UV185+254 nm photolysis of typical thiol collectors: decomposition efficiency, mineralization and formation of sulfur byproducts

The decomposition of toxic flotation reagents upon UV185+254 nm irradiation was attractive due to operational simplicity and no dosage of oxidants. In this work, the degradation of typical thiol collectors (potassium ethyl xanthate (PEX), sodium diethyl dithiocarbamate (SDD), O-isopropyl-N-ethyl thionocarbamate (IET) and dianilino dithiophoshoric acid (DDA)) was investigated by UV185+254 nm photolysis. The degradation efficiencies and mineralization extents of collectors were assessed. The formation of CS2 and H2S byproducts was studied, and the mechanisms of collector degradation were proposed under UV185+254 nm irradiation. The PEX, SDD and IET were decomposed with nearly 100% removal upon 75 min of UV185+254 nm irradiation. The decomposition rate constants decreased in the order SDD > PEX > IET ≫ DDA, and the DDA was the refractory collector. After 120 min of UV185+254 nm irradiation, 15−45% of carbon and 25−75% of sulfur of collectors were completely mineralized, and the mineralization extent decreased in the order PEX > SDD > IET > DDA. The percentage of gaseous sulfur (CS2 and H2S) ranged from 0.48 to 4.85% for four collectors, showing the fraction of emitted sulfur byproducts was small. The aqueous CS2 concentration increased in the first 10−20 min, and was decreased to a low level of 0.05–0.1 mg l−1 at 120 min. Two mechanisms, i.e. direct UV254 nm photolysis and indirect oxidation with free radicals, were responsible for collector decomposition in the UV185+254 nm photolysis.


Introduction
Froth flotation has become the most widely applied process for separating valuable minerals from ores in mines around the world. Thiol collectors, such as xanthates, dithiophosphates and dithiocarbamates, are important flotation reagents to render sulfide minerals hydrophobic and facilitate bubble attachments [1,2]. To achieve high recovery of non-ferrous metals, the dosage of collectors is frequently ranged from 30 to 300 g ton 21 (ore). Therefore, the consumption of thiol collectors becomes very large due to the extremely high amount of treated ores. Even in the 1980s, the global consumption of xanthates had reached above 52 000 tons per year [2]. Nevertheless, nearly 50% of collectors dosed in flotation circuits would be discharged in wastewaters after the mineral flotation [3]. Some collectors and their byproducts are found to be toxic to soil microbes, biota, animals and human beings [4][5][6][7]. Previously, the hazards of xanthate to frog embryos have been reviewed [7]. Accordingly, the discharge of collectors from mineral flotation may cause serious environmental pollution. Therefore, it has been of great concern to remove potentially toxic collectors from flotation wastewaters for the sustainable development of mining industry.
Recently, some processes have been developed to remove organic reagents from flotation wastewaters, including adsorption [8], chemical oxidation [9,10], ozonation [11,12] and biodegradation [13,14]. Chemical oxidation and ozonation have exhibited high efficiency in decomposing flotation reagents. But these processes need the dosage of chemical oxidants, such as sodium hypochlorite [9], persulfate [10], hydrogen peroxide [15] and ozone [11]. The usage of highly reactive oxidants in mines makes them less attractive and sometimes expensive. Since most of the mines are located far from industrial zones, there are potential risks associated with long distance transportation and storage of dangerous chemicals. The biodegradation is widely accepted as a low cost process in wastewater treatment. However, previous reports indicate that some flotation reagents are toxic to microbes, considerably reducing microbial activities in the biodegradation of flotation reagents [13,16].
Advanced oxidation processes (AOPs) involving ozone [10], Fenton's reagent [17], hydrogen peroxide [15], persulfate [18,19] and photocatalyst [20][21][22] have been investigated to effectively degrade flotation collectors, pharmaceuticals and textile dyes. However, most of these studies only concerned the removal efficiencies of xanthates [11,12,15]. Up to now, research on the removal of other thiol collectors such as dithiophosphates and dithiocarbamates has been scarce. Because thiol collectors have different molecular structures [1,23], their decomposition behaviours by the same AOP method may differ. Additionally, the functional groups of all thiol collectors have sulfur atoms [1,23]. So, it is possible to generate toxic sulfur byproducts such as CS 2 and H 2 S, while organic sulfur of collectors is mineralized to sulfate. For example, the concentrations of aqueous CS 2 were determined in the ozonation of xanthates [24]. However, the generation of sulfur byproducts receives little concern in the decomposition of thiol collectors by the AOPs [10,12].
The UV-based AOPs, such as UV 254 nm /O 3 and UV 254 nm /TiO 2 photocatalysis, can effectively decompose flotation collectors and subsequently mineralize various byproducts [11,20]. But these AOPs involve the dosage of oxidants or powder catalysts, making them expensive and complex. The UV 185þ254 nm photolysis, irradiated by a low-pressure Hg lamp emitting both vacuum-UV (VUV) at 185 nm and UVC at 254 nm, is a simple and promising AOP. The 185 nm VUV irradiation of water can directly generate hydroxyl radicals (OH †), hydrogen radicals (H †), solvated free electrons (e aq 2 ), superoxide radicals (HO 2 †, O 2 †2 ) and H 2 O 2 as shown in equations (1.1)2(1.6) [25]. The formation of OH † radicals and superoxide anion (O 2 †2 ) in the VUV irradiation of water have been proven by the electron paramagnetic resonance spectroscopy [26]. The UV 185þ254 nm photolysis has shown high oxidation capacity for pollutants, such as odour compounds [27], pharmaceuticals [28] and pesticides [29]. Moreover, the operation of UV 185þ254 nm photolysis is much simpler than the UV 254 nm /O 3 and UV 254 nm /TiO 2 photocatalysis because no oxidant or catalyst is required. In terms of the practical treatment of flotation wastewaters, the transportation and storage of dangerous oxidants in mines can also be avoided if the UV 185þ254 nm photolysis is applied. Thus, it is quite necessary to assess the removal performances of flotation reagents by the UV 185þ254 nm photolysis. However, as far as we know, there is no report on the UV 185þ254 nm photolysis of organic flotation reagents In this work, the UV 185þ254 nm photolysis of thiol collectors is investigated by using a 40 W low-pressure Hg lamp, which emits about 10% radiation at 185 nm and 90% radiation at 254 nm. Four thiol collectors, potassium ethyl xanthate (PEX), sodium diethyl dithiocarbamate (SDD), Oisopropyl-N-ethyl thionocarbamate (IET) and dianilino dithiophoshoric acid (DDA), are selected as typical sulfide mineral collectors. The objectives of this work are (i) to assess the feasibility of decomposing thiol collectors by the UV 185þ254 nm photolysis, (ii) to determine the generation of CS 2 and H 2 S byproducts, as well as (iii) to propose the collector decomposition mechanisms under UV 185þ254 nm irradiation. The mineralization of collectors is examined by measuring total organic carbon (TOC) and SO 4 22 concentrations. The amounts of gaseous CS 2 and H 2 S, as well as the concentration of aqueous CS 2 , were measured to study the generation of sulfur byproducts.

Experimental set-up and degradation procedures
All degradation experiments were conducted with a batch mode in a jacket glass reactor connected to a thermostatic bath. The schematic diagram of the experimental set-up is shown in figure 1. The cylindrical reactor, with 1240 mm height and 53 mm internal diameter, was installed at its axis with a quartz tube with a height of 1220 mm and outer diameter of 21 mm. The UV 185þ254 nm or UV 254 nm lamp was inserted into the quartz tube. The optical distance from tube surface to the internal surface of the reactor was 16 mm. The air was continuously purged into the reactor through a porous glass plate with a flow rate of 1.67 l min 21 . The injected air stream not only provided dissolved oxygen for degradation reactions, but also stirred the collector solutions. The degradation experiments were conducted at 25 + 28C. Prior to photolysis experiments, the collector (0.2 g) was dissolved in 2 l deionized water to prepare a PEX (SDD, IET or DDA) solution of 100 mg l 21 concentration. The initial pH was adjusted to 7.0212.0 using 0.05 mol l 21 NaOH or HCl solution. When the collector solution was introduced into the reactor purged with an air stream, the UV 185þ254 nm or UV 254 nm lamp was turned on to carry out photolysis experiments. The irradiation time was controlled to be 120 min. Because CS 2 , H 2 S and other sulfur byproducts were emitted into gas phase, the emission gas was introduced into an activated carbon column to remove toxic byproducts. The aqueous samples were taken at designated intervals to determine the concentrations of collector, chemical oxygen demand (COD), TOC, CS 2 and SO 4 22 ions.

Analysis and calculation 2.3.1. Determination of collector concentration, COD and TOC
The concentration of thiol collectors (PEX, SDD, IET and DDA) was determined by a UV-vis spectroscopic method [10,11,13,30]. The absorbance of collector solution was recorded by a UV -vis spectrophotometer (UV-5500PC, Shanghai Metash Instruments Co. Ltd, China). The COD was determined by the standard dichromate method (HJ/T 399-2007). The TOC of water samples was measured by using a Shimadzu TOC-V organic carbon analyzer. The carbon mineralization extent (g c ) of the collector (PEX, SDD, IET and DDA) was calculated as the below equation where TOC 0 and TOC t (mg l 21 ) were the TOC concentration at initial and time t, respectively. The aqueous CS 2 concentration was measured by a diethylamine cupric acetate spectrophotometric method (GB/T 15504-1995). Fifty millilitres of water sample were continuously purged by a 100 ml min 21 N 2 stream for 1 h to volatilize CS 2 , which was absorbed by a diethylamine and cupric acetate mixed liquid. Then the absorbance of absorption liquid was measured at 430 nm using dehydrated alcohol as a reference solution. The concentration of sulfate was determined by a barium chromate spectrophotometry method (HJ/T 342-2007). Because SO 4 22 ions, with the highest valence of sulfur, were the final oxidization product of organic sulfur in collectors, the sulfur mineralization extent (g s ) of the collector was defined as the below equation [31] where M and n were the molecular weight and number of sulfur atom in the collector (PEX, SDD, IET or DDA), respectively, C SO 2À 4 ,t (mg l 21 ) was the concentration of SO 4 22 ions at time t, and C 0 (mg l 21 ) was the initial concentration of the collector.

Measurement of the amount of gaseous CS 2 and H 2 S
The amount of gaseous CS 2 emitted from the reactor for 120 min was measured by a diethylamine spectrophotometric method (GB/T 14680-93). By adding 5.0 mg cupric acetate, 2.5 ml diethylamine and 2.5 ml triethanolamine into a 500 ml volumetric flask, the CS 2 absorption liquid was prepared by diluting the mixture with dehydrated alcohol to 500 ml. As shown in figure 1, before the emission gas was introduced into CS 2 absorption liquid, it was first passed through a glass tube filled by absorbent cotton coated with lead acetate to remove H 2 S. The amount of gaseous H 2 S for 120 min was measured by a methylene blue spectrophotometric method (GB/T 11742-89). By adding 4.3 g cadmium sulfate, 0.30 g NaOH and 10.0 g polyvinyl alcohol ammonium phosphate into a 1000 ml volumetric flask, the H 2 S absorption liquid was prepared by diluting the mixture with deionized water to 1000 ml.

Effect of UV wavelength
The photolysis of thiol collectors (PEX, SDD, IET and DDA) upon UV 254 nm or UV 185þ254 nm irradiation is shown in figure 2. The removal efficiency of collectors and decomposition rate constants are summarized in table 2. In this work, overall kinetics of the collector degradation can be described by a simple pseudofirst-order rate law [32]. As shown in figure 2a and   IET and DDA upon UV 254 nm irradiation. In UV-irradiated solutions, the decomposition of pollutants is frequently initiated by excited molecules with absorbing UV irradiation [33]. Among four collectors, the PEX and SDD may have higher mole absorptivity of UV 254 nm irradiation than IET and DDA, resulting in higher removal efficiencies. In most mines, residual flotation reagents are usually removed in tailing ponds by a natural degradation process involving sunlight and oxygen. As given in table 2, the halflife of IET and DDA was almost one order of magnitude larger than that of PEX and SDD. Therefore, it can reasonably be inferred that after the natural degradation, residual IET and DDA in tailing ponds would have higher concentrations than the PEX and SDD. As exhibited in figure 2b and table 2, the removal efficiency of PEX and SDD had reached above 96% even at 20 min under UV 185þ254 nm irradiation. At 75 min, nearly 100% removal of PEX, SDD and IET was achieved, revealing that most thiol collectors could be effectively degraded by the UV 185þ254 nm photolysis without adding any oxidant. The decomposition rate constant (k collector ) of collectors upon UV 185þ254 nm irradiation decreased in the order k SDD (0.2005 min 21 ) . k PEX (0.1565 min 21 ) . k IET (0.04485 min 21 ) ) k DDA (0.00701 min 21 ). As listed in table 2, the k collector values for four collectors under UV 185þ254 nm irradiation were 1.60210.03 times higher than those achieved upon UV 254 nm irradiation, showing the enhanced degradation of thiol collectors by the UV 185þ254 nm photolysis. However, the DDA, with very low k collector value, can be considered as a refractory collector for the UV 185þ254 nm photolysis.

Effect of initial pH
The flotation of sulfide minerals is usually conducted in alkaline pulps. Thus, residual collectors are present in the flotation wastewaters at pH range from neutral to alkaline [34]. In this work, the effect of initial pH on the UV 185þ254 nm photolysis of PEX and DDA was investigated. As shown in figure 3, both the k PEX and k DDA values gradually decreased as initial pH rose. The decomposition rate constants obtained at pH 7.0 were 2.4 times for PEX and 4.9 folds for DDA higher than those achieved at pH 12.0, respectively. It suggests that the neutral pH can facilitate the decomposition of flotation collectors upon UV 185þ254 nm irradiation. In the previous studies, the higher decomposition efficiencies were also achieved at lower pH during the UV 185þ254 nm photolysis of 1,4-dioxane [30] and 4-tert-octylphenol [35].
Upon 185 nm VUV irradiation, the homolysis and photochemical ionization of H 2 O occur with the generation of OH † radicals (equations (1.1) and (1.2)), and OH † radicals (or HO 2 †) can recombine to form H 2 O 2 (equations (1.3) and (1.6)) [25]. Under UV 185þ254 nm irradiation, H 2 O 2 can be decomposed by absorbing UV 254 nm light to form OH † radicals (equation (3.1)). Accordingly, OH † radicals and H 2 O 2 are in equilibrium under UV irradiation offered by the UV 185þ254 nm lamp. On the other hand, the concentration of H 2 O 2 is also dependent on the pH value of the reaction system. As given in equation (3.2), H 2 O 2 itself is in equilibrium with OH 2 anions [36]. Additionally, HO 2 † radicals are also in equilibrium with protons ((equation (1.5)). According to equations (1.5), (1.6) and (3.2), it can be seen that high H þ concentration (i.e. low pH) promotes the equilibrium to form H 2 O 2 . In turn, Table 2. The removal efficiencies of collectors, decomposition rate constants (k collector ), half-lives (t 1/2 ) and correlation coefficients (R 2 ) in the degradation of thiol collectors under UV 185þ254 nm and UV 254 nm irradiation, respectively. In addition, the equilibrium of carbonate may also influence the pH dependence of the collector decomposition. The mineralization of contaminants by UV-based AOPs results in the formation of CO 2 , subsequently increasing the aqueous concentration of CO 3 22 anions. According to the equilibrium of CO 3 2 in water, HCO 3 2 and CO 3 22 anions dominate the equilibrium approximately at neutral to weak basic pH (6 , pH , 10) and at strong basic pH (pH . 10), respectively. These carbonate species can scavenge OH † radicals. However, CO 3 22 anions have a larger scavenging capacity than HCO 3 2 [37].
Thus, at higher initial pH, more OH † radicals would be scavenged by CO 3 22 anions, resulting in lower decomposition rate constants of PEX and DDA.

The mineralization of thiol collectors
To investigate the mineralization of thiol collectors under UV 185þ254 nm irradiation, the concentrations of COD, TOC and SO 4 22 ions were measured with the results shown in figure 4. For each collector, the relative concentrations of COD and TOC declined, and the SO 4 22 concentration increased as the collector was decomposed. It suggested that the byproducts had been further degraded with the formation of CO 2 and SO 4 22 ions under UV 185þ254 nm irradiation. As indicated in figure 4, the TOC was reduced just a bit while the evident decrease in COD was observed for each collector. Thus, the removal efficiency of COD was higher than the mineralization extent of carbon as given in table 3. By comparing the decomposition rate constants for collector and COD removal, as summarized in tables 2 and 3, the k collecor for each collector was 3221 folds higher than the k COD value. It clearly indicated that only a small fraction of organic carbon in collectors was mineralized to CO 2 upon UV 185þ254 nm irradiation although nearly 100% removal of PEX, SDD and IET was achieved. For four collectors, the mineralization extent of carbon (g c ) at 120 min increased in the order DDA , IET , SDD , PEX. In particular, the g c for the PEX was much higher than that for SDD, IET and DDA, indicating that the byproducts derived from PEX seemed to be readily decomposed by OH † radicals. As shown in figure 4, the concentration of SO 4 22 ions increased up to 90.36 mg l 21 for PEX, 64.79 mg l 21 for SDD, 26.90 mg l 21 for IET and 11.48 mg l 21 for DDA, respectively, under UV 185þ254 nm irradiation for 120 min. The SO 4 22 ions were the final sulfur product in the oxidation of organic sulfur in thiol collectors [11,12,24]. The occurrence of SO 4 22 ions with increased concentrations well demonstrated that organic sulfur of collectors could be completely mineralized [24]. As summarized in  16.75% for DDA, respectively. Thiol collectors except for DDA had much higher g s values than the g c , suggesting that the conversion of organic sulfur to SO 4 22 was more efficient than the mineralization of organic carbon to CO 2 . By considering g c and g s together, the mineralization extent of thiol collectors decreased in the order PEX . SDD . IET . DDA upon UV 185þ254 nm irradiation. In particular, the PEX and SDD had much larger extent of mineralization when compared with IET and DDA at the same degradation conditions.    For example, CS 2 is considered as a hazard air pollutant under the Title III of the 1990 Clean Air Act Amendment (CAAA) of the USA [42]. Accordingly, while toxic sulfur byproducts are emitted from air purged flotation wastewaters into the atmosphere or indoor environment, it may pose a significant hazard to safety, health and the environment (SHE). However, to the best of our knowledge, no work is available in the literature on the quantitative determination of emitted CS 2 and H 2 S in gas phase when thiol collectors are decomposed by the AOPs. For the aqueous CS 2 byproduct, Fu et al. [11,24] had measured the CS 2 concentration while degrading xanthates by the O 3 and UV/O 3 processes. However, the evolutions of aqueous CS 2 byproduct has not yet been investigated during the degradation of thiol collectors, except for xanthates by the AOPs. In this work, the amounts of gaseous CS 2 and H 2 S emitted from air purged solutions were measured using the CS 2 and H 2 S absorption liquids, respectively. As illustrated in figure 5, upon UV 185þ254 nm irradiation of 100 mg l 21 collector solutions for 120 min, the amount of gaseous CS 2 reached 1.405 mg for PEX, 3.784 mg for SDD, 0.154 mg for IET and 0.202 mg for DDA, respectively. The results indicated that the SDD and PEX released much larger amount of gaseous CS 2 than IET and DDA. For the emission of H 2 S byproduct, gaseous H 2 S reached 0.0836 mg for PEX, 0.468 mg for SDD, 0.101 mg for IET and 0.0502 mg for DDA, respectively. Among four collectors, the SDD had released the largest amount of H 2 S into gas phase. For each collector, the amount of gaseous CS 2 released was larger than that of H 2 S, and the trends became more evident for PEX and SDD. By considering CS 2 and H 2 S together, the amount of gaseous sulfur byproducts released upon UV 185þ254 nm irradiation increased in the order IET % DDA , PEX , SDD. According to these results, it could be seen that the release of gaseous sulfur byproducts was quite diverse due to their different molecular structures and sulfur contents of thiol collectors.

Generation of CS 2 and H 2 S byproducts
As mentioned above, the emitted CS 2 and H 2 S are toxic gases. Thus, it is essential to indicate the percentage of gaseous sulfur to total sulfur in collectors. By assuming that no volatile sulfur byproducts except for CS 2 and H 2 S are present in emission gas, the percentage of gaseous sulfur (b s,g ) can be defined according to the below equation where m 1 and m 2 (mg) were the amount of emitted CS 2 and H 2 S into gas phase for 120 min, n was the number of S atom in collectors and V (l) was the volume of collector solutions. Thus, the b s,g values for four collectors are summarized in table 4. Except for the SDD, the b s,g for PEX, IET and DDA was very low, showing that only a small fraction of sulfur in thiol collector was released into emission gas in the UV 185þ254 nm photolysis. However, a previous study had estimated that 20.6% of total sulfur in n-butyl xanthate was released into gas phase after the oxidation by O 3 [12]. These data were achieved by subtracting total sulfur in n-butyl xanthate with sulfur in SO 4 22 ions. Since other sulfur byproducts such as sulfite, thiosulfate as well as organic sulfur compounds would be also presented in aqueous solutions, 20.6% of sulfur released in gas phase might be overestimated by Yan et al. [12].
In this work, aqueous CS 2 concentrations were also measured. As presented in figure 6, the CS 2 concentration for each collector rapidly increased to its maximum value in the first 10220 min, and then gradually decreased to a low level under UV 185þ254 nm irradiation. The measured maximum concentration of CS 2 was 0.37 mg l 21 for PEX, 1.88 mg l 21 for SDD, 0.18 mg l 21 for IET and 0.15 mg l 21 for DDA, respectively. It can be clearly seen that the decomposition of SDD can generate much more CS 2 in comparison to other collectors. Additionally, the order of maximum CS 2 concentration for four collectors was well consistent with that for the amount of gaseous CS 2 as shown in figure 5. As shown in figure 6, residual CS 2 concentrations for all collectors were decreased to a low level of 0.0520.1 mg l 21 after 120 min of UV 185þ254 nm irradiation. By considering the remarkable increase in SO 4 22 concentrations shown in figure 4, it can be reasonably inferred that most of CS 2 byproduct was converted to SO 4 22 ions by OH † radicals with the pathways as elucidated in our previous work [24]. As discussed above, the decomposition of SDD and PEX had generated a larger amount of CS 2 in comparison to IET and DDA. As summarized in table 1, both the sulfur content and molecular structure of thiol collectors are quite different from each other. Accordingly, the different sulfur contents and decomposition mechanisms caused by different molecular structures would be associated with the formation of sulfur byproducts under UV 185þ254 nm irradiation. The sulfur contents of PEX, SDD, IET and DDA are 40.25%, 37.42%, 21.77% and 22.85%, respectively. The high sulfur contents of PEX and SDD may well elucidate their large amount of generated CS 2 byproduct.
Additionally, the generation mechanisms of CS 2 from thiol collectors might also significantly influence its amount. However, up to now, the reports on the decomposition mechanisms of thiol collectors by the AOPs are scare. For the ozonation of xanthates, it is suggested that the attack on the C2O bond of xanthates by OH † radicals and nucleophilic reactions occurring for the C¼S bond of the 2CSS 2 group could result in the generation of the CS 2 [24,39]. In this case, the C2O bond in the PEX and C2N bond in the SDD have high electron density, which may be preferentially attacked by OH † radicals. Thus, the CS 2 may be generated from the PEX and SDD under UV 185þ254 nm irradiation according to equations (3.4) and (3.5), respectively. As presented in table 1, the IET and DDA had molecular structures that were different to the PEX and SDD. The pathways of CS 2 generated from IET and DDA may be different from that for the PEX and SDD. In our future work, attention should be paid to elucidating the generation pathways of sulfur byproducts  3.5. The mechanisms of UV 185þ254 nm photolysis of thiol collectors As shown in figure 2 and table 2, thiol collectors could be degraded by both UV 254 nm and UV 185þ254 nm photolysis, but with higher degradation rate constants under UV 185þ254 nm irradiation. As mentioned above, free radicals such as OH † can be effectively generated in air purged water under 185 nm VUV irradiation as given in equations (1.1)2(1.6). OH † radicals, with an oxidation-reduction standard potential of 2.80 V, are non-selective and vigorous oxidants. The rate constants for OH † reacting with most organic compounds are within the range of 10 6 210 9 l/(mol s) [43]. Therefore, under UV 185þ254 nm irradiation, two degradation mechanisms as shown in figure 7, i.e. direct UV 254 nm photolysis and indirect oxidation with free radicals such as OH †, should be responsible for the decomposition of collectors [44]. Nevertheless, the contributions of UV 254 nm photolysis and indirect oxidation with free radicals to collector decomposition were quite dependent on molecular structures of thiol collectors. For the PEX, SDD and DDA, both UV 254 nm photolysis and indirect oxidation with free radicals had contributed greatly as the k collector values under UV 185þ254 nm irradiation were 1.6022.63 times higher than those under UV 254 nm irradiation. But for the IET, the k IET for UV 185þ254 nm photolysis was 10.03 times higher than that for UV 254 nm photolysis. This suggested that indirect oxidation with free radicals was the main mechanism for the UV 185þ254 nm photolysis of IET.

Conclusion
Thiol collectors (PEX, SDD, IET and DDA) could be effectively degraded by the UV 185þ254 nm photolysis without dosing any oxidant. The removal efficiencies of PEX, SDD and IET reached nearly 100% upon 75 min of UV 185þ254 nm irradiation. The k collector for four collectors decreased in the order k SDD . k PEX . k IET ) k DDA . The DDA was the typical refractory flotation collector for UV 185þ254 nm photolysis. In the UV 185þ254 nm photolysis of the PEX and DDA, the k collector values were decreased at high initial pH, indicating neutral pH of flotation wastewaters can facilitate the collector decomposition. After 120 min of UV 185þ254 nm irradiation, the g c and g s for four collectors reached 15245% and 25275%, respectively, with the mineralization extent of PEX . SDD . IET . DDA. The effective degradation of thiol collectors was attained under UV 254 nm irradiation alone. Thus, two mechanisms, i.e. direct UV 254 nm photolysis and indirect oxidation with free radicals such as OH †, were responsible for the decomposition of collectors by the UV 185þ254 nm photolysis.
After UV 185þ254 nm irradiation for 120 min, the percentage of gaseous sulfur was 1.57% for PEX, 4.85% for SDD, 0.53% for IET and 0.48% for DDA, respectively, indicating that only a small fraction of sulfur in collectors was released into emission gas. For each collector, the amount of emitted CS 2 in gas phase was larger than that of H 2 S. The aqueous CS 2 concentration increased rapidly in the first 10220 min for each collector, and was then reduced to a low level of 0.0520.1 mg l 21 at 120 min under UV 185þ254 nm irradiation.
Data accessibility. The datasets supporting this article have been uploaded as part of the electronic supplementary material.
Authors' contributions. G.L. and X.W. set up the degradation system. G.L. and X.L. carried out the UV photolysis of thiol collectors. B.L. and X.W. measured the aqueous concentration of CS 2 and amount of emitted CS 2 and H 2 S in gas phase. P.F., G.L. and X.L. contributed to analysis data and wrote the draft of the manuscript. Finally, P.F. revised the manuscript.