3.1 Laser flash photolysis of potassium persulfate in the presence of p-ASA
To generate a sulfate radical and study its reaction with p-ASA monoanion (pKa values: 2.00, 4.02, and 8.92 (Jaafar, 2001)), the flash (266 nm) photolysis of potassium persulfate at pH 7 was performed. To minimize a contribution of the intrinsic pASA photochemistry (Tyutereva et al., 2019) we used laser pulses of low energy (typically, around 1 mJ/pulse). Figure 1 demonstrates the evolution of the absorption spectra of PS in aqueous solution in the presence of p-ASA at pH 7 and the kinetic curves for three characteristic wavelengths. Immediately after the laser pulse, a wide absorption band was observed with a maximum at 440 nm, corresponding to the spectrum of the SO4•- radical (Ivanov et al., 2000). Then this band narrows and grows in amplitude within first two microseconds with a subsequent decrease of the signal on the time scale of tens of microseconds (Fig. 1).
The secondary intermediate observed at time delays more than 3 μs after the laser pulse was assigned to the p-ASA radical-cation (RNH2+•) formed in the reaction of a sulfate radical with a p-ASA (2). A similar intermediate, with an absorption maximum at 430 nm, was registered earlier in the direct photolysis of p-ASA (Tyutereva et al., 2019).
The decay of RNH2+• radical at the initial times can be described by a first-order equation (3) with an effective rate constant k3:
Since the SO4-• radical (Ivanov et al., 2000) undergoes effective recombination in reaction (1), an analytical solution for the schemes (1-3) cannot be realized. However, kinetic curves of transient absorption decay were calculated numerically by the 4th order Runge-Kutta method using the kinetic schemes (1-3) (see Fig. S1 and additional information in ESI for details). The influence of p-ASA concentration on the sulfate radical decay is the most pronounced at 490 nm (Fig. 2). For calculations we used the absorption coefficient of the sulfate radical at 490 nm, ε(SO42-•) = 1000 M-1cm-1 (Ivanov et al., 2000) and the recombination constant of this radical, 2k2 = 2.9×109 M-1s-1, determined from the second-order fit of the kinetic curve at 490 nm in the absence of p-ASA (Fig. 2, curve 1); these parameters were fixed during the subsequent calculations. The remaining parameters were fitted to achieve the best agreement between the calculated and experimental curves in a wide range of p-ASA concentrations (Fig. 2).
The numerical calculations gave the value of the rate constant for the reaction of SO4-• radical with p-ASA monoanion as k2 = (2.4 ± 0.4)×109 M-1s-1; the observed decay rate constant of RNH2+• radical and its absorption coefficient at 490 nm as k3 = (5.7 ± 1.5)×103 M-1s-1 and e(RNH2+•) = (3.4 ± 0.5)×102 M-1cm-1, respectively. At pH 3 the reaction rate constant of the sulfate radical with the neutral form of p-ASA is k2 = (7.3 ± 0.6) ×109 M-1s-1 that is three times greater than that obtained for the anionic one (Fig. S2). This indicates the strong influence of Coulomb repulsion during the interaction of two negatively charged species at pH 7. However both rate constants are close to diffusion-controlled limit, so one can conclude that SO4-• radical could readily oxidize p-ASA at a wide pH range even at low concentration of the target molecule.
3.2 Stationary photolysis of the Fe(Ox)33- complex with PS in the presence of p-ASA.
In our previous article (Tyutereva et all., 2020) we demonstrated that at neutral pH the Fe(III) oxalate complexes effectively generate •OH radical under UV irradiation, which reacts with p-ASA with high rate constant, (8.6 ± 0.5) × 109 M-1s-1. Subsequent oxidation of organic radical formed leads to complete degradation of both p-ASA and basic aromatic photoproducts with formation of inorganic As(V) mainly, under optimal conditions. It is also worth to note, that p-ASA has own photochemistry under 308 nm irradiation but it is negligible in our experimental conditions due to high absorption and photoactivity of Fe(III) complexes.
The presence of both Fe(III) oxalate and PS in the solution accelerates the photodegradation of p-ASA and gives an opportunity to reduce the working concentration of oxalate ions in the solution (from 0.5 to 0.12 mM) without changes in the high degradation efficiency of not only the target compound but the aromatic photoproducts as well (Fig. 3). This can be explained by the catalytic decomposition of PS by photogenerated Fe(II) ions with the regeneration of the initial Fe(III) oxalate complex and the formation of additional oxidative sulfate radical:
Fe(II) + S2O82- = Fe(III) + SO42- + SO4-• (4)
Addition of PS also allows to oxidize p-ASA at high concentration of the pollutant that cannot be done in the presence of iron oxalate alone. Figure 4 shows the dependence of the yield of different arsenic species on the irradiation time of p-ASA – Fe(III) oxalate system in the presence and absence of PS at initial concentration of p-ASA about 3 ppm (4×10-5 M). Without PS only a partial (about 50%) degradation of p-ASA was observed after 40 min of irradiation (Fig. 4A). The total content of arsenic including p-ASA and inorganic species at the end of irradiation (~1.7 ppm) was significantly less than the total arcenic concentration (~2.8 ppm) before irradiation that indicates an accumulation of some organic arcenic byproducts with questionable toxicity.
In the presence of PS (Fig. 4B) the complete degradation of p-ASA was observed with generation of As(V) mainly, which can be removed by standard water treatment procedures (Vircikova et all., 1996; Lawrence and Higgs, 1999). It worth to note that the concentration of As(V) measured by CZE method shows evident decrease after 20 minutes of irradiation comparatively with the experiment withot PS (Fig. 4B). To explain such puzzling effect the test experiments with model As(V) - Fe(III) oxalate – PS system without and with irradiation were carried out under the same conditions.
Figure 5A demonstrates the difference between actual (by preparation) concentration of As(V) measured using methods of ICP-AES and CZE before and after irradiation. Prior to irradiation both methods give results consistent with each other and actual As(V) concentration (Fig. 5A). In the case of photolysed samples, ICP-AES shows no influence of irradiation on the measured concentrations of As(V), while CZE demonstrates a pronounced decrease of As(V) by 0.7 – 2 ppm, depending on the initial arsenic content. This discrepancy increases with exposure (Fig. 5B) that gives a right to assume that a portion of As(V) transforms into another form or product, which is out of registration by CZE, e.g., conjugate or complex, on account of Fe(III)-oxalate degradation. Analogous loss of As(V) measured by CZE was reported in our previous work (Tyutereva et al., 2020) where oxidation of p-ASA was studied in Fe(III)-oxalate system in the absence of PS. According to (Wang et all., 2020), this unknown product is most likely an As(V) – ferric oxyhydroxides colloid.
Another argument for the assumption that part of arsenic is sorbed on photogenerated iron oxyhydroxides are the results of total As determination using ICP-AES analysis of undisturbed samples few days after photolysis (Fig. S3). The total arsenic concentration is restored by a shaking of samples prior the measurements. Figure 5B also illustrates this effect (curves 1 and 1'). This important observation indicates that both processes including p-ASA oxidation and As(V) sorption occurs during a single photochemical process. The significance of this effect for further application of this photochemical approach for p-ASA removal from natural waters is in the reduction of the number and the cost of water purification steps. We intend to continue research in this prospective direction in the nearest future.
Assuming that p-ASA is completely converted to inorganic forms of arsenic we can estimate the loss of As(V) in CZE measurements after 20 and 40 min of irradiation, accordingly. Result was lower that one predicted by photolysis of the model system with As(V) (³1 ppm, Fig. 5B). However, it should be taken into account that in the model system ROS, generated by excitation of Fe(III) oxalate, react immediately with the complex itself. In the presence of p-ASA these active species firstly react with target pollutant and start the degradation of complex only at the final stage of photolysis. So, we can expect the stabilization of Fe(III)-oxalate system in presence of p-ASA as compared with As(V).
After correction of As(V) concentrations measured by CZE, we can conclude that in the presence of 1 mM PS the complete degradation of p-ASA to inorganic arsenic occurs at concentrations of p-ASA up to 3 ppm. The ratio of As(III)/As(V) formed during p-ASA oxidation depends on the initial concentration of the pollutant. Figure 6 demonstrates the arsenic species distribution in solution after 20 min of irradiation depending on the starting concentration of p-ASA. Almost complete conversion of p-ASA to inorganic arsenic could be seen (Fig. 6). Another important result is the decrease of As(III)/As(V) ratio from 0.21 to negligible value (Fig. 6, inset) with the lowering of p-ASA concentration from 3.3 to 0.44 ppm. Apparently, at low p-ASA concentrations all photogenerated As(III) is completely oxidized by PS. High concentrations of p-ASA lead to higher consumption of PS that, in turn, results in incomplete degradation of photogenerated As(III). At real concentration of the organic arsenicals in contaminated surface water (~ tens ppb, Mangalgiri et all., 2015) all p-ASA is expected to fully converts to As(V) in Fe(III)-oxalate – PS system, which significantly reduces the toxic effect of p-ASA on the environment.