Advanced oxidation processes for the removal of [bmim][Sal] third generation ionic liquids: effect of water matrices and intermediates identification

Unique properties of ionic liquids make them green alternatives for conventional volatile organic compounds. Due to increased production and the high stability of these substances, they could be classified as persistent pollutants and could break through classical treatment systems into natural waters. A preliminary ionic liquid hydrolysis study demonstrated a pH dependent degradation profile with a significant decrease in hydrolysis efficiency as pH lowered from 10.0 to 2.8. In order to examine future prospects for ionic liquid removal, different advanced oxidation processes (TiO2 Degussa P25/ H2O2, TiO2 Degussa P25, 7.2Fe/TiO2/H2O2, and H2O2) were studied for their applicability in the degradation of imidazolium-based ionic liquids in aqueous solution. These processes were conducted in the dark as well as in the presence of UVA and simulated sunlight (SS) radiation. Among the investigated dark processes, the 7.2Fe/TiO2/H2O2 system showed the highest efficiency, which can be attributed to a dark heterogeneous Fenton process. Otherwise, the most efficient among all the studied degradation processes was the UVA/TiO2 Degussa P25/H2O2 process. In order to make degradation processes more similar to that of the practical process SS radiation was used. Among studied processes, the 7.2Fe/TiO2/H2O2 system showed the greatest potential for the removal of ionic liquids. Also, it was observed that the impact of anions on the cation degradation efficiency was much more pronounced. Due to the possible fate of ionic liquids in the environment, for five different waters (pond, rain, tap, river, and condensate) degradations in the dark and under simulated sunlight were studied. For all processes, and all water types in the presence of SS radiation a remarkable positive effect of naturally dissolved organic matter on the degradation efficiency was observed. Also, in all experiments, the anion was less stable than the cation. The major photodegradation products identified using liquid chromatography-mass spectrometry (HPLC-MS/MS) techniques were hydroxylated compounds.


Introduction
Room-temperature ionic liquids are receiving considerable attention as green, high-tech reaction media of the future. 1 This can be explained by the fact that ionic liquids possess negligible vapor pressure, so that they can be a good alternative to the emissions of toxic vapors from conventional molecular organic solvents. 2 In addition to these features, ionic liquids have other good properties, such as good thermal stability, a wide electrochemical potential window, high electric conductivity and miscibility with water or organic solvents which make them interesting for application in different elds. 3,4 Up to now the ionic liquids have been used in lithium ion batteries, 5 gas separation, compression and handling, 6,7 metal plating, 5,8,9 dyesensitized solar cells, and as paint or polymer additives. 10,11 Furthermore, the third generation of ionic liquids includes salts with a melting point below 100 C that contain biologically active cations and/or anions in the structure. 12,13 A special attention of the scientic community in the recent years has been paid to ionic liquids with salicylate anion, since they show possibility of a wide range of applications in the eld of pharmaceutical industry, 14 the synthesis of nanoparticles, 15 heat storage, 16 electrodeposition of transition metals 17 etc. Complexing (chelating) ability of salicylate anions was used for the synthesis of the hydrophobic ionic liquids for selective metal (metalloid) extraction from the industrial and the communal wastewaters. [18][19][20][21] On the other hand a broad implementation of ionic liquids leads us to the following question: what will be the outcome when these substances get through the environment? It can be assumed that their low vapor pressure provides air protection from the emission of toxic vapors, but on the other hand their stability makes them hazardous to water and soil. 22 Due to its high stability in the near future it can be expected the presence of signicant amounts of ionic liquids in technological wastewaters and their break through classical treatment systems into natural waters. 1 Therefore, information concerning their fate in the aquatic environment is very important.
Also, some studies showed that ionic liquids are more toxic to the cells in comparison to conventional solvents. 1 These studies are in good agreement with other studies in which the most common imidazolium based compounds showed high resistance toward microbial degradation. 22 It can be also assumed that among the various transformation processes (biotic and abiotic), photostability is an important factor inuencing the fate of ionic liquids in the environment. Thus, it is very important to nd an efficient method for a complete and safe removal of ionic liquids from the environment. [23][24][25] The advanced oxidation processes (AOPs), for example, UVA/H 2 O 2 , UVA/Fe 2+ /H 2 O 2 , and UVA/TiO 2 are potentially useful for wastewater treating because they generate hydroxyl radicals, 26-28 a powerful nonspecic oxidant, and can provide an almost total degradation. 29 Among numerous AOPs, Fenton's process is one of the promising technologies because of the high oxidation potential and relatively low cost. 30 This process can be enhanced by visible light due to the decomposition of the photo-active Fe(OH) 2+ species producing additional OH-radicals in solution. 31 In these processes participation of iron oxide minerals (hematite a-Fe 2 O 3 , goethite a-FeOOH, magnetite Fe 3 O 4 and ferrihydrite) as the catalysts in the presence of hydrogen peroxide was also effective in the past, because uniform rates of hydroxyl radical generation at high concentration of hydrogen peroxide used. Moreover a-FeOOH/H 2 O 2 system has a number of advantages as compared to a homogenous Fenton's system with respect to reaction pH range and the removal of iron. 32 In this study efficiency of ionic liquid degradation was investigated and compared by different AOPs, including direct and H 2 O 2 indirect photolysis, as well as heterogeneous photocatalysis using TiO 2 and 7.2Fe/TiO 2 catalysts in the presence of UVA/simulated sunlight (SS) radiation. In contrast to previously studied ionic liquids that contain an inorganic anion i.e. chloride 1,22,[33][34][35] or bromide, 36 in this paper an ionic liquid in which both components were organic ions was investigated. Also, mutual inuence of anion and cation was studied in detail. For broaden the application of used AOPs, the photocatalytic degradation of different kind of ionic liquids were also studied. Because there is no information in the available literature on the photodegradation of ionic liquid under natural aquatic conditions, the present study deals with the inuence of water matrices on the ionic liquid photodegradation efficiency. In this study ve different types of waters was used HPLC-MS/MS technique was used to track the 23 intermediate products formed during the course of ionic liquid degradation.

Degradation experiments
All experiments were carried out in a batch reactor made of Pyrex glass (total volume of cca 100 mL, solution depth 46 mm). The UVA radiation was provided from a 125 W high-pressure mercury lamp (l > 290 nm, Philips HPL-N, Netherlands, emission bands in UV region at 304, 314, 335, and 366 nm, with the maximum emission at 366 nm). Intensity of UVA radiation was 3.57 Â 10 À3 W cm À2 . Irradiation with SS light was performed using a set of four 35 W halogen lamps (l > 300 nm, VITO, MR 16, China), placed in a symmetrical way around the reactor. UVA and vis radiation intensities for applied SS light were 1.8 Â 10 À4 W cm À2 and 116.2 Â 10 À3 W cm À2 , respectively.
A solution of 30 mL [bmim][Sal] containing 1.67 g L À1 of catalyst and/or 0.45 mmol L À1 H 2 O 2 with successive addition of H 2 O 2 (70 mL each hour during the degradation) was used (except for the study of hydrolysis and direct photolysis). All experiments were performed at pH 2.8 (except in the case of hydrolysis study). A value of 2.8 was chosen because it was optimal value for 7.2Fe/TiO 2 /H 2 O 2 system. 27 The H 2 SO 4 and NaOH solutions were used to adjust the pH of reaction solutions. The lamps were turned on for 15 min prior to degradation to obtain a constant light intensity output. At the same time reaction solution was stirred continuously in the dark in order to ensure equilibrium adsorption of [bmim][Sal] on the photocatalyst.
To ensure a constant O 2 the reaction solution was continuously purged to the bottom of the reactor with molecular oxygen (99.99% purity) at a constant ow rate of 5 mL min À1 . Furthermore the solution was homogenized with the aid of a stirring bar, to ensure completely mixed batch conditions. The temperature of the reaction solution was kept at 25 AE 0.5 C throughout the experiment by water circulation system. Dark experiments were conducted under identical conditions but without radiation.
The matrix effects of ve water types on the efficiency of ionic liquid degradation were evaluated. For this purpose we used the following samples: river water, rain water, pond water, tap water, and condensate from air conditioner. The river water sample was collected from the Danube (Novi Sad, Serbia) in July 2014. Of the same year in the urban area of Novi Sad, rain water and pond water samples were collected during the rainy periods between the ninth and fourteenth of July. In the same month tap water sample was collected from public drinking water supply system of Bački Jarak (Serbia) while the condensate water from the air conditioner harvested in one at that is also located at the territory of Novi Sad. To avoid microbial growth, water samples were sterilized by ltration and all glassware was sterilized by autoclaving for 30 min at 140 C.

Analytical methods
For the kinetic studies of ionic liquid removal, 0.5 mL of the reaction mixture were sampled at the beginning of the experiment and at regular time intervals, followed by ltration through Millipore (Millex-GV, 0.22 mm) membrane lter (in the presence of catalyst). The lack of ionic liquid adsorption onto the lters was conrmed by a preliminary test. The HPLC-DAD analyses were performed on a Shimadzu 20A series ultra-fast liquid chromatograph (UFLC, Shimadzu Cooperation, Kyoto, Japan), equipped with UV-vis diode array detection and polar CN column (Inertsil CN-3 column; 150 mm Â 4.6 mm i.d., particle size 3 mm). The injection volume was 20 mL. Column temperature was held at 25 C, mobile phase was a mixture of 0.1% H 3 PO 4 -ACN (6 : 4, v/v), pH 2.56, and the total ow rate was 0.8 mL min À1 . Ionic liquid elution was monitored at 210 nm. The retention time of ionic liquid's cation was 1.44 min while the same for the anion was 6.58 min.
For collected water samples physicochemical characteristics are given in Table 1.
Dissolved organic carbon (DOC) of different water samples (ltered through 0.22 micrometer lter) was determined on an Elementar Liqui TOC II (Germany) in accordance with Standard US EPA Method 9060A. Before measurements samples were acidied.
The radiation uxes were measured using a Delta Ohm HD 2102.2 (Padova, Italy) radiometer which was tted with the LP 471 UVA (spectral range 315-400 nm) and LP 471 RAD (spectral range 400-1050 nm) sensors.
The conductivity measurements were performed using a conductivity meter type Cond 3210 (WTW, Weilheim, Germany) equipped with a standard conductivity measuring cell TetraCon 325.
The pH measurements were made using a Hanna Instruments combined glass electrode (Kehl, Gemany), on a previously calibrated pH-meter (Iskra, Kranj, SFRJ).
In addition, in order to gain insight into the relationship between the elements present in the different water matrices and the efficiency of ionic liquid degradation, the signal intensities obtained by scanning the waters by inductively coupled plasma mass spectrometry (ICP-MS) are given in Table  S1, ESI. † The ICP-MS analysis was carried out on an Agilent 7700e ICP-MS (Hachioji, Japan), with a collision cell operated in helium mode. General purpose plasma conditions were applied, with a power of 1550 W and a sampling depth of 10 mm. A full mass scan was carried out, from Li 7 to Pb 208, and the resulting spectra analysed using the Agilent MassHunter soware.
For the HPLC-ESI-MS/MS identication of intermediates, 0.38 mmol L À1 of ionic liquid solution was prepared, and UVA/ TiO 2 P25/H 2 O 2 system was used. Filtration was carried out to separate the photocatalyst particles as previously described. Then, a 20 mL of sample was injected and analyzed by an Agilent Technologies 1200 series HPLC with Agilent Technologies 6410A series electrospray ionization triple-quadrupole MS/MS, using the aforementioned polar CN column held at 25 C. The mobile phase, consisting of 0.1% aqueous formic acid and ACN, was delivered at ow rate 0.8 mL min À1 .
In negative mode, isocratic elution (20% ACN) was used. In positive mode, 100 mmol L À1 ammonium acetate was added to aqueous phase to enhance retention of charged quaternary nitrogen compounds (that otherwise elute at dead time), and gradient mode (0 min 20% ACN, 4 min 20% ACN, 6 min 100%, post time 3 min) was employed to facilitate separation. Analytes were ionized using the electrospray ion source, with nitrogen as the drying gas (temperature 350 C, ow 10 L À1 min À1 ) and nebulizer gas (50 psi), and a capillary voltage of 4.0 kV. High purity nitrogen was used as the collision gas. Full scan mode (m/z range 50-500, scan time 100 ms, fragmentor voltage 100 V), using both polarities, was used to detect degradation intermediates, determine their molecular weight and select precursor ions for MS 2 experiments. To obtain structural information and identify the detected compounds, product ion scan analysis (MS 2 experiment) was used, with [M + H] + and [M À H] À ions as precursors, and collision energies in 10-40 V interval (10 V increments).

Effect of pH on ionic liquid hydrolysis
The degradation of ionic liquid in aqueous solution in the absence of light at ambient temperature of 25 C was monitored at different pH values. As can be seen from Fig. 1a

Preliminary dark experiments
Since the suspension was stirred continuously in the dark for 15 min prior to the degradation, it was examined whether this contact time was sufficient for [bmim][Sal] to reach adsorptiondesorption equilibrium on the photocatalyst. Even aer 180 min, the removal of [bmim][Sal] by dark adsorption at natural pH and pH 2.8 was almost negligible (Fig. 2).
Furthermore, obtained results for the experiments conducted in dark showed that the ionic liquid was stable in the presence of H 2 O 2 alone (Fig. 3, curve 1), as well as in a combination of the TiO 2 photocatalysts (TiO 2 P25 or TiO 2 support) and H 2 O 2 (Fig. 3, curves 3 and 4). Also, using 7.2Fe/TiO 2 did not cause the ionic liquid degradation (Fig. 3, curve 5).
However, in the presence of H 2 O 2 and 7.2Fe/TiO 2 aer 180 minutes 27% of [bmim] + (Fig. 3a, curve 2), and 92% of [Sal] À (Fig. 3b, curve 2) has been degraded. This behavior can be attributed to the Fenton process.  According to our previous examinations and since all of the experiments were conducted at pH 2.8, for 7.2Fe/TiO 2 /H 2 O 2 system is expected that the concentration of iron in solution will not exceed 8 ppb. In compliance with this, the degradation efficiency was studied in the presence of 8 ppb dissolved iron(III) and appropriate amounts of hydrogen peroxide. Aer 180 min it has been found that the ionic liquid degradation efficiency at this level of dissolved iron concentration (homogeneous Fenton processes) was negligible.
The same conclusion was derived from the experiment with 60 ppb of Fe(III). This experimental evidence implies that the overall degradation efficiency originates from the heterogeneous Fenton process (reactions (1)- (5)). 38 Finally, it can be concluded that within this process existence of surface bonded iron oxide was crucial. This statement is based on our previous experiments 27 which were carried out with PrecFe (iron particles contained in 7.2Fe/TiO 2 photocatalyst) in the presence of H 2 O 2 , and it was found that the degradation efficiency was practically the same to that in its absence. This was explained with the fact that the solution did not practically contain Fe 3+ ions. However, the presence of H 2 O 2 was increased the catalytic efficiency of 7.2Fe/TiO 2 to a higher extent than of the TiO 2 support and PrecFe, which suggests that H 2 O 2 plays a different role, i.e., in the case of 7.2Fe/TiO 2 /H 2 O 2 , in addition to the electron-acceptor role, H 2 O 2 is also involved in the heterogeneous photo-Fenton process.

UVA photodegradation studies
The [bmim] + strongly absorbs UV radiation over a range of wavelengths, from 200 to 235 nm, with the maximum absorption band at 210 nm. On the other hand, the [Sal] À showed UV absorption at a broader range between 200 and 335 nm, with three absorption maxima at 205, 236, and the lowest at 303 nm. Since the UV radiation of used lamp below of closely 275 nm cannot pass through Pyrex glass, 25 no direct photolysis of [bmim] + could be expected. Also, observing the radiation of lamp which passed through the wall of the reactor and the absorption spectrum of absorbed radiation by [Sal] À in the range of 275 to 335 nm, a direct photolysis of the [Sal] À could be possible.
Besides direct photolysis in Fig. 4 was presented the efficiency of other AOPs that can be potentially used for the ionic liquid degradation. This includes the results obtained using indirect photolysis by H 2 O 2 (Fig. 4, curve 2), and photocatalysis with TiO 2 P25 (with and without H 2 O 2 , Fig. 4, curves 3 and 5), as well as with 7.2Fe/TiO 2 /H 2 O 2 system (Fig. 4, curve 4). In experiments in which H 2 O 2 was employed, because of its depletion, H 2 O 2 was added successively (70 mL each hour during the radiation). By comparing the results presented in Fig. 3    Also increased efficiency of indirect photolysis can be explained by the fact that radiation with wavelengths less than 300 nm can cause photolysis of hydrogen peroxide. 39

Simulated sunlight photodegradation studies
Further tests were conducted to examine the performance of different AOPs through degradation of ionic liquid under the SS radiation for the rst time. Using this kind of the light source, the degradation process was more similar to that of the practical process which proceeds under sunlight.
Comparing obtained results under SS radiation (Fig. 5) with those obtained for the same processes which were carried out in the dark (Fig. 3), in all cases increasing efficiency can be observed.
For the studied SS/H 2 O 2 , SS/TiO 2 P25 and SS/7.2Fe/TiO 2 / H 2 O 2 processes, these increases of efficiency were different for cation and anion. Namely, for the [bmim] + (Fig. 5a) were 17%, 9%, and 4.5%, respectively, while the same in the case of [Sal] À (Fig. 5b) were 74%, 52%, and 5.3%, respectively. In this case, the spectral energy distribution of used SS source was more suited  for conducting indirect photolysis in relation to the photocatalysis. On the other hand, for the SS/7.2Fe/TiO 2 /H 2 O 2 process it was noticed that the contribution of the heterogeneous photo-Fenton oxidation was negligible compared to the dark heterogeneous Fenton oxidation.
Also, if compared Fig. 5 and 4, for all of the studied processes, ionic liquid photodegradation efficiency was significantly lower using SS radiation, compared to that in the presence of UVA radiation.
Further research has been focused on studying mutual facilitation of [bmim] + and [Sal] À on the photodegradation efficiency in the presence of the SS radiation. For this purpose, the SS/TiO 2 P25 process was chosen. By applying this process the degradation efficiency in the presence of only [bmim] + or only [Sal] À has been studied (Fig. 6) and compared with those obtained when they are in a mixture (referring to the ionic liquid, Fig. 5). When the degradation efficiency of individual ions was studied, [bmim] + was present in the form of chlorides, while the [Sal] À was present in the form of the sodium salt. Comparing results presented in Fig. 5 and 6 it can be seen that aer 180 min of only [Sal] À degradation, efficiency decreased by 5.4%, while in the case of [bmim] + , efficiency was increased for 36%. Based on previous conclusions it can be established that the impact of [Sal] À on the [bmim] + degradation efficiency was much more pronounced. It was also noted that in all experiments when both ions are present in the mixture, the [Sal] À showed signicantly higher degradation tendency in comparison with [bmim] + .

Degradation in different water types
It was very important to study the stability properties of ionic liquids toward AOPs in different types of water, since it is very likely that these substances will get into the different types of water that can be oen found in the environment.
The studied degradation processes included heterogeneous Fenton and photo-Fenton processes. Fig. 7 depicts the degradation curves of the ionic liquid in different waters by applying dark heterogeneous Fenton process. In the case of [bmim] + (Fig. 7a), experimental data obtained aer 180 min indicated that efficiency of degradation decreases in the following order: pond < rain < tap < river < condensate water, showing a strong dependence on the constitution of the radiated media.
During the [bmim] + degradation in acidic media (pH 2.8) the surface of 7.2Fe/TiO 2 particles was positively charged, so that the extent of adsorption of metal ions was relatively insignicant. From the literature it is known that Na + , K + , Ca 2+ , and Mg 2+ ions are the common cations in aquatic environment. They are all in the highest and stable oxidation state and cannot capture electrons or holes in solution. 40 In the case of dark Fenton process it can be assumed that the [bmim] + degradation efficiency mostly depended on the P, S, and Cl content (Table

S1
, ESI †). The highest degradation efficiency was found in pond water and it is probably due to the lowest total content of P, S, and Cl. By comparison of composite of pond water with rainwater in which the efficiency of the [bmim] + degradation is some lower, it can be seen that in rainwater was 4 times higher content of P, 1.5 times higher S content, but 1.3 times lower Cl content. In the case of tap water also in comparison with pond water, the S content was 8 times lower, but the contents of P and Cl were eight and two times higher, respectively. Furthermore, the Danube water had the highest content of P, S and Cl, so its degradation efficiency was the lowest. It can be assumed that the impact of content on the efficiency of degradation process decreases in the following order: P > Cl > S. Bali et al. came to a similar conclusion of negligible effect of SO 4 2À and Cl À , in comparison with the effect of PO 4 3À which was remarkable. 41 Condensate water has the lowest P content, followed by pond water. The lower efficiency of degradation observed in condensate water can be attributed to the highest content of copper which leads to complexation with humic acids. 42 On the other hand, in the case of [Sal] À aer 180 min of degradation approximately equal efficacy was observed for all types of used water (Fig. 7b). By applying a dark Fenton process, for all types of investigated water, a higher degradation efficiency of [Sal] À compared to [bmim] + was observed (Fig. 7b vs. a). This can be explained by the fact that salicylic acid could present in different formsnon-dissociated at pH < 3.0, and dissociated at pH 6.0-12 (the pKa for salicylic acid is 2.97). 43 It was also found that the adsorption of salicylate was the highest at pH 3 on iron containing surface. 44 Besides, higher degradation efficiency as recorded in our case, can be explained by strong ability of salicylic acid to complex with surface iron atoms of photocatalyst under acidic conditions. 32 Further studies were focused on examining of ionic liquids degradation efficiency in different types of water by using heterogeneous photo-Fenton process (Fig. 8). For both [bmim] + and [Sal] À a signicant increase in the degradation efficiency by SS/7.2Fe/TiO 2 /H 2 O 2 process was observed (Fig. 5 vs. 8). The small contribution to degradation efficiency comes from the heterogeneous photo-Fenton process (see simulated sunlight photodegradation studies). However, signicantly greater contribution to degradation efficiency comes from presence of SS radiation due of dissolved organic carbon (DOC). By absorption of sunlight radiation DOC could provide a rich variety of photochemical reactions. 45 The resulting excited states of the DOC could participate in energy transfer, electron transfer, and free radical reaction, which affect the fate of aquatic pollutants. 46 In this case, it can be concluded that the present DOC together with SS radiation provided a powerful sensitization effect (DOC was signicant source of OH radicals production) and thus increase the ionic liquid degradation efficiency. At the end it can also be concluded that the amount range and composition of DOC (Table 1) in different types of water was such that the ionic liquid degradation efficiency in presence SS were nearly the same.

Identication of intermediates
Identication of the intermediates was carried out using the HPLC-ESI-MS/MS technique. Based on their MS/MS fragmentation data, twenty-two intermediates originating from [bmim] + were identied in positive mode, and one intermediate originating from [Sal] À was identied in negative mode (Table 2; S2, ESI †).
P1 is a compound with M mi ¼ 137, which is 2 mass units lower than the molecular mass of 3-butyl-1-methylimidazolium (PI), indicating loss of two hydrogen atoms, most likely from butyl-chain, indicating that P1 is 3-(2-but-2-en-1-yl)-1-methyl-1H-imidazol-3-ium. P2 and P3 are ionic compounds with M mi ¼ 153. DM mi of 14 indicates the presence of oxo group in both compounds, thus it seems likely that P2 is 3-butanoyl-1-methyl-1H-imidazol-3-ium and P3 is 1-methyl-3-(2-oxobutyl)-1Himidazol-3-ium. P4 and P5 are ionic compounds with M mi ¼ 155. Mass difference of 16 units suggests hydroxylation, with compound P4 likely standing for 3-(1-hydroxybutyl)-1-methyl-1H-imidazol-3-ium, and compound P5 being 3-(4-hydroxybutyl)-1-methyl-1H-imidazol-3-ium. P6 and P7 are isobaric compounds with M mi ¼ 171, which is 32 units higher than the monoisotopic mass of 3-butyl-1-methylimidazolium (PI), indicating the presence of two hydroxyl groups. P6 could possibly represent 3-(3,4dihydroxybutyl)-1-methyl-1H-imidazol-3-ium and P7 is 3-(2,3dihydroxybutyl)-1-methyl-1H-imidazol-3-ium. Compounds P8 and P9 both have M mi ¼ 173. The monoisotopic mass, 34 mass units greater than that of 3-butyl-1-methylimidazolium, points out to the presence of two hydroxyl groups and either reduction of one double bond in imidazolium ring or ring cleavage. P8 likely stands for [formyl(methyl)amino]-N-(hydroxybutyl)-Nmethylmethaniminium and P9 is likely N-butyl[formyl(methyl) amino]-N-(hydroxymethyl)methaniminium. Although compounds P10, P11 and P12 have identical monoisotopic masses of M mi ¼ 187, their fragmentation patterns suggest that they have various distribution of three hydroxyl groups (Dm/z ¼ 48) present in them. P10 likely stands for 3-(dihydroxybutyl)-1-  (hydroxymethyl)-1H-imidazol-3-ium, intermediate P11 was identied as 1-methyl-3-(trihydroxybutyl)-1H-imidazol-3-ium and P12 is 3-(dihydroxybutyl)-1-(hydroxymethyl)-1H-imidazol-3-ium. P13 and P14 both correspond to compounds with monoisotopic mass M mi ¼ 189, 50 mass units higher than the molecular mass of 3-butyl-1-methylimidazolium. Based on their PI MS 2 spectra and specic patterns of fragmentation, it was concluded that these compounds could have either three hydroxyl groups and reduction of one double bond in imidazolium ring, or cleavage of imidazolium ring with hydroxyl and oxo group attached, along with another hydroxyl on butyl or Nmethyl moiety. It was not possible to elucidate the structure of P13, while P14 most likely stands for hydroxy-N- (3- Comparing obtained results, with those obtained for the same processes which were carried out in the dark, in all cases increasing efficiency can be observed. Also, for all of the studied processes, ionic liquid photodegradation efficiency was signicantly lower in the presence SS radiation, compared to that in the presence of UVA radiation. By studying mutual facilitation of [bmim] + and [Sal] À on the photodegradation efficiency, the impact of [Sal] À on the [bmim] + degradation efficiency was much more pronounced. From a practical point of view, it was very important to study the stability properties of ionic liquids toward AOPs in different types of water. Applying dark heterogeneous Fenton process on [bmim] + indicated that efficiency of degradation decreases in the following order: pond < rain < tap < river < condensate water showing a strong dependence on the constitution of the radiated media. On the other hand, in the case of [Sal] À approximately equal efficacy was observed for all types of used water. Also, for all types of water, a higher degradation efficiency of [Sal] À compared to [bmim] + was observed. Furthermore, in presence of SS radiation the present DOC provided a powerful sensitization effect and thus increases the ionic liquid degradation efficiency.
LC-MS/MS analysis indicated that intermediates with one or more OH groups were produced during the photocatalytic degradation process. In addition, for the most efficient UVA/ TiO 2 P25/H 2 O 2 system structure of nineteen intermediates are assumed.