3.1. Dissipation of ATZ in water
ATZ residues in water decreased with incubation time for all tested concentrations regardless of whether adding GO or not, and the dynamic dissipation fitted the first-order kinetics model well (Fig. 1). ATZ dissipation rate on day 21 was significantly higher in treatments with GO than in treatments without GO (Fig. 2). The treatments of ATZ+GO+F had significantly greater dissipation rate constants than treatments of ATZ and ATZ+GO, while there was no significant difference between ATZ and ATZ+GO (Fig. 1). ATZ alone showed slow dissipation, with the half-lives from 64.7 to 184.4 days over the 21-d period; ATZ with GO showed significantly faster dissipation with the half-lives from 54.9 to 110.8 days (Fig. 3a). These results indicated that GO significantly accelerated ATZ dissipation and thus ATZ exhibited decreased persistence in water.
3.2. Adsorption capacity for GO towards ATZ
The adsorption capacity for GO towards ATZ reached the maximum value within 14 days when ATZ initial concentration was not above 1.0 mg·L-1, while it gradually increased at level up to 2.0 mg·L-1 during the whole incubation period (Fig. 3b). In addition, the adsorption equilibrium was shifted toward higher adsorption capacity by the increasing ATZ concentration. This result was in agreement with previous work that reported the larger ATZ adsorption amount at higher GO concentration owing to more available adsorption sites for ATZ binding (Muthusaravanan et al., 2021). The adsorption equilibrium time also exhibited an extending trend with the increase in ATZ concentration possibly due to the enhanced competition for the active binding sites on GO surface among ATZ molecules. As adsorption gradually reached an equilibrium, GO contribution to ATZ dissipation decreased as incubation time went on (Fig. 3c). Therefore, ATZ dissipation in the late stages of the experiment (after 14 days) mainly resulted from the natural degradation of ATZ, as reflected by the smaller increase in ATZ dissipation rate on day 21 compared with day 14.
3.4. ATZ metabolites in water
For all treatments with or without GO during the 21-day experimental period, the contents of ATZ metabolites in the water increased with the initial ATZ concentrations and incubation time (Fig. 4). The metabolites at higher levels were detected in the treatments of higher initial concentrations of ATZ. For the same initial concentration of ATZ, the ordering of metabolite levels was HYA<DIA<DEA. DEA and DIA are two primary dealkylatrazine derivatives, and DEA is a significant metabolite in surface water owing to preferential removal of ethyl side chains relative to isopropyl groups in the N-dealkylation pathway of ATZ degradation process (Thurman and Fallon, 1996; Lutze et al., 2015). Thus similar to ATZ alone, deethylation significantly dominated over deisopropylation in formation of the production of N-dealkylated forms of ATZ combined with GO. HYA, a non-herbicidal metabolite of ATZ, is formed mainly through hydrolytic dechlorination process (Mandelbaum et al., 1993). Therefore, the comparatively lower HYA concentration indicated that ATZ dealkylation preferentially occurred compared to its hydrolysis, regardless of whether or not GO coexisted. It follows that GO does not alter the preferential pattern of ATZ degradation in water.
The results of ATZ metabolite contents obtained with and without GO were summarized in Fig. 4. When ATZ initial concentrations were 0.1 and 0.5 mg·L−1, ATZ alone and ATZ combined with GO showed no significant difference in DEA contents within the first 14 days after incubation; while significantly higher DEA contents were observed in treatments with GO addition than in treatments of ATZ alone (Fig. 4a, d). When ATZ initial concentration was 1.0 and 2.0 mg·L−1, DEA contents of ATZ with GO were significantly lower than ATZ alone after 7 days of incubation (Fig. 4g, j). DIA contents of ATZ alone was generally higher than those of ATZ+GO (Fig. 4b, e, h, k), HYA contents of ATZ alone was lower than those of ATZ+GO while during the whole incubation period (Fig. 4c, f, i, l). These results demonstrated that the presence of GO inhibited ATZ dealkylation while facilitated hydroxylation. ATZ dealkylation reactions occurred mainly by microorganisms intermediation (Mandelbaum et al., 1993). The engineered nanoparticles inhibited ATZ biodegradation through the toxic effects on microbial activity; at the same time, the decreased contents of DEA and DIA may be related to the decreasing bioavailability resulted from the adsorption of ATZ on GO (Zhang et al., 2015). The adsorption by GO decreased ATZ utilization by the microbial population, and thus led to a reduced rate of transformation by dealkylation. Adding of other sorbent, such as biochar, also had similar effects on the ATZ degradation into DEA and DIA in the soil (Huang et al., 2018). Both DEA and DIA with greater mobility exerted phytotoxic activity, and DEA was considered to be nearly toxic as ATZ (Winkelmann and Klaine, 1991; Seybold and Mersie, 1996). The higher dissipation rates of ATZ with limited transformation into chloroderivatives indicated that ATZ combined with GO posed a lower environmental risk than ATZ alone did. Concerning the promoting effect of GO on ATZ hydrolysis, the available evidence indicated that hydrogen bonding, between the adsorbent protonated carboxyl groups and ATZ ring nitrogen atom, resulted in apparently catalyzed hydrolysis by adsorption (Armstrong and Chesters, 1968). The abundant carboxyl groups at GO surface can provide acidic hydrogen atoms for hydrogen bonding (Shayimova et al., 2021). The previous study has confirmed that the hydrogen bonding mechanism was an important contributor of ATZ adsorption onto GO (Muthusaravanan et al., 2021). Thus, GO also may catalyze ATZ hydrolysis through the specific adsorption mechanism mentioned above. Given that HYA is a nontoxic metabolite, hydroxylation of ATZ to HYA is generally considered an efficient remedia1 measure for ATZ residue problem (Mu et al., 2019). It provides ideas on usage of carboxyl functionalized GO for larger ATZ transformation into HYA and decreasing generation of toxic chloro-dealkylated intermediates. Additionally, the generation rate of HYA was closely related to pH of the environment medium (Borggaard and Streibig, 1988). In this study, pH values of the solution increased with incubation time, but pH increase rate of solution with ATZ and GO was lower than that with ATZ alone (Fig. 5). Especially in treatments of high ATZ concentrations (1.0 and 2.0 mg·L-1), this difference was particularly significant (Fig. 5c, d). The solution with lower pH can provide more hydrogen ions which may also promote ATZ hydrolysis. In addition, the contents of three ATZ metabolites were all greater for ATZ+GO than for ATZ+GO+F, which was presumably attributed to adsorption of GO to them. This result was expected because these metabolites had very similar structures to parent ATZ and consequently being adsorbed by GO in a similar adsorption way.
The incubation time for ATZ metabolites being detected in water (Td) more clearly reflected the features of ATZ degradation in the presence of GO. DEA can be detected on day 1 of incubation in all treatments. Furthermore, of the three determined metabolites, it was also the first one to be detected for ATZ at the lowest initial level (0.1 mg·L-1). HYA was the last metabolite detected. These findings further confirmed the domination of ATZ degradation pathway of dealkylation. GO had no effects on Td of DEA and DIA, while HYA was detected earlier in the presence of GO than in absence of GO, indicating that GO accelerated the hydroxylation of ATZ. This inference was supported by data of HYA contents in water discussed above.