Performance Comparison of Different Constructed Wetlands Designs for the Removal of Personal Care Products

This research investigates the performance of four types of constructed wetlands (CWs): free water surface CW (FWSCW), horizontal flow CW (HFCW), vertical flow CW (VFCW), and hybrid CW (HCW) for the removal of 20 personal care products (PCPs), based on secondary data compiled for 137 CWs reported in 39 peer reviewed journal papers. In spite of considerable variation in the re-moval efficiency of PCPs, CWs prove to be a promising treatment technology. The average removal efficiency of 15 widely studied PCPs ranged from 9.0% to 84%. Although CWs effectively reduced the environmental risks caused by many PCPs, triclosan was still classified under high risk category based on effluent concentration. Five other PCPs were classified under medium risk category (triclocarban > methylparaben > galaxolide > oxybenzone > methyl dihydrojasmonate). In most of the examined PCPs, adsorption and/or sorption is the most common removal mechanism followed by biodegradation and plant uptake. The comparatively better performance of HCW followed by VFCW, HFCW, and FWSCW might be due to the co-existence of aerobic and anaerobic conditions, and longer hydraulic retention time enhancing the removal of PCPs (e.g., triclosan, methyl dihydro-jasmonate, galaxolide, tonalide, and oxybenzone), which are removed under both conditions and by adsorption/sorption processes.

Domestic (55%) and Industrial discharge (45%) (*); The Population equivalent (PE) is calculated based on the common relation 1 PE = 60 g BOD d -1 . BOD values were approximated using the ratio COD/BOD = 2 in the studies where BOD was not reported [10,13], and COD values were approximated using the ratio COD = 2BOD in the studies where COD was not reported [5,12,14,17,19].  The Population equivalent (PE) is calculated based on the common relation 1 PE = 60 g BOD d -1 . BOD values were approximated using the ratio COD/BOD = 2 in the studies where BOD was not reported [23,25,26], and COD values were approximated using the ratio COD = 2BOD in the studies where COD was not reported [14,19,21].  The Population equivalent (PE) is calculated based on the common relation 1 PE = 60 g BOD d -1 . BOD values were approximated using the ratio COD/BOD = 2 in the studies where BOD was not reported [36,37], and COD values were approximated using the ratio COD = 2BOD in the studies where COD was not reported [5,14,19,29].

Supplementary materials 2: Contribution of removal mechanisms of PCPs in
experimental studies and CWs. Supplementary materials 3: The estimated statistics (mean and standard deviation of concentration and removal) for personal care products for which three or more data points were available. Page | 39

Supplementary materials 4: The selected statistics on risk quotient for nine PCPs based on effluent concentration in
CWs.   Dietz and Schnoor [47]; Alvarez et al. [44]; Pilon-Smits [48]; Le-Minh et al. [45] Plant uptake Although, the Log Dow of organic compounds is nearly the same as Log Kow but the modification in ionizable functional groups may affect their removal by biological treatment processes.
Zhang et al. [50] Log Koc This is the ratio of the mass of a compound that is adsorbed in the soil per unit mass of organic carbon in the soil.   [53] reported that it is persistent to biodegradation in biological treatment. However, some recent studies attributed its removal to biodegradation in biological wastewater treatment [54][55][56]. This can be explained by its higher removal efficiency in HCW containing two types of substrate media (gravel-based VFCW and subsequent sand-based VFCW). This is due to the reason that sand media provided a larger available surface area for microbial growth and higher oxygen to promote the elimination of substances that are majorly removed by aerobic biodegradation pathways. It was poorly removed by gravel-based HFCW and negatively removed in gravel-based VFCW.

Preservatives Propylparaben
The removal efficiency of propylparaben was moderate in FWSCW (75±20%) and it was negatively removed in HFCW (-21±21%) (  (Table S1). In addition to direct uptake, the indirect positive effects of plants presence such as degradation by enzymatic exudates and aerobic biodegradation facilitated by the release of root exudates (such as carbohydrates and amino acids), and oxygen by the plant roots in the rhizosphere [17,37,40,57] might contribute to its removal [40]. These processes might take place in HFCW but its negative removal suggests that photodegradation could contribute to its removal in FWSCW.

Insect repellents N,N-diethyl-meta-toluamide
The removal efficiency of N,N-diethyl-meta-toluamide was higher in HFCW (98%) compared with HCW (80%). It was poorly removed in VFCW (28%), but FWSCW did not show any removal (0.0%) ( uptake. In spite of this, the lower contribution of plants (9.1%) in hydroponic system (Spirodela polyrhiza) compared with the control without plants (17% and 7.9%, respectively) reveals that this removal pathway is contributing very less in its removal ( Figure 3 and Table S5). Furthermore, its organic carbon sorption capacity is also low (Log Koc = 1.76) ( Table 4). Thus, the efficient removal in CWs might be due to enhancement in biodegradation processes in the presence of plants.
However, in the E-coli biodegradation experiment the highest removal observed was very low (4.5%) [7] (Figure 3 and Table S5). The highest removal in HFCW followed by HCW indicates that predominantly anaerobic and slightly aerobic conditions might favor its removal [15,24]. In FWSCW, the major process contributes in PCPs removal is photodegradation. This compound is not removed in FWSCWs, which can be explained by its persistence against light. In photodegradation experiment, it has been demonstrated that this compound is not light sensitive (1.2% removal efficiency at highest by photodegradation) [7] (Figure 3 and Table S5).

Antiseptics Triclocarban
The removal efficiency of triclocarban was higher in HFCW (81±13%) but it was negatively removed in HCW (-14%) (Table S10). It is almost insoluble in water (0.11 mg L -1 at 25 °C), highly hydrophobic (Log Kow = 4.90; Log Dow = 4.90) with moderate molecular weight (315.6 g mol -1 ) (Table 4), which suggest its removal by adsorption onto soil particles. Its high organic carbon sorption capacity (Log Koc = 3.73) also favors its removal by sorption onto organic surfaces. This can be seen by its significant positive correlation with the removal of chemical oxygen demand (COD) [16]. Consistent with that, Zhu and Chen [33] estimated its high concentration in sediments reaching an average value of 6.6 mg kg -1 , which indicates that sorption to sediments or sludge mainly contributed to its removal.

Cashmeran
The removal efficiency of cashmeran was moderate in HCW (56±21%) ( This can be explained by its higher removal efficiency in winter compared with summer (50% and 39%, respectively) [36] (Table S4). Since adsorption is an exothermic process, which is favored by low temperature (in winter). Its high organic carbon sorption capacity (Log Koc = 2.99) also favors its removal by sorption onto organic surfaces. This can be seen by the efficient removal in FWSCW compared with pond of the HCW (combination of two ponds and one FWSCW). The roots of the plants in the CWs increase the accumulation of organic matter as well as the sorption capacity [31].

Tributyl phosphate
The removal efficiency of tributyl phosphate was moderate in HFCW (48±19%) and HCW (43±15%) (Table S10). It is highly hydrophobic (Log Kow = 4.00) with slight water solubility (280 mg L -1 at 25 °C), and moderate molecular weight (266.32 g mol -1 ), which favors its removal by adsorption to substrate. Its high organic carbon sorption capacity (Log Koc = 3.60) ( Table 4) suggests its removal by sorption onto organic surfaces. This is explicit by its significant positive correlation with the removal of total suspended solids (TSS) and COD [13]. This can also be seen by its slightly better removal efficiency in summer compared with winter (37% and 32%, respectively) [36] (Table S4) due to the higher activity of the rooted plants in the warm season, since the roots of the plants in the CWs increase the accumulation of organic matter and the sorption capacity [31]. The other indirect positive effect of plants presence such as enhancement in biodegradation might also contribute to its removal, which is revealed by its strong positive relationship with ammonium-nitrogen (NH4 + -N) removal [36].

Triphenyl phosphate
The removal efficiency of triphenyl phosphate was moderate in HFCW (79±5%) and HCW (53±21%) ( Table S10). It is almost insoluble in water (1.9 mg L -1 at 25 °C), highly hydrophobic (Log Kow = 4.59) with moderate molecular weight (326.29 g mol -1 ), which suggest its adsorption to substrate as a major removal mechanism. Its high organic carbon sorption capacity (Log Koc = 4.03) ( Table 4) reveals its removal by sorption onto organic surfaces, which can be explained by its significant positive correlation with the removal of TSS and COD [13]. It shows a higher removal efficiency in summer compared with winter (67% and 38%, respectively) [36] (Table S4) because of efficient growth of the rooted plants in the warm season also reveals the contribution of this removal pathway. Furthermore, the enhancement in biodegradation in the presence of plants might also contribute to its removal, which is shown by its significant positive correlation with NH4 + -N removal [36].

Tris (2-chloroethyl) phosphate
The removal efficiency of tris (2-chloroethyl) phosphate was very low in HFCW (24±6%) and HCW (17±10%) ( Table S10). It is highly water soluble (7.82 g L -1 at 25 °C) with moderate hydrophobicity (Log Kow = 1.44), and moderate molecular weight (285.48 g mol -1 ). It is neutral in nature under neutral conditions (pH = 7) (Table 4), which favors its removal by plant uptake. Its moderate organic carbon sorption capacity (Log Koc = 2.48) also suggests its removal by sorption onto organic surfaces. This is explicit by its significant positive correlation with the removal of TSS and COD [13]. Considering the efficient growth of the rooted plants in summer and the ability of roots to increase the accumulation of organic matter as well as the sorption capacity [31], its higher removal efficiency was observed in summer compared with winter (24% and 7.0%, respectively) [36] (Table S4). However, this compound showed very low removal because of its high recalcitrance [13,58], which might be due to the presence of chlorine in its structure.

Sulisobenzone
The removal efficiency of sulisobenzone was very low in HFCW (9.0±1.9%) (Table S10). It is highly water soluble (20.3 g L -1 at 25 °C) with moderate molecular weight (308.31 g mol -1 ), highly hydrophilic (Log Kow = 0.37), and anionic under neutral conditions (pH = 7) (Table 4), which might hinder its uptake by the plants as well as adsorption to the substrate. Its anionic form and very low lipophilicity (Log Dow = -0.53) ( Table 4) hinders its partition into lipophilic cell structure of negatively charged biomembranes of the plant roots because of the charge repulsion. Although its molecular weight is moderate, but it is most hydrophilic (Log Kow < 1.0) and the most water soluble (WS > 1000 mg L -1 ), thus adsorption cannot be considered as a removal mechanism because for the sorption/sedimentation of these types of compounds more time is required [38]. Therefore, this low removal can be attributed to biodegradation, which has been established in other wastewater treatment technologies. For instance, in contact with activated sludge it was degraded in aerobic batch experiments forming at least nine transformation products (TPs) [59]. Beel et al. [59] proposed biodegradation pathway based on the structure of the TPs identified and the sequence of their formation.

Supplementary materials 7:
The removal efficiency of 15 widely studied PCPs and the results of one-way ANOVA and z-Test for comparison of means of six selected PCPs.