Different Role of Bisulfite/Sulfite in UVC-S(IV)-O2 System for Arsenite Oxidation in Water

It is of interest to use UV-sulfite based processes to degrade pollutants in wastewater treatment process. In this work, arsenic (As(III)) has been selected as a target pollutant to verify the efficacy of such a hypothesized process. The results showed that As(III) was quickly oxidized by a UV-sulfite system at neutral or alkaline pH and especially at pH 9.5, which can be mainly attributed to the generated oxysulfur radicals. In laser flash photolysis (LFP) experiments (λex = 266 nm), the signals of SO3•− and eaq− generated by photolysis of sulfite at 266 nm were discerned. Quantum yields for photoionization of HSO3− (0.01) and SO32− (0.06) were also measured. It has been established that eaq− does not react with SO32−, but reacts with HSO3− with a rate constant 8 × 107 M−1s−1.


Introduction
Recently, advanced oxidation processes (AOPs) using sulfite (S(IV)), as effective strategies for the removal of contaminants, have attracted a lot of attention [1][2][3]. Though there have been some reports on the UVC(254 nm lamp)-S(IV) system, the hydrated electrons (e aq − ) generated have mainly been used for reductive dehalogenation [4,5]. The e aq − could be only formed in sufficiently high concentration under oxygen-free conditions according to reactions (1) and (2). However, in most natural aqueous environments, dissolved oxygen (ca. 0.2 mM) is invariably present, and this may impose some limits on reductive dehalogenation using UVC-S(IV) system. SO 3 •− , another product of S(IV) photolysis, a relatively mild oxidant [6], reacts rapidly with oxygen to produce SO 5 •− radicals and then form SO 4 •− and even HO• (reactions (3)-(7)) [7][8][9][10][11][12]. These radicals possess higher redox potential and can effectively degrade pollutants [2,13]. Hence, it may also be possible to oxidize or degrade contaminants in an oxygen-containing sulfite system under UVC irradiation (namely, a UVC-S(IV)-O 2 system).
The environmental chemistry and pollution control of arsenic have been extensively studied due to its high toxicity and carcinogenicity. In an investigation of 445,000 well water samples in China, it has been estimated that 5.6 million people are exposed to high concentrations of arsenic in drinking water (>50 µg L −1 ) and that some 14.7 million are exposed to arsenic concentrations of >10 µg L −1 [14]. A similar problem has happened in the United States, The US geological survey reported that 1% of the 54,000 US public water supplies exceed 50 µg L −1 of arsenic concentrations, 3% exceed 20 µg L −1 , 8% exceed 10 µg L −1 , and 14% exceed 5 µg L −1 (Water Resources Investigations Report 99-4279) [15]. Many As(III) oxidation processes have been devised, such as direct photolysis [16,17] and homogeneous or heterogeneous metal-sulfite systems [2,[18][19][20]. However, the former is time-consuming while the latter introduces secondary pollution in the form of the metal in the metal-S(IV) systems. Inorganic As(III) is much more toxic than inorganic As(V) and accounts for about 20% of the arsenic present in the natural environment [21]. Therefore, it is very attractive to develop a method for oxidizing As(III) to As(V) quickly and safely. Zhang et al. [22] reported As(III) oxidation with SO 2 /O 2 and UV light, which was mainly accomplished through the action of oxysulfur radicals (especially SO 4 •− ). However, the reaction mechanism in UVC-S(IV)-O 2 system is still unclear with respect to the performance of two different S(IV) species (HSO 3 − /SO 3 2− ions) under UVC light.
In this work, As(III) oxidation by UVC-S(IV)-O 2 system has been investigated with the aim of revealing the different role/performances of HSO 3 − /SO 3 2− at pH 7 and 9.5, respectively. Laser flash photolysis has been utilized to obtain the quantum yields of photolysis of S(IV) species under 266 nm irradiation and the rate constant for the reaction between e aq − and S(IV).

Experimental Procedures
All photodegradation experiments were conducted in a 400 mL cylindrical reactor cooled by external jacket water circulation at a constant temperature of 25 • C ( Figure S1). A 254 nm lamp was used as the excitation source and was placed at the center of the reactor. A Na 2 SO 3 stock solution was freshly prepared prior to the reaction, using cooled boiling water to prevent the oxidation of sulfite. Predetermined amounts of As(III) were added to the reactor, and the solution was constantly stirred with a poly tetra fluoroethylene(PTFE)-coated magnetic stirrer. A pH meter (Mettler Toleod LE409) was used to measure the pH value before the reaction. After the pH and temperature of solution had stabilized, dilute H 2 SO 4 or NaOH was used to adjust the pH to the desired value. Each experiment was started by spiking a certain volume of fresh Na 2 SO 3 solution and immediately switching on the lamp. The pH was not controlled during the reaction. Aliquots (2 ml) of the solution were withdrawn at fixed time intervals, and then a specific amount of HCl (1:1) was added to terminate the reaction. The As(III) concentration were measured by liquid chromatography-hydride generation-atomic fluorescence spectrometry (LC−HG−AFS, Bohui Innovation Technology Co., Ltd., Beijing, China).

Analysis
Arsenic speciation was simultaneously analyzed by LC−HG−AFS. Phosphate buffer (45 mM, pH 5.6) was used as mobile phase to separate inorganic As(III) and As(V) on a Hamilton PRP-X100 anion-exchange column (Switzerland) in LC. Solutions of 5% HCl-2% KBH 4 were used for the determination of arsenic species concentration in HG-AFS. Argon (99.99%) was used as the carrier gas and shielding gas during the determination.
All laser flash photolysis (LFP) experiments were conducted in a 1 cm quartz cell in air-equilibrated or argon-saturated solutions at an initial pH of 7 or 9.5, at 298 K, under atmospheric pressure. Argon-saturated solutions were obtained by constantly bubbling argon through the sample. The LFP setup in the time-resolved experiments was based on an LS-2137U Nd:YAG laser (Lotis TII, Belarus) with an excitation wavelength of 266 nm, a pulse duration of 5-6 ns, an illumination spot area of 0.03 cm 2 and an energy per pulse of up to 10 mJ. The time resolution of the setup was ca. 50 ns. Solutions in LFP experiments were refreshed after every 100-150 pulses to maintain their degradation less than 15% during the measurements. Spectra of the sulfite solution were recorded on an Agilent 8453 spectrophotometer (Agilent Technologies) using a 1 cm cell. Figure 1 shows the efficiencies of As(III) oxidation in the UVC-S(IV)-O 2 system and the related control systems at pH 7 and 9.5, respectively. Whether in neutral or alkaline solution, As(III) alone under UVC irradiation showed no obvious oxidation in 10 min. In a dark experiment with sulfite, only about 13% of As(III) was oxidized at pH 7, but this amount was doubled at pH 9.5 (about 26%), suggesting that alkaline pH may activate sulfite in some way to oxidize As(III). In the UVC-S(IV) system, along with rapid S(IV) oxidation caused by UVC irradiation ( Figure S2), As(III) could also be oxidized to As(V) to some extent at pH 9.5 or 7. Indeed, we found that alkaline solution strongly facilitated As(III) oxidation from only 23% at pH 7 to 72% at pH 9.5 within 10 min. Three aspects could explain this marked difference. First, the critical oxidation-reduction potential (ORP) of As(V)/As(III) couples drop with decreasing pH value, such that the oxidation of As(III) to As(V) was more feasible in alkaline than in acidic solutions [23,24]. In addition, Hayon etc. [25] reported the pKa of HSO 3 − as 7.2, and so at pH 7 about 40% of sulfite should be present as SO 3 2− , increasing to almost 100% at pH 9.5. SO 3 2− has a better quantum yield under 254 nm irradiation compared to HSO 3 − (see Section 3.2 for details). Lastly, we noticed that in the experiment started at neutral pH, the solution became more acidic during the reaction time, whereas the alkaline pH was well maintained ( Figure S3), consistent with the results of As(III) oxidation at pH 7 or 9.5. High sulfite concentration (2 mM) could induce an anaerobic environment in solution within an extremely short time, since SO 5

As(III) Oxidation in the UVC-S(IV)-O 2 System
•− , SO 4 •− , HO• formation and self-oxidation of sulfites all consume oxygen [26,27]. Once the dissolved oxygen concentration dropped to a low level, reaction (3) could be a rate-controlling step in the chain reactions and hence influence the As(III) oxidation. We conducted pumping experiments with synthetic gas to prove the influence of oxygen on As(III) oxidation. As shown in Figure 1, when synthetic gas (21% O 2 /79% N 2 ) was constantly pumped into the reaction solution, As(III) oxidation efficiency was greatly enhanced at both pH 7 and 9.5, demonstrating that oxygen was indeed necessary for the chain reaction process and promoted the formation of radicals for the As(III) oxidation. Radical-scavenging experiments were employed to prove the existence of relevant radicals in the UVC-S(IV)-O 2 system ( Figure 2). As demonstrated above, oxygen was necessary for the chain reaction process and hence synthetic gas was constantly pumped into the reaction solution. Commonly, alcohols (EtOH and TBA) have been selected as probes of SO 4 •  [26]). In Figure 2, it can be seen that As(III) oxidation was not inhibited in the presence of TBA at pH 7, but the initial oxidation rate (r) decreased from 0.227 min −1 to 0.214 min −1 following addition of the same TBA concentration at pH 9.5. An alkaline solution could promote the HO• formation according to reactions (6) and (7). Besides, EtOH (5 mM or 177 mM) only partly inhibited As(III) oxidation, especially at pH 9.5 (23%-28% inhibition). Hence, other reactive species must be responsible for the As(III) oxidation. These results demonstrated that oxysulfur radicals generated in the UVC-sulfite system were the main reason for As(III) oxidation. Notably, about 18% and 26% of As(III) were still oxidized under anaerobic conditions at pH 7 and pH 9.5. This partial oxidation was clearly not due to SO 3 •− or e aq − because of the weak oxidation ability of the former and the reducing capacity of the latter. However, SO 3 •− could generate dithionate (S 2 O 6 2− ) according to reaction (8) [28] and then oxidize As(III). This is relevant because SO 3 •− accumulates in anaerobic environments, as proved in our previous work [26]. According to the results of radical-scavenging experiments and subsequent LFP experiments, a transformation process of S(IV) under UVC irradiation is proposed (Scheme 1), which includes three sections. Under anaerobic conditions, SO 3 •− and e aq − are the main products of S(IV) photolysis, which would be proved by LFP experiments. Under aerobic conditions, SO 3 •− forms reactive sulfur species (RSS) and e aq − forms reactive oxygen species (ROS), respectively. These reactive species showed enough ability to oxidize As(III). Hence, all three sections contribute to As(III) oxidation.

Scheme 1.
Proposed mechanism of action of the UVC/S(IV) system with/without oxygen.

LFP Studies of SO 3 •− and Hydrated Electrons
As shown in Figure S4 (Figure 3), which was mainly attributed to hydrated electrons [29]. The lifetimes of hydrated electrons (e aq − ) under these conditions were about 7-9 µs and did not depend on sulfite concentration, in agreement with literature estimates (k(e aq − + SO 3 2− ) < 1.5 × 10 6 M −1 s −1 [2]). The e aq − absorbance at 720 nm ( Figure S5) showed a good linear dependence on the excitation energy, which allowed estimation of the quantum yields of monophotonic ionizations of SO 3 2− (φ ion 266nm = 0.06) and HSO 3 − ions (φ ion 266nm = 0.01). From this, it could be concluded that HSO 3 − ions produced far fewer e aq − due to photoionization, in full agreement with the lower degradation efficiency of As(III) at pH 7 ( Figure 1).  [29]. Insert:kinetic curve at 720 nm with the first-order best fit with lifetime 8 µs.

Decay of e aq − in Aqueous Solution
Lowering the pH from 9.5 to 7.0 led to a decrease not only of photoionization quantum yield, but also lifetime of e aq − ( Figure 5). Moreover, the observed rate constant (k obs 720 nm ) of e aq − decay at pH 7 exhibited linear dependence on sulfite concentration ( Figure 5), indicating that this species was consumed by the reaction with HSO 3 − ions. Therefore, using the data of Figure 4 and the fact that at pH 7 about 60% of S(IV) was in the form of HSO 3 − ions, one could calculate the rate constant for e aq − quenching by HSO 3 − (k = 8 × 10 7 M −1 s −1 ) which is consistent with that in a previous literature report [28].

Conclusions
The UV-sulfite system has been successfully used to oxidize As(III). Photolysis of sulfite by 254 nm lamp irradiation induced the production of SO 3 •− and secondary SO 5 •− , SO 4 •− and HO•.
Oxysulfur radicals were responsible for As(III) oxidation at neutral or alkaline pH. Oxygen played a vital role in promoting As(III) oxidation. Through LFP experiments, we observed the signals of e aq − and SO 3 •− at 720 and 255 nm, respectively, providing evidence for sulfite photoionization. SO 3 2− ions (φ ion 266nm = 0.06) exhibited a much higher quantum yield of photoionization than HSO 3 − ions (φ ion 266nm = 0.01), indicating that alkaline pH was more favorable for application of the UV-S(IV)-O 2 system. The rate constant (k = 8 × 10 7 m −1 s −1 ) for reaction between e aq − and HSO 3 − has been measured.