3.1. LC-MS/MS analysis
All parameters were optimized by directly inject each standard at a level of 1 mg/L, prepared in acetonitrile, to identify the ionization mode for detection of all the analytes. In detail, among the 18 compounds, seven (4-Aminobenzoic Acid, Sulisobenzone, Avobenzone, Ecamsule, Bemotrizinole, Iscotrizinole, Ethylhexyl triazone) showed efficient ionization in ESI+. Afterwards, the precursor and the product ions for each compound were identified by optimizing the collision energy and the cone voltage. Consequently, the conditions of liquid chromatography were optimized. An Agilent Zorbax Eclipse XDB-C8 column was chosen for analyzing UV filters with particles size of 3.5 µm and a length of 100 mm. In addition, to optimize the mobile phase, a series of preliminary experiments were carried out, consisting of water and acetonitrile with formic acid at different concentrations (0.01%, 0.02%, 0.03%, 0.05%, and 0.1%) or with ammonium acetate (5mM) and water/methanol. Gradient elution with pure water and acetonitrile has been found to be the most efficient mobile phase for compounds analyzed in positive mode. This result can be related to the lower viscosity of the mixture of water / acetonitrile compared to water / MeOH, which reduce the pressure problems mostly frequent in liquid phase chromatography. Then, the intensities were better without additives (formic acid and ammonium acetate) for all compounds. The parameters of the drying gas temperature (250, 300 and 350°C) and the nebulizing gas pressure (5, 10, 15, 20 and 30 psi) were adjusted and the optimized intensities were established at 350 °C and 20 psi respectively. Multiple reaction monitoring (MRM) was utilized, and Table 1 shows the specific MS/MS parameters for the targeted compounds and their retention times.
3.2. GC-MS analysis
Data acquisition was done using the selected ion monitoring (SIM) and the full-scan modes. Full scan mode was selected with a wide range of m/z. Two monitored ions were chosen for each target compound, the first for confirmation while the second for quantitation, based on the scanning chromatograms of UV filter standards (Table 1).
3.3. Optimization of the extraction procedure
The SPE was chosen for extraction among the other extraction techniques. This technique is considered as the key step in sample pre-concentration due to its best extraction recovery and low solvent consumption rendering it the most utilized to extract pharmaceuticals compounds from aquatic samples. The filtration step did not influence the SPE extraction. Subsequently, the extraction process was assessed in terms of nature of sorbent, pH of sample, elution solvent and breakthrough volume.
3.3.1. Choice of cartridge type
The extraction efficacy of the SPE technique is highly affected by the type of the adsorbing material. For this reason, two different adsorbent materials were tested: Oasis HLB (200 mg) and Chromo Bond C18 (200 mg). Oasis HLB cartridge is constituted of two combined monomers, lipophilic divinylbenzene and hydrophilic N-vinylpyrrolidone , permitting the retention of polar and non-polar compounds (Khalikova et al. 2018). Moreover, this cartridge type has shown also a good performance in extracting acidic, neutral and basic compounds (Khalikova et al. 2018). Chromo Bond C18 is made by Silica material, permitting the retention polar and non-polar compounds (Khalikova et al. 2018). In this study, the results exhibited good extraction recoveries for all compounds with acceptable range (60–120%) using Chromo Bond C18 compared to Oasis HLB (Fig.1). Moreover, the addition of EDTA salt, which act as metals chelates, in the water sample was studied. The Fig.1 shows that the extraction recoveries of most of the UV-filter compounds were enhanced due to addition of EDTA salt. This result can be explained by the disruption of bonds between the studied target compounds and the metals presented in the water and by the decrease of the pH of the water samples allowing better retention of the compounds of interest on the cartridges.
3.3.2. Effect of pH
The adjustment of the pH of the water samples before being loaded into the cartridges consists another critical point of SPE procedure to be optimized. Therefore, the extraction efficiency of target compounds from adjusted water samples using formic acid was studied at four pH: 2, 4, 7 and 9. The adjusted water samples with pH 4 revealed the best extraction recoveries for all the target compounds (Fig.2). The obtained results can be explicated by the fact that when increasing the pH, the target compounds acquired an ionization charge, which impar their interaction with the solid phase. Instead, when the pH becomes acidic, the ionic charge of the target compounds reduced, allowing then the obtention of best extraction recoveries.
3.3.3. Optimization of the elution solvent
Furthermore, to eluate the target compounds from the adsorbents, it is necessary to use a solvent whose interaction with compounds is higher than the adsorbents allowing then their desorption. Then, in order to choose the best elution solvent, three solvents were tested (MeOH 100%, ACN 100% and ACN / MeOH 60/40), the yields obtained in the case of MeOH 100% and ACN / MeOH 60 / 40 are between 40 and 50%, while with pure ACN the recovery increases up to 78% (Fig.3). In the procedure of extraction, the pure acetonitrile was adopted due to the following reasons: it insured the highest recovery of all compound and a short evaporation time in the final step compared to the other eluents.
3.3.4. Optimization of breakthrough volume
Finally, a last experiment was carried out to assess the breakthrough volumes. It corresponds to the maximum volume, which can be percolated on an SPE cartridge without modifying the extraction yields. In this study, different sample volumes (100 mL, 250 mL, 500 mL and l L) were investigated. Samples were doped at the same concentration for all analytes (300 μg /L) and then extraction procedures were performed. As shown in Fig.4, the recoveries tended to decrease with a sample volume higher than 250 mL. For this reason, 250 mL has been selected as a sample volume to perform the extractions, thus a maximum enrichment.
3.4. Performance evaluation of our method
The developed method for the extraction of 18 UV filters in water samples was assessed for its analytical performance by evaluating the quality parameters (the linearity, precision, limit of detection (LOD) and limit of quantification (LOQ)) (Table 2). According to the optimum conditions, the method displayed good linearity with coefficients of determination (R2) higher than 0.9 for all the UV filters studied in the 0.13- 400 μg/L range, and repeatability, considered as percentage relative standard deviations (RSD < 20%).
The minimum concentration of a target analyte detected in a spiked water sample with a signal-to-noise ratio (S/N) of 3 is defined as LOD. LOQ is the lowest concentration of a compound that can be quantified in a sample with acceptable precision under the stated operational conditions of the method. The concentration of target compound equivalenting to a S/N of 10 is considered as LOQ, that can be determined, from the less intense MRM transition calculated, using an extract of ultrapure water spiked at the 50 μg/L level. Calculated LOD were between 0.03 and 30 μg/L and calculated LOQ were between 0.13 and 60 μg/L. Then, water samples were spiked with low concentrations of UV filters to determine the MDL. Calculated MDL were between 2.5 and 50 ng/L.
3.5. Application to water samples from swimming pool
To assure the suitability of the validated method, 10 recreational water samples, collected from swimming pools located in the region of Bekaa in Lebanon were worked and alayzed (Al Rihâb, Bekaa Joy, Sunny Land, Al Tilal, Park Hôtel, Kadery, Water Park, Serenity, Shams Palace and Mountajaa Al Sharek). Nine out of the 18 target UV filters were detected in the analyzed samples (Dioxybenzone, Octorylene, Padimate-O, Ethylhexyl Methoxycinnamate, Iscotrizinol, Oxybenzone, Bemotrizinol and Isoamyl Methoxycinnamate). As showed in (Fig.5), Padimate-O is found in 90% of swimming pools at low concentrations: 1 to 20 ng/l. Others filters: Oxybenzone, Dioxybenzone, Isoamyl Methoxycinnamate and Bemotrizinol were detected at concentrations between 70 and 180 ng/L. The last category including Octocrylene, Ethylhexyl Methoxycinnamate and Iscotrizinol were showed in high concentration between 1500 and 3000 ng/L. Octocylene has been reported to be found in 75% of the studied swimming pools water. The presence of UV filters in swimming pool water samples varies according to the persistence of each filter in the light and its release in water.