Assessing the efficacy and mechanisms of glycol-contaminated water treatment through floating treatment wetlands

Constructed wetlands (CWs), as a nature-based solution, have emerged as an important approach for improving water quality while simultaneously delivering co-benefits such as biodiversity enhancement (Oral et al. 2020; Castellar et al. 2022). Floating treatment wetlands (FTWs) represent one of many diverse applications of CWs, in which emergent wetland plants grow hydroponically on mats or structures floating on the surface of a pond (Pavlineri et al. 2017). FTWs offer some advantages over other subsurface flow CWs, such as the absence of requirements for substrate materials, which significantly reduces construction costs and eliminates clogging risks because no substrate pores are present (Colares et al. 2020). Moreover, FTWs are advantageous as a stormwater management tool since they can fit easily into existing retention ponds and adapt to varying water depths characteristic of event-driven stormwater systems (Headley & Tanner 2011).

Previous studies have shown the efficacy of FTWs in reducing organics (COD and BOD₅), nutrients, TSS, and toxic metals (e.g. Cd, Cu, and Zn) in wastewater treatment (Afzal et al. 2019; Park et al. 2019; Colares et al. 2020). With increasing concerns about water pollution and stringent regulations, there has been a growing interest in the removal of emerging contaminants (Lv et al. 2016a; Zhang et al. 2017). One such pollutant is propylene glycol, which is widely used in the manufacture of polyester resins and formulation into functional fluids such as anti-freeze, aircraft anti-icing and de-icing fluids, cosmetics, pharmaceuticals, personal care products, and pesticides (West et al. 2014; Rogers et al. 2019). Although low concentrations of glycol are not directly toxic to aquatic organisms, elevated concentrations can lead to increased microbial activity that drives a decrease in oxygen levels, which can be very harmful to aquatic life (Freeman et al. 2015; Nott et al. 2020; Exton et al. 2023).

Subsurface flow CWs have demonstrated promising results in the removal of glycol from wastewater, achieving up to 99% removal under initial concentrations of over 1,000 mg/L (Higgins & Maclean 2002; Murphy et al. 2015). However, one of the main sources of environmental contamination of glycol is from de-icer applications, such as in airports, which enters into environment through weather-driven storm water runoff (Corsi et al. 2012; Nott et al. 2020). As such, FTWs have been tested and operated at Heathrow Airport to treat glycol, owing their ability to cope with varying water levels. The average glycol removal efficiency achieved 54% in the initial pilot-scale FTW (3 × 5 m) trial (Chong et al. 1999). However, limited published data were available regarding the treatment performance in their applied FTW system (Richter et al. 2003). Therefore, despite numerous studies on the effectiveness of FTWs in removing nutrients and other pollutants (Afzal et al. 2019), research on their ability to eliminate glycol remains scarce.

In FTWs, elucidating the removal mechanisms and quantifying different removal pathways of the pollutants through biological processes (e.g. plant uptake and microbial biodegradation) and physicochemical processes (e.g. photodegradation and evaporation) are essential for determining optimal implementation strategies (Barco et al. 2021). Sriprapat et al. (2011) reported that the aquatic macrophyte (Echinodorus cordifolius) was able to uptake diethylene glycol while treating the wastewater. The selection of plant species can substantially impact treatment performance by modulating plant uptake capabilities (Oliveira et al. 2021). In terms of removing ethylene glycol from industrial wastewater, Acorus calamus demonstrated the best treatment performance when compared to Phragmites australis and Typha latifolia during phytoremediation experiments (Marecik et al. 2013). Apart from plant uptake, the plant rhizosphere provides a conducive environment, oxygen, and carbon sources (root exudates) for microbial growth (Lv et al. 2016b; Zhang et al. 2017), which can significantly contributes to the biodegradation of glycols (Higgins & Maclean 2002; Murphy et al. 2015). Previous FTW research highlighted that the mechanisms underlying nutrient removal in FTWs remain incompletely understood (Colares et al. 2020), and this knowledge gap also pertains to glycol removal. We hypothesize that FTWs can effectively remove glycol from contaminated water, and the evaporation process may play a significant role in removal, considering that glycol is a volatile organic compound (VOC). Consequently, a quantitative investigation of removal mechanisms is vital for future FTW implementations.

In this study, three native wetland species, namely Juncus effusus, Carex acutiformis, and A. calamus, and their mixed planting were employed in mesocosm-scale FTWs to evaluate the effects of plant species and planting strategies (monoculture vs. polyculture) on propylene glycol removal from the water column. Moreover, a separate experiment was conducted to quantitatively investigate the removal pathways of propylene glycol, including microbial biodegradation, plant uptake, and the combined effects of photodegradation and evaporation. Additionally, the role of FTWs in supporting the biodiversity and carbon sequestration was discussed.

Three commonly used native wetland plant species, namely J. effusus, C. acutiformis, and A. calamus, were selected for this study due to their abilities to thrive in wetland environments and their potential for wastewater treatment applications. The leisis floating mat were purchased from a local botanic shop (Beaver Plants Ltd, Lingfield, UK). The propylene glycol and antibiotic sulfamethoxazole were sourced from Fisher Scientific and Scientific Lab Supplies, respectively.

The mesocosm-scale experiment was conducted over 6 weeks from 21 July to 1 September 2022, with the initial 2 weeks allocated for stabilisation to facilitate the growth of plants and biofilm. Prior to the experiment, the roots of all plants were washed, the wet biomass was measured and documented. To facilitate the transition of plants from soil to hydroponics, all cleaned plants were submerged in a 1% concentration of Formulex nutrient solution (Growth Technology Ltd, Somerset, UK) for a week. Healthy plants were selected for the experiment. The experimental setup was placed in a large glasshouse to maintain ambient environment while avoiding precipitation.

The second experiment was conducted in smaller 1-L buckets, aimed at quantifying the different pathways involved in propylene glycol removal in FTWs planted with different plant species. Three different FTW configurations (groups) were established within this experiment. A common FTW setup was applied to group 1 (plant + floating mat), with the plant supported by a 5 × 5 cm floating mat. To reduce propylene glycol loss through evaporation and photodegradation, each bucket in group 2 was covered with a lid, leaving a Ø 4 cm hole for the plant (Figure 1(c), plant + floating mat + lid). To assess the contribution of microbial biofilm on glycol treatment, 30 mg/L of sulfamethoxazole was added to the solution to kill or inhibit bacterial growth in group 3 (plant + floating mat + antibiotics). This antibiotic was chosen because it is frequently present in the environment and has been used to inhibit microbial activity in previous studies (Schmidt et al. 2012). The antibiotic dosage used in this experiment was based on a prior study that evaluated the impact of antibiotics on FTWs (Garcia Chance et al. 2020). The initial working solution and plant wet biomass were 0.8 L and 100 g in each bucket, respectively. The experimental procedures were identical to those in the first series of experiments.

The initial solution for both experiment series was prepared using local river water (Chicheley Brook, Buckinghamshire, England) and supplemented with propylene glycol. The initial concentration of propylene glycol was set at 1,250 mg/L with a total organic carbon (TOC) of 1,680 mg/L to replicate typical levels of this contaminant in airport runoffs (Murphy et al. 2015). Furthermore, a 1% concentration of Formulex nutrient solution (Growth Technology Ltd, Somerset, UK) was added to each bucket weekly to support plant growth.

Water levels were checked, physicochemical water quality assessed with a multiparameter meter, and water samples taken for glycol concentration at day 2, day 7, and then weekly. Change in water level was used to determine water loss due to evapotranspiration. To assess the physicochemical water quality, HANNA HI-9813-6 and HI-9147-04 probes were used to measure pH, temperature, conductivity, and dissolved oxygen (DO) directly in the buckets. Prior to measurement, all probes were calibrated according to the instrumentation manual, and electrodes were carefully cleaned between measurements to prevent cross-contamination. Water samples (20 mL) were collected using a 25-mL syringe, and prior to each sampling, the buckets were gently shaken for thorough mixing. The samples were than filtered through a 0.2 μm retention cellulose acetate membrane syringe filter in preparation for propylene glycol analysis.

The wet biomass of each plant species per experimental unit was quantified at the beginning and end of the experiment using a precision weighing scale. Furthermore, the physical changes in root mass and the presence of different species were recorded.

The normality of the mean glycol concentration data was evaluated using the Shapiro-Wilk W test, which was followed by an Analysis of Variance (ANOVA) evaluation to assess the effect of plant species in FTWs on glycol concentration changes. JMP v13 was employed to perform this statistical analysis, and p ≤ 0.05 were used as the threshold for statistical significance.

The experiment was carried out in the summer season in the UK, where the water temperature remained between 21 and 25 °C throughout the study period (Table 1). The pH level in all FTW mesocosms significantly increased from approximately 6.4 at the beginning to 7.2–7.4 after 1 week, regardless of the plant species differences in the system. The electrical conductivity (EC) levels in the solution remained at a similar level (1.9–2.0 mS/cm) after 1 week in the control group. However, in the presence of wetland plants (including the mixed planting FTW), the EC significantly dropped to 1.2–1.5 mS/cm after 1 week. The DO concentration was 33.9 ± 0.2 mg/L in the freshly prepared solution. After 1 week, the DO decreased to 1.3–1.7 mg/L in the FTWs with plants, but no significant differences were observed between each group. The final DO level in the control group was significantly lower (0.7 ± 0.3 mg/L) than in all other systems.

Although the final concentration of glycol is always the most important parameter in terms of regulatory and potential environmental risk, water evaporation effects, particularly in the open water surface FTW, may significantly concentrate the pollutant and lead to a misleading understanding of the system's functionality. In this study, the removed mass of propylene glycol from all systems was analysed (Figure 2(c)). Around 11–18% percentage removal of glycol was obtained in the control group (Figure 2(a)), and 7–11% of mass removal could be accounted for after considering the water loss in the system. Other planted FTWs showed a constant high mass removal (89–99%) of glycol after 7 days of treatment throughout the four-week study (Figure 2(c)). Similar to the results of percentage removal, FTW planted with J. effusus had a lower average mass removal on average (4,659 mg, 93%) than with C. acutiformis (4,912 mg, 98%), A. calamus (4,939 mg, 99%), and polyculture (4,934 mg, 99%) (Figure 2(d)).

A similar trend was observed after 7 days of treatment across all groups (Figure 4(c)), wherein 90–98% of the glycol mass was removed from the FTWs with standard setups (Group 1 in Figure 4(c)). However, groups with the additional lid (Group 2) and antibiotic treatment (Group 3) exhibited residual glycol masses of 169–237 mg (76–83% removal) and 381–425 mg (58–62% removal), respectively. The contributions of microbial biodegradation, plant uptake and evaporation increased to 37.9, 45.2, and 15.5% for A. calamus, 33.3, 43.1, and 18.9% for J. effusus, and 39.7, 37.9, and 19.5% for C. acutiformis, respectively (Figure 4(d)). The results confirmed that microbial biodegradation and plant uptake persist as the primary pathways for propylene glycol removal.

The observed reduction in EC (Table 1) in the solution was attributed to the removal of dissolved glycols (Figure 2) and nutrients, such as organics, nitrogen, and phosphorus (Wang et al. 2021). In the absence of plants in the control group, the EC remained similar to the initial level, with less glycol removal (Table 1). The relatively low DO levels observed in this experiment (Table 1) highlight the occurrence of oxygen-consuming processes required for glycol removal through microbial biodegradation, which can have toxic effects on aquatic biota (Freeman et al. 2015). The DO in the control group dropped to an extremely low level (0.7 mg/L), however, wetland plant roots can release oxygen into the solution (Brix 1997), and the DO in the planted FTWs remained at a relatively higher level (1.3–1.7 mg/L) compared to the control group.

This study demonstrated that all of the selected plant species could remove glycol up to 99% within 7 days (Figure 2). The results were generally consistent with previous findings in glycol phytoremediation, which were supported by earlier studies showing above 91% removal within 12–15 day under a higher initial concentration of 2,000 mg/L (Teamkao & Thiravetyan 2010; Sriprapat et al. 2011). A. calamus was the most efficient species, especially when planted as a single species, despite having a smaller wet biomass and root structure at the beginning of the experiment (Figure 3(a)). However, the A. calamus achieved the highest biomass growth (Figure 4(a)) and root density (Figure 4(d)), which may contribute to the higher removal performance of glycol through plant uptake and microbial biodegradation processes. In contrast, the polyculture FTW did not exhibit such high biomass growth during the experiment but still achieved nearly complete removal of glycol. Previous studies have shown that mixed cultures of plants can enhance microbial biodiversity and activity in wetlands (Button et al. 2016; Wang et al. 2022), which may have contributed to the high removal efficiency of glycol observed in this study. In summary, this study confirms that the selected UK native macrophytes are suitable and effective for the removal of propylene glycol and that A. calamus or mixed species could provide the best removal performance in FTWs.

Propylene glycol is a VOC that should be removed via evaporation and photodegradation (Chen et al. 2017). Nevertheless, only 10.8–13.8% (Figure 4(b)) and 15.5–19.5% (Figure 4(d)) of the glycol mass were removed via these processes after 2 days and 7 days of treatment, respectively. These low values may be due to the low exposed water surface, as much of the water surface was covered by the floating mat and plant within the experimental mesocosms. Previous studies have reported that FTWs can effectively reduce evaporation and inhibit phytoplankton growth through shielding and nutrient uptake (Jones et al. 2017). Additionally, lower removal rates within mesocosm experiments may be attributed to lower heat transfer between air and stagnant water compared to flowing systems, which can lower evaporation rates (Garcia Chance et al. 2020).

The contributions of plant uptake and microbial biodegradation of glycol in this study, accounting for over 80% of the total, were supported by previous studies that claimed that microbial biodegradation and plant's effect could be the key factors in removing glycols (Chong et al. 1999; Higgins & Maclean 2002; Murphy et al. 2015). Moreover, the contributions of microbial biodegradation and plant uptake exhibited similar patterns during both the 2- and 7-day investigations (Figure 4). The reduced removal efficiencies of glycol (Figure 4(a)) compared to normally operated FTWs (Figure 2(b)) indicated that bacterial composition had been altered as a result of the antibiotic addition. Tong et al. (2020) also reported that exposure to ofloxacin and tetracycline can suppress the growth and activity of overall microbial bacteria. The substantial removal of glycol through plant uptake may be attributed to the rapid growth of the plants during that specific period. With regard to the long-term operation of the FTW in the applied system, as the plant biomass reaches a stable period, microbial biodegradation is expected to become the dominant process for glycol removal in the system. Long-term monitoring, with the support of microbial community analysis, is essential to gain a more accurate understanding of the contributions, which is recommended for future studies.

A whole lake experiment showed that the application of FTWs can aesthetically improve the lake environment and increase biodiversity by attracting more local fish and other wildlife (Henny et al. 2020). In this study, the presence of small wild plant species did not affect the FTW's performance as averagely 93% glycol removal was achieved throughout the observation period (Figure 2). The results supported the diverse benefits of macrophytes in increasing biodiversity while removing glycols. Moreover, FTWs are a feasible net zero carbon strategy that should be encouraged within WWTPs, as they have a low carbon footprint with 50% fewer greenhouse gas (GHG) emissions (Nahlik & Mitsch 2006). The previous study has demonstrated that FTWs can act as sinks for carbon, considering different types of GHGs, with carbon storage capability up to 25 g CO₂-eq/m²/d (Sun et al. 2019). As GHG emissions from wetlands vary widely, depending mostly on location, the source of water received by the wetlands and wetland vegetation (D'Acunha & Johnson 2019). Therefore, directly monitoring at larger scale applied FTWs in the context of glycol removal is needed to quantify the co-benefits.

This study provides evidence that FTWs are highly effective in mitigating the contaminant propylene glycol. All plant species tested in the FTWs were able to remove up to 99% of the glycol within 1 week under a high initial concentration of 1,250 mg/L. A. calamus and the mixed species were found to be the most effective planting strategy for rapidly reducing glycol concentrations compared to other plant species. The key mechanisms responsible for glycol removal were found to be microbial biodegradation and plant uptake, which together accounted for nearly 80% of glycol removal. The contributions of evaporation and photodegradation were responsible for the remaining glycol mass removal; however, their contributions are estimated to be higher in applied FTWs with moving waters. Overall, this study demonstrates the potential of FTWs for treating glycol-contaminated water while potentially delivering additional co-benefits such as supporting biodiversity and carbon sequestration.

This research was funded by Heathrow Airport Ltd and conducted at Cranfield University. The authors gratefully acknowledge the support of Andy Freeman and Clodagh Murphy from ARM Ltd, who provided valuable insights and consultation based on their experience with treatment wetlands and glycol management.

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

The authors declare there is no conflict.