Simultaneous Determination of Isothiazolinones and Parabens in Cosmetic Products Using Solid-Phase Extraction and Ultra-High Performance Liquid Chromatography/Diode Array Detector

Isothiazolinones methylisothiazolinone (MI) and methylchloroisothiazolinone (MCI), and parabens methylparaben (MP), ethylparaben (EP), propylparaben (PP) and butylparaben (BP) are the most common synthetic preservatives. They are all known to be potential skin allergens that lead to contact dermatitis. Thus, the identification of these unsafe chemicals in cosmetic products is of high importance. In the present study, solid-phase extraction (SPE) based on HyperSep reversed-phase C8/benzene sulfonic acid ion exchanger (HyperSep C8/BSAIE) and Sep-Pak C18 sorbents, and ultra-high performance liquid chromatography/diode array detector (UHPLC/DAD) were optimized for the simultaneous determination of MI, MCI, MP, EP, PP and BP in cosmetic products. HyperSep C8/BSAIE and UHPLC/DAD with the eluting solvent mixture (acetonitrile/methanol, 2:1, v/v) and detection wavelength (255 nm) were found to be the optimal conditions, respectively. The method illustrates the excellent linearity range (0.008–20 μg/mL) with coefficient of determination (R2, 0.997–0.999), limits of detection (LOD, 0.001–0.002 μg/mL), precision in terms of relative standard deviation (RSD < 3%, intra-day and <6%, inter-day) when examining a standard mixture at low (0.07 µg/mL), medium (3 µg/mL) and high (15 µg/mL) concentrations. A total of 31 cosmetic samples were studied, achieving concentrations (MI, not detected (nd)-0.89 µg/g), (MCI, nd-0.62 µg/g), (MP, nd-6.53 µg/g), (EP, nd-0.90 µg/g), (PP, nd-9.69 µg/g) and (BP, nd-17.80 µg/g). Recovery values ranged from 92.33 to 101.43% depending on the types of sample. To our knowledge, this is the first specific method which covers the theme and describes background amounts of such preservatives in cosmetics.

Relating to the determination of parabens, different methods have been reported earlier in various matrices, for instance, HPLC-fluorescence/UV/DAD (cosmetics, toothpaste and mouthwash) [23,24,34], paper spray-MS/MS (cosmetics and drugs) [35], capillary liquid chromatography with UV detection (cosmetics, food and pharmaceuticals) [11], UHPLC-high-resolution mass spectrometry (human urine) [36], UHPLC-MS/MS (human milk) [37] HPLC-MS/MS (domestic sewage) [38] and so on. Some common methods had also been recently reported; Abad-Gil et al., (2021) optimized the simultaneous determination of MI, MCI, 4-hydroxybenzoic acid, phenoxyethanol and MP in cosmetics, using an HPLC/DAD/FL system [21]. In another study, Hefnawy et al., (2017) reported the simultaneous analysis of MI, MP, EP, PP and salicylic acid in cosmetics by monolithic HPLC-PDA [39]. In both methods, the studied compounds and applied methods were found to be different than those used in the current study. The current method (UHPLC/DAD) was found to be rapid, sensitive and economical, especially for the simultaneous determination of MI, MCI, MP, EP, PP and BP in cosmetics.
Saudi Arabia is the main marketplace for cosmetic products in the Arab and African countries, and it has one of the world's utmost cosmetics consumption rates. Recently, the Saudi Arabia personal care and beauty market was forecasted to attain $5.5 billion by 2025, rising at a compound annual growth rate of 10.49% during the forecast period (https://www.mordorintelligence.com/industryreports/saudi-arabia-beauty-and-personal-care-marketm).
Up until now, there has been no earlier analytical system for the identification of MI, MCI and parabens (MP, EP, PP and BP) in cosmetic products by using a single extraction and determination method. Thus, our investigation aimed for the development and validation of a specific method based on solid-phase extraction (SPE) and ultra-high-performance liquid chromatography/diode array detector (UHPLC/DAD) for the simultaneous determination of MI, MCI, MP, EP, PP and BP in cosmetic products.

Optimization of SPE Method
At present, several extraction methods are available which deal either with MI and MCI [2,3,22] or four potential parabens (MP, EP, PP and BP) [11,23,24,34,35] determination in cosmetics. Thus, the most important aim of the current study was to develop a single extraction and determination method for MI, MCI, MP, EP, PP and BP, which usually co-occur in cosmetic samples. This is the first approach relating to the extraction and determination of these compounds in cosmetics by using a single method.
Owing to the low amounts of MI, MCI, MP, EP, PP and BP present in the cosmetic samples, the optimization of a reliable method for their analysis is of high importance, and thus, it required a very functional extraction and cleanup system which can eliminate the sample matrix interferences that typically interfere with the determination of target compounds by UHPLC/DAD system.
According to the nature of the analyzed compounds, initially, we selected two types of SPE extraction cartridges, namely HyperSep™ Verify CX Cartridges (Thermo Fisher Scientific, San Jose, CA, USA) of HyperSep C 8 /BSAIE (200 mg/mL) and Sep-Pak C 18 Classic Cartridge, 360 mg, particle size 55-105 µm (Waters (Milford, MA, USA). Preliminary studies were performed by using a 20 mL mixed solution (3 µg/mL) prepared in methanol and water of all the targeted compounds (MI, MCI, MP, EP, PP and BP). A series of experiments were carried out by the passing sample mixture solution at a controlled flow rate (1 mL/min) through two SPE cartridges separately. Once the sample solution passed completely, the targeted compounds from SPE sorbents were eluted by using different solvent mixtures (10 mL) at various proportions: water/methanol, water/acetonitrile and acetonitrile/methanol. After that, the solution was evaporated under nitrogen gas, until there remained 3 mL of the total solution volume, followed by filtration by using a polytetrafluoroethylene (PTFE) syringe filter Pharmaceuticals 2020, 13, 412 4 of 15 (0.45 µm). Finally, the filtrate was injected to UHPLC/DAD, for the determination of MI, MCI, MP, EP, PP and BP. Among them, the SPE cartridge HyperSep C 8 /BSAIE and eluting solvent mixture (acetonitrile/methanol, 2:1, v/v) were found to be the optimal extraction parameters and used for the analysis of real samples. Figure 1 demonstrates the UHPLC/DAD chromatograms of MI, MCI, MP, EP, PP and BP (standard solution mixture, 3 µg/mL), obtained using different SPE cartridges and eluting solvent mixtures at proportion (2:1, v/v). It can be observed from Figure 1 that using (A1) water/methanol 2:1, v/v and Sep-Pak C18; (A2) water/acetonitrile 2:1, v/v and Sep-Pak C18; and (A3) acetonitrile/methanol 2:1, v/v and Sep-Pak C18 conditions, the compounds were either not detected or found below LOD. Nevertheless, by using (B1) water/methanol 2:1, v/v and HyperSep C8/BSAIE; (B2) water/acetonitrile 2:1, v/v and HyperSep C8/BSAIE; and (B3) acetonitrile/methanol 2:1, v/v and HyperSep C8/BSAIE conditions, the compounds have been identified in all cases. In B1 and B2 conditions, the compounds were identified with the poor resolution with low peak intensity. However, in B3 conditions, the compounds were identified with excellent resolution and symmetrical with high intensity.

Optimization of UHPLC/DAD Method
The most important challenge on the new UHPLC/DAD system was to separate MI, MCI, MP, EP, PP and BP in a single run, with the advantage of high peak resolution, symmetry and short analysis time. Due to differing in their polarity, many determination methods have been reported earlier which deal either with MI and MCI [2,3,22] or parabens (MP, EP, PP and BP) [11,23,24,34,35] in cosmetics. For the optimization of the UHPLC/DAD system, a standard mixture solution (3 µg/mL) was analyzed by using ACCLAIM™ 120 C 8 analytical column and mobile phase with different solvent proportions, such as water (0.1% formic acid) with acetonitrile/methanol; water (0.05% trifluoroacetic acid) with acetonitrile/methanol; and water (0.1% trifluoroacetic acid) with acetonitrile/methanol. During the assessment of the method parameters, the absorbance was studied in the range of 250-280 nm. The most favorable chromatographic conditions for the analysis of MI, MCI, MP, EP, PP and BP was water (0.1% trifluoroacetic acid) with acetonitrile (mobile phase) and absorbance 255 nm, selected as the final method for real sample analysis. Figure 2 displays the UHPLC/DAD chromatograms obtained at optimal chromatographic conditions. The method offers excellent peak resolution and symmetry and a total analysis time lower than 25 min. The influence of column temperature on analysis was also established in the range from room temperature, 25 • C, to 50 • C, with 5 • C variations. The analysis time was reduced with increasing column temperature beyond 35 • C, which offered a poorer compounds' separation. Consequently, the column temperature of 35 • C was selected as an optimal condition.

Performance of the Method
The performance of the proposed method was investigated in terms of linearity (R 2 ), limit of detection (LOD, signal-to-noise ratio 3:1) and limit of quantification (LOQ, signal-to-noise ratio 10:1), precision (intra-and inter-day) and accuracy. The achieved values have been presented in Tables 1 and 2. Linearity was determined by analyzing standard mixture at different concentrations, ranging from 0.008 to 20 µg/mL. The analysis was performed in triplicates (n = 3). Calibration curves were found to be linear over the broad range of concentrations with the coefficient of determination (R 2 , 0.997-0.999). LOD (signal-to-noise ratio 3:1) and LOQ (signal-to-noise ratio 10:1) values were calculated from the calibration equations using formula 3*standard deviation of the response/slope. LOD and LOQ values were found in the range of 0.001 to 0.002 µg/mL and 0.004 to 0.007 µg/mL, respectively. Precision (intra-and inter-day) was estimated in terms of relative standard deviation (RSD%) and achieved < 3% for intra-day and <6% for inter-day when examining a standard mixture of targeted compounds at concentrations of low (0.07 µg/mL), medium (3 µg/mL) and high (15 µg/mL) levels. Recovery values of targeted compounds were assessed at low, medium and high levels in all of the analyzed samples, and obtained from the added and found concentrations of each compound. The recovery values were achieved between 92.33% and 101.43% depending on the types of sample. The excellent quality conditions were obtained and can be proposed for the determination of these compounds in cosmetics.
Sep-Pak C18; and (A3) acetonitrile/methanol 2:1, v/v and Sep-Pak C18 conditions, the compounds were either not detected or found below LOD. Nevertheless, by using (B1) water/methanol 2:1, v/v and HyperSep C8/BSAIE; (B2) water/acetonitrile 2:1, v/v and HyperSep C8/BSAIE; and (B3) acetonitrile/methanol 2:1, v/v and HyperSep C8/BSAIE conditions, the compounds have been identified in all cases. In B1 and B2 conditions, the compounds were identified with the poor resolution with low peak intensity. However, in B3 conditions, the compounds were identified with excellent resolution and symmetrical with high intensity.  excellent peak resolution and symmetry and a total analysis time lower than 25 min. The influence of column temperature on analysis was also established in the range from room temperature, 25 °C, to 50 °C, with 5 °C variations. The analysis time was reduced with increasing column temperature beyond 35 °C, which offered a poorer compounds' separation. Consequently, the column temperature of 35 °C was selected as an optimal condition.

Performance of the Method
The performance of the proposed method was investigated in terms of linearity (R 2 ), limit of detection (LOD, signal-to-noise ratio 3:1) and limit of quantification (LOQ, signal-to-noise ratio 10:1), precision (intra-and inter-day) and accuracy. The achieved values have been presented in Tables 1  and 2. Linearity was determined by analyzing standard mixture at different concentrations, ranging from 0.008 to 20 μg/mL. The analysis was performed in triplicates (n = 3). Calibration curves were found to be linear over the broad range of concentrations with the coefficient of determination (R 2 , 0.997-0.999). LOD (signal-to-noise ratio 3:1) and LOQ (signal-to-noise ratio 10:1) values were  LOD, signal-to-noise (s/n, 3:1); LOQ, signal-to-noise (s/n, 10:1); SD, standard deviation, obtained from three replicates.

Comparison of Proposed Method with the Previous Works
A comparison of the proposed method with the earlier reported analytical methods is presented in Table 3. Earlier methods have individually identified MI and MCI [2] or parabens mixed with other compounds in cosmetics, environmental, biological, pharmaceuticals and personal care samples [11,23,25,33,40], but never the simultaneous identification of the MI, MCI, MP, EP, PP and BP in cosmetics by SPE/UHPLC/DAD. The reason for not identifying these two classes of compounds in a single analysis is the differing polarities, especially when applying a mass spectrometric system. In addition, the reported sample-preparation techniques for the chromatographic determination have only dealt with particular product types. In another approach, Lin et al.  [42]. These established values [41,42] were also found in good agreement with those archived in the current study. Moreover, on the basis of the achieved outcomes from the current study, the present method could be applied for the determination of MI, MCI, MP, EP, PP and BP in various kinds of matrices. At present, there is no common method available for such kinds of determination, either in sample preparation technique or chromatographic system. The reported SPE/UHPLC/DAD method can simultaneously analyze MI, MCI and parabens (MP, EP, PP and BP) in a single chromatographic method in cosmetic products. The performance of the reported method (linearity, LOD, precision and accuracy) was found to be in good agreement with those reported in earlier works [2,11,23,25,33,40].

Chemical and Reagents
HPLC grade acetonitrile and methanol were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trifluoroacetic acid, MI, MCI, MP, EP, PP and BP were obtained from Merck (Darmstadt, Germany). All chemicals were of high purity (>99%). The structures of the studied compounds are demonstrated in Figure 4. Ultrapure water was prepared by using a Barnstead TM Smart2Pure TM water purification system (Thermo Scientific, Göteborg, Sweden). Solid-phase extraction cartridges,
The individual stock solution was prepared at a concentration of 200 mg/L, by dissolving an adequate weight in methanol, used for further dilutions. To produce the linearity range and calibration curves, standard mixtures of the studied compounds at concentrations ranging from 0.008 to 20 µg/mL were prepared. Standard solutions and cosmetic samples were filtered by a syringe PTFE filter (0.45 µm) (Macherey-Nagel GmbH, Düren, Germany) before being analyzed by the UHPLC/DAD system.
To assess the SPE efficiency and prevent the matrix influence on peak intensity, retention time and symmetry, the MI, MCI, MP, EP, PP and BP quantification was performed by a standard addition procedure (a quantitative analysis method applied to reduce matrix effects that obstruct with compound measurement signals) consisting of non-fortified (two, zero levels) and fortified (three levels, 50%, 100% and 500%) samples. The levels values demonstrated the increase of compounds in the sample after fortifying. The fortifying of samples was carried out at the start of the extraction method. Cosmetic samples were studied in triplicates (three different extractions of the same sample), and statistical data analysis of the studied samples was performed by means of ANOVA (analysis of variance). electronic (Berlin, Germany). Whatman ® qualitative filter paper, Grade 1 circles, diameter 90 mm, was purchased from Merck (Darmstadt, Germany). Polytetrafluoroethylene (PTFE) syringe filter (0.45 μm) was purchased from Macherey-Nagel GmbH (Düren, Germany).
The individual stock solution was prepared at a concentration of 200 mg/L, by dissolving an adequate weight in methanol, used for further dilutions. To produce the linearity range and calibration curves, standard mixtures of the studied compounds at concentrations ranging from 0.008 to 20 μg/mL were prepared. Standard solutions and cosmetic samples were filtered by a syringe PTFE filter (0.45 μm) (Macherey-Nagel GmbH, Düren, Germany) before being analyzed by the UHPLC/DAD system. To assess the SPE efficiency and prevent the matrix influence on peak intensity, retention time and symmetry, the MI, MCI, MP, EP, PP and BP quantification was performed by a standard addition procedure (a quantitative analysis method applied to reduce matrix effects that obstruct with compound measurement signals) consisting of non-fortified (two, zero levels) and fortified (three levels, 50%, 100% and 500%) samples. The levels values demonstrated the increase of compounds in the sample after fortifying. The fortifying of samples was carried out at the start of the extraction method. Cosmetic samples were studied in triplicates (three different extractions of the same sample), and statistical data analysis of the studied samples was performed by means of ANOVA (analysis of variance).
LOD and LOQ were calculated from the calibration equation, i.e., 3*standard deviation of the response/slope. Recovery of MI, MCI, MP, EP, PP and BP was assessed at low, medium and high levels in all of the analyzed samples, and obtained from the added and found concentrations of each analyte.

Extraction Method
For the identification and quantification of MI and MCI, and four potential parabens, comprising MP, EP, PP and BP, cosmetics of diverse trademark and country of origin were obtained from cosmetic and pharmacy retail superstore based in Al-Jouf and Riyadh, Saudi Arabia. The sample LOD and LOQ were calculated from the calibration equation, i.e., 3*standard deviation of the response/slope. Recovery of MI, MCI, MP, EP, PP and BP was assessed at low, medium and high levels in all of the analyzed samples, and obtained from the added and found concentrations of each analyte.

Extraction Method
For the identification and quantification of MI and MCI, and four potential parabens, comprising MP, EP, PP and BP, cosmetics of diverse trademark and country of origin were obtained from cosmetic and pharmacy retail superstore based in Al-Jouf and Riyadh, Saudi Arabia. The sample description was presented in Table 4. Subsequent to purchase, cosmetic samples were immediately stored at 4 • C, and studied at the earliest time, to avoid any chemical loss or contamination. To examine the selective extraction by using SPE method, 0.5 g of cosmetic samples was added to a mixture solution of water and methanol (50:50, v/v, 20 mL), followed by mixing (10 min) by using a magnetic stirrer. Afterward, the sample mixture was sonicated (10 min) in ultrasonic baths, followed by filtration through Whatman ® qualitative filter paper (grade 1 circles, diameter 90 mm). Then, the sample filtrate was eluted through the SPE cartridge (HyperSep C 8 /BSAIE), at a controlled flow rate (1 mL/min). Finally, the analyte was eluted with a mixture solution of acetonitrile and methanol (2:1, v/v, 10 mL). The sample solution was evaporated, under nitrogen gas, to a final volume of 3 mL. Prior to the analysis by using UHPLC/DAD, the sample extract (3 mL) was filtered through a PTFE syringe filter (0.45 µm). The volume of sample injection was 10 µL. The samples were extracted in triplicates (three different extractions of the same sample). In order to verify the sample contamination and method sensitivity maintained throughout the study, quality control samples were analyzed. Besides this, the sampling steps were carried out with safety measures to reduce sample contamination.
Because of the complexity of cosmetics preparations, a precise pretreatment of the cosmetic samples is typically needed prior to the identification of these compounds by using the UHPLC/DAD technique. The present SPE method using HyperSep C8/BSAIE cartridge was found to be precise and selective for the analysis of MI, MCI, MP, EP, PP and BP in cosmetics. Nevertheless, in earlier studies, the authors have reported various extraction methods based on ultrasound-assisted extraction, solid-phase microextraction, vortex-assisted dispersive liquid-liquid microextraction and liquid-liquid extraction for the analysis of preservatives in different matrices [3, [44][45][46]. These methods were also found to be precise and selective for different types of compounds extracted from different matrices.

Instrumentation
The sample analysis was performed by using a Dionex UltiMate 3000 UHPLC system (Thermo Scientific, San Jose, CA, USA), comprising a LPG-3400SD binary pump, WPS-3000TSL thermostat autosampler, TCC-3000SD thermostat column compartment and DAD-3000 diode array detector. The data were recorded and analyzed by Chromeleon™ 7.2 Chromatography Data System Software (Thermo Scientific, San Jose, CA, USA).
The chromatographic separation of MI, MCI, MP, EP, PP and BP was achieved through an ACCLAIM™ 120 C 8 analytical column with the dimensions 150 mm × 2.1 mm and 5 µm of particles size (Thermo Scientific, San Jose, USA). The optimal separation was obtained by using binary mobile phase: water (0.1% trifluoroacetic acid, pH 2.1, solvent A) and acetonitrile (solvent B) at a flow rate of 0.5 mL/min. The gradient mobile phase elution was 0-2 min (B, 12.5%), 2-4 min (B 20-30%), 4-16 min (B, 30-50%), 16-22 min (B, 50-100%), return to its equilibrium conditions and 22-30 min. The column temperature was kept at 35 • C, and the sample injection volume was 10 µL. The column was also washed with a mixture (50:50, v/v) of methanol and Milli-Q water solution, for five minutes, following the analysis of every ten samples. The optimal detection wavelength was performed in the UV range at 255 nm.

Conclusions
A HyperSep C 8 /BSAIE SPE and UHPLC/DAD method for the simultaneous identification of MI, MCI, MP, EP, PP and BP in cosmetic products was optimized and validated, using 31 cosmetic samples of various trademarks and origin. These unsafe chemicals are the most common preservatives that manufacturers frequently apply in such products. In addition, excellent method performance parameters, namely linearity (R 2 , 0.997-0.999), LOD (0.001-0.002 µg/mL), precision (<6%) and accuracy as percent recovery (92.33-101.43%), were achieved. These outcomes revealed that the developed method offers an alternative method for the quality control of MI, MCI, MP, EP, PP and BP in cosmetic products. The present method can be practically used for an extensive range of cosmetic products, for the identification of MI, MCI, MP, EP, PP and BP. For instance, this procedure will be appropriate to screen the frequency of wrong labeling of such unsafe chemicals (MI, MCI, MP, EP, PP and BP) on cosmetics' ingredients lists. Ingredients labels that are incorrect or omitted can go against clients and health care experts and when seeing for a causative agent to elucidate skin reactions and demanding to evade cosmetics that comprise these potential skin allergens leading to contact dermatitis.