components of C. indica and H. littoralis before and after the study can be observed in Figure 7. These two plants exhibited high levels of metal accumulation in different parts, such as the roots, stems, and leaves. The concentration of metals in the roots was significantly higher than in the aboveground parts, as indicated in Table 2. Specifically, the roots of Canna and Hymenocallis contained concentrations of all metals ranging from 938.4 to 6,095 mg/kg and 328 to 2,920 mg/kg, respectively. On the other hand, the shoots of Canna and Hymenocallis contained concentrations of metals ranging from 11.7 to 1,761.6 mg/kg and 5.2 to 1,108 mg/kg, respectively. Although the concentration of Fe, Zn, and Cu was found to be relatively high, their actual accumulation in the plants was relatively low compared to other metals due to their higher initial concentrations (Figure 7). According to Table 2, the BCF values for all metals are ranged from 69.4 to 158 for Canna and 24 to 70.3 for Hymenocallis. Additionally, the TF values for Canna and Hymenocallis ranged between 0.01 to 0.29 and 0.02 to 0.38, respectively, for all metals. The TF values for all studied metals in both plants were less than 1, which could be attributed to the high concentration of HMs in the influents (Deng et al. 2004). It is plausible that the presence of elevated levels of concentrated metals in the roots has impaired the capacity for metal translocation (Deng et al. 2004; Saeed et al. 2021a). (Ghezali et al. 2022) found high BCF values (>1) and reduced TF values (<1) in C. indica, even when subjected to relatively low levels (0.03–7.56 mg/L) of metals stress. The sequestration of metals within the root vacuoles, where they are rendered non-toxic, could explain the considerable accumulation of metals in the roots and limited translocation to other plant tissues (Shanker et al. 2005; Ang et al. 2023).

Both plants exhibited rapid growth in the challenging environmental conditions. Following the 3 months operational period, C. indica displayed a biomass increase of 142, 85, and 226 g of dry weight in CW2, CW4, and CW5, respectively. Similarly, the H. littoralis plants experienced an increase of 83 and 95 g in CW3 and CW6, respectively. The growth ratio, tolerance mechanisms, and significant accumulation of HMs in the roots, along with the high BCF values, indicate that both C. indica and H. littoralis possess proficient Phyto stabilization capabilities (Hernández-Pérez et al. 2021; Ghezali et al. 2022). Therefore, these plants are suitable for implementation in CWs for the treatment of HMs, particularly in high-concentration scenarios.

According to the study by (Zhang et al. 2020), the uptake of metals by plants only represented 1.6–6.5% of the total amounts present in the influent HMs. Conversely, plant contribution might exhibit a significantly higher influence in CWs operated for long periods with low-concentration HM loadings. After 7 months of conducting experiments, (Nguyen et al. 2022) found that the planted CWs exhibited a notable biomass accumulation in the plant, with a range of 8.8–44.3% of total HM distribution.

Except for direct accumulation in tissues, plants influence metal removal in CWs through the release of organic molecules from roots, as well as by offering additional surface areas to support the growth of microorganisms (Sultana et al. 2014; Saeed et al. 2021a). Additionally, the functions performed by plants in efficiently transferring DO and organic substances play a significant role in the prevailing aerobic mechanism for the removal of HMs (Oustriere et al. 2017; Sinha et al. 2017; Yang et al. 2017; Yu et al. 2022).

In eggshell-filled CW5, the release of HCO₃₋ and might be the primary removal pathways for Cr, Pb, Fe, Cu, and Ni by forming insoluble metal carbonates (Figure 8). Additionally, Ca²⁺ released from the surfaces can also react with metal ions to form a suspension and therefore make it immobile (Nguyen et al. 2022).

It is important to note that calculating mass balance demonstrates insight into the HM distribution pathways within CWs. However, a detailed and precise representation of how metals behave in different parts of a CW system is still limited due to their complex relationships. Furthermore, the results of this study were obtained over 3 months, so it is necessary to conduct long-term studies to fully understand the removal mechanism and associated microbial diversity to support the overall HM removal mechanism described in this study. Nevertheless, the data obtained from this study will provide baseline information for future research on removing HMs in CWs using alternative substrates, specifically waste-derived materials.

The current study emphasizes the efficacy of four different types of substrates for treating wastewater containing high concentrations of HMs. Additionally, the study showcases how these organic and inorganic substrates remove HMs in CWs. The results indicated that CWs filled with eggshell, DPF, and Jhama brick were more effective in removing metals than the widely utilized gravel-filled CWs. Furthermore, the organic substrates (such as eggshell and DPF) consistently exhibited higher removal levels than the studied inorganic substrates (such as brick and gravel), even with elevated HM loadings. This study also demonstrates a high concentration of metal accumulation in various parts of C. indica and H. littoralis, particularly in the roots, and reveals their tolerance to increased HM loadings. The distribution models identified substrate accumulation as the main pathway for removing metals from wastewater in most CWs. Following substrate accumulation, the immobilization of metals through complex chemical formation with organic molecules might be the principal removal mechanism, particularly in the organic substrates-filled CWs. The findings of this study suggest that investigated waste-derived substrates are practical materials for removing HMs from wastewater in CWs. Additionally, the comprehensive results will assist in efficiently designing CWs for the remediation of high-concentration HMs contained in wastewater.

This study was partly supported by the Research and Development (R&D) Project from the Ministry of Science and Technology, Government of the People's Republic of Bangladesh [grant number: 33; FY: 2022–2023]. The authors also appreciate all the personnel of the following laboratories for their cooperation during their analytical work: Climate and Disaster Management Dept. Laboratory, Genome Center, JUST, and CSIRL Laboratory, JUST.

F.M.R. contributed to conceptualization, writing-original draft, data curation, methodology. Md.K.H. contributed to investigation, formal analysis, validation, software. Md.A.R. contributed to conceptualization, writing-original draft, review and editing, validation. Md.S.I. did investigation and formal analysis. P.K.S., Md.I.A., Md.W.K., and S.M.T.R. did formal analysis and validation. M.M.R. contributed to conceptualization and validation. Md.H.R. contributed to supervision, project administration, funding acquisition, and review and editing. J.Z. contributed to supervision, review and editing.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.