Solid-phase extraction coupled with ultra performance liquid chromatography tandem mass spectrometry to determine seven halogenated salicylanilides in cosmetics

Abstract

A solid-phase extraction (SPE) purification coupled with ultra performance liquid chromatography tandem mass spectrometry (SPE-UPLC-MS/MS) method was developed for simultaneous determination of seven halogenated salicylanilides (including 3,3′,4′,5-tetrachlorosalicylanilide, 4′-bromosalicylanilide, tribomsalan, dibromsalon, metabromsalon, 5′-bromo-4′-chlorosalicylanilide and 5′-chlorosalicylanilide) in cosmetics. Samples were firstly extracted using dichloromethane and purified by an amino (NH₂) SPE cartridge. The target analytes were separated by a Waters UPLC™ HSS T3 (100 mm × 2.1 mm, 1.8 μm) column and subsequently detected through an electrospray ionization (ESI) source in the negative mode with multi-reaction monitoring (MRM) conditions. An external standard method was adopted for the quantification. Finally, validation of the proposed method was performed in terms of linearity, limits of detection (LOD), selectivity, matrix effect (ME), accuracy and precision. Under the optimized conditions, linear relationships were favorable over the selected concentration ranges of 0.5–200.0 μg L⁻¹ for all the seven analytes, with correlation coefficients (R²) greater than 0.997. The limits of detections (LODs) were in the range of 1.5–2.5 μg kg⁻¹. Recovery experiments were conducted at three concentration levels spiked in two kinds of cosmetics (cream class and water class). The average recoveries were calculated between 83.9% and 107% with relative standard deviations (RSDs) of 2.9–6.5% for intra-day precision (n = 6) and 4.9–7.0% for inter-day precision (n = 5). The validated method was successfully applied to determine the concentrations of halogenated salicylanilides in thirty complex cosmetics samples, and dibromsalan (DBS) and 3,3′,4′,5-tetrachlorosalicylanilide (TCSA) were detected in two samples with the concentrations of 584 μg kg⁻¹ and 422 μg kg⁻¹, respectively.

---

1. Introduction

Salicylanilide is a kind of antimycotic antibacterial agent which has been widely applied in industrial products such as plastics, leather products, coatings, fibres and cosmetics. Halogenated salicylanilide is formed via hydrogen replacement with a halogen on the benzene ring of salicylanilide. It has a more effective antimycotic ability and antibacterial ability, therefore results in a more considerable application.^1–6 However, halogenated salicylanilides are harmful and may threaten human health. For example, they have been found to be closely related to the photosensitized reactions, the use of cosmetics containing such compounds can cause allergy or severe skin disease.^7,8 In the field of cosmetics, the use of halogenated salicylanilides has been explicitly and strictly stipulated in Cosmetic Handbook promulgated by U.S. Food and Drug Administration (FDA),⁹ EU Cosmetics Regulations (EC 1223/2009)¹⁰ and Cosmetics Health Standards of China,¹¹ tribomsalan, dibromsalon, metabromsalon and 3,3′,4′,5-tetrachlorosalicylanilide, were classified as forbidden components in cosmetics particularly. However, to the authors' best knowledge, no literature have been reported for the determination of halogenated salicylanilides in cosmetics. Therefore, appropriate and feasible analytical method to determine the concentrations of halogenated salicylanilides in cosmetics is extremely required, which is also of importance for monitoring and controlling the quality and safety of cosmetics.

The research about halogenated salicylanilides mainly focused on the drug resistance and ultraviolet (UV) light conversion mechanism, while very little reports involving the determination of them.^12,13 Cukor, et al.⁵ developed a high performance liquid chromatography-ultraviolet detection (HPLC-UVD) method to determine ramifications of halogenated salicylanilides in soap and sanitizer, however, the pretreatment was unperfect and the target analytes involved was not comprehensive enough. Another method, based on liquid chromatography tandem mass spectrometry (LC-MS/MS), was established by Caldow, et al.¹⁴ to determine salicylanilide anthelmintics residuals in bovine kidney, the investigated compounds contained oxyclozanide, closantel, rafoxanide, niclosamide and salicylanide. In a study conducted by Sakamoto et al.¹⁵ The salicylanilide anthelmintics residuals such as nitroxynil, oxyclozanide, and tribromsalan in milk were firstly determined by using LC-MS/MS method. It's worth noting that separation were performed on C18 chromatographic column and ionization was finished through electrospray ion (ESI) source in negative mode with selected reaction monitoring (SRM) acquisition mode both in researches of Caldow, et al.¹⁴ and Sakamoto, et al.¹⁵ However, the target compounds analyzed by these two methods are quite different with the seven halogenated salicylanilides (tribomsalan, dibromsalon, metabromsalon, tetrachlorosalicylanilide, 5′-bromo-4′-chlorosalicylanilide, 4′-bromosalicylanilide, 5′-chlorosalicylanilide) investigated in this work. As can be seen in Fig. 1, all these compounds have the ultraviolet absorption groups (e.g. π–π conjugated double bond and –C=O), which can be detected by ultraviolet detector (UVD) or diode array detector (DAD). Moreover, these compounds are sensitive to ultraviolet light and can absorb photons to reach the excited state thus fluorescence would be emitted, therefore, fluorescence detector (FLD) can also be used to detect these analytes. However, the sensitivity of UV method is relatively lower than that of LC-MS/MS. In addition, both UV and FLD have the defects of relatively imprecise qualitation. Fortunately, the chemical structures of these seven halogenated salicylanilides are similar to those of salicylanilide anthelmintics,^14,15 for example, both of them have –OH. They can be ionized by ESI source in negative mode of triple quadrupole mass spectrometry and parent ion [M − H]⁻ was formed, thus dissociation and rearrangement were performed in MS/MS because of collision energy and the secondary ion fragments were formed. Therefore, these seven target analytes in this work can also be separated and detected by LC-MS/MS.

Fig. 1 Proposed fragmentation patterns of characteristic ions of 7 halogenated salicylanilides.

UPLC-MS/MS is one of the most important analytical techniques because of its high selectivity, specificity, accuracy and sensitivity in analyzing complicated samples by using multiple reaction monitoring (MRM) mode, which is widely used in the detection of banned and restricted substances in consumer goods.^16–19 In the present work, the extraction conditions and the UPLC-MS/MS conditions were optimized for the detection of seven halogenated salicylanilides in cosmetics. In addition, the SPE technology was applied to purify the samples aimed at the complex nature of cosmetics. Finally, a SPE purification coupled with ultra performance liquid chromatography tandem mass spectrometry (SPE-UPLC-MS/MS) (In MRM mode) method was developed and validated for simultaneous determination of seven halogenated salicylanilides.

2. Materials and methods

2.1. Chemicals, reagents and standards

Methanol, dichloromethane, acetonitrile, n-hexane and acetone (HPLC grade) were obtained from Merck (Darmstadt, Germany). Formic acid of HPLC grade (purity: 98–100%) and chromatographic quality ammonium acetate were purchased from Fluka (Buchs, Switzerland). The other reagents were of analytical pure and provided by Guangzhou Chemical Reagent Factory (Guangzhou, China). Ultrapure water (18.2 MΩ) was generated using a Milli-Q system (Millipore, Bedford, USA) and used in all the experiments. Standards of tribomsalan (purity: 99.0%), dibromsalon (purity: 99.0%), metabromsalon (purity: 99.0%) and 3,3′,4′,5-tetrachlorosalicylanilide (purity: 99.0%) were purchased from Dr Ehrenstorfer GmbH (Augsburg, Germany), the other standards, 5′-bromo-4′-chlorosalicylanilide (purity: 98.0%), 4′-bromosalicylanilide (purity: 98.0%) and 5′-chlorosalicylanilide (purity: 98.0%), were obtained from CNW Technologies (Düsseldorf, Germany).

2.2. Standard solutions

Individual standard stock solutions of all target analytes were prepared at the concentration of 1000 mg L⁻¹ by using acetonitrile. The mixed standards stock solution of 100 mg L⁻¹ was prepared by diluting the each individual standard stock solution with acetonitrile. All individual standard stock solutions and the mixed standards stock solution were stored in the dark at −18 °C prior to use. Appropriate concentrations of mixed standard work solutions were obtained by diluting the mixed standards stock solution with acetonitrile.

2.3. Instrumentation

Chromatographic separation was performed on an Acquity™ ultra performance liquid chromatography (UPLC) system (Waters Technologies, Milford, MA, USA). An AB 4000 QTRAP triple quadrupole mass spectrometer (Massachusetts, USA) was used for the detection of target analytes. Solid phase extraction (SPE) vacuum manifolds were obtained from Waters Technologies (Milford, MA, USA). A Turbo Vap™ LV enrichment workstation (Biotage Company, Uppsala, Sweden) was used for enrichment of sample solutions. The 5404R high speed centrifuge, MS3 basic vortex mixer, and Milli-Q Gradient system used in this work were obtained from Eppendorf Corp. (Hamburg, Germany), IKA Corp. (Staufen, Germany) and Millipore Corp. (Bedford, USA), respectively. Mixed anion exchange SPE cartridge (Oasis MAX, 3 mL/60 mg), hydrophilic and lipotropy reverse phase adsorption SPE cartridge (Oasis HLB, 3 mL/60 mg) and C18 SPE cartridge (6 mL/500 mg) were purchased from Waters Technologies (Milford, MA, USA). The other SPE cartridge, amine SPE cartridge (3 mL/200 mg), was obtained from Supelco Company (Bellefonte, Pennsylvania, USA).

2.4. Samples and sample preparation

Thirty cosmetic samples, including ten water class cosmetic samples (three aftershave water samples, three acne removing water samples, two toner samples and two skin perfecting water samples), five moisturizer, three shaving cream, three facial cleanser, five skin care cream and four body wash samples, were purchased from local markets and local cosmetics manufacturers of Guangzhou, China.

0.50 g cosmetic sample and 10 mL of dichloromethane were orderly added into a 10 mL stoppered colorimetric tube. Sufficient extraction was performed by vortexing the sample solution for 1 min and ultrasonic extraction for 10 min. Sample solution was then centrifuged for 3 min at 3000 rpm, and the liquid supernatant was moved to another 10 mL glass centrifuge tube and blowed to approximately drying by slow nitrogen gas in a 40 °C water-bath. After which, 5 mL of dichloromethane and n-hexane (1 : 9, v/v) mixed solvent was added to dissolve it.

The above dissolved sample solution was poured to an amino SPE cartridge which was activated by 10 mL methanol–acetone (1 : 1, v/v) mixed solvent and 5 mL n-hexane in advance. After sample solution discharged naturally, 5 mL n-hexane was used to wash the cartridge and the effluent was also discarded. Then 8 mL of 5% formic acid–acetone solution was adopted to elute and the eluant was subsequently blowed to approximately drying by slow nitrogen gas in a 40 °C water-bath. After that, the residue was dissolved by using 1.0 mL acetonitrile and the reconstituted solution was at 15 000 rpm for 3 min. Finally, the upper layer solution was transferred into a sample tube for the UPLC-MS/MS analysis.

2.5. Chromatographic conditions

Analytes were separated on a Waters UPLC™ HSS T3 column (2.1 mm × 100 mm, 1.8 μm) with the column temperature of 30 °C. The injection volume was 2 μL. 0.1% formic acid–water (A) and methanol (B) were adopted as moving phases with a flow rate of 0.30 mL min⁻¹. The gradient elution program was used to separate target analytes and it was optimized as follows: mobile phase A was 40% in the beginning and a linear gradient to 10% in 2 min was performed, then it was maintained at 10% for 4 min, 40% A was obtained from 10% A in the following 0.1 min, finally mobile phase A was maintained at 40% for the last 1.9 min, the total run time was 8.0 min.

2.6. MS/MS conditions

The electrospray ionization (ESI) source in negative ion mode was used for the MS/MS detection. Data acquisition was performed by using the multiple reaction monitor (MRM) mode. The ion source temperature was set as 400 °C, and the other parameters of the mass spectrometer were as follows: ESI: −4500 V, curtain gas pressure: 25 psi, atomization gas pressure: 50 psi, aux gas pressure: 50 psi, dwell time: 50 ms. The declustering potential and collision energy were presented in Table 1.

Table 1 The multiple reaction monitoring conditions of target compounds

Analytes	Declustering potential (V)	MRM 1 (m/z)	CE^b 1 (eV)	MRM 2 (m/z)	CE 2 (eV)
a Quantitative ion.b CE represents “collision energy”.
Tribromsalan (TBS)	−115	449.8/250.9^a	−40	449.8/171.0	−55
Dibromsalan (DBS)	−115	369.8/171.0^a	−40	369.8/81.00	−65
Metabromsalon (MBS)	−115	369.8/250.9^a	−34	369.8/171.0	−40
3,3′,4′,5-Tetrachlorosalicylanilide (TCSA)	−110	349.8/160.9^a	−36	349.8/125.03	−47
5-Bromo-4′-chlorosalicylanilide (BCSA)	−90	325.8/172.8^a	−34	325.8/126.0	−43
4′-Bromosalicylanilide (BSA)	−80	292.0/172.0^a	−33	292.0/92.9	−40
5-Chlorosalicylanilide (CSA)	−75	245.8/127.0^a	−30	245.8/92.0	−50

---

3. Results and discussion

3.1. Optimization of MS/MS conditions

The MS/MS acquisition parameters were optimized using ESI in negative ion mode by directly infusion of standard solution (500 μg L⁻¹) via the syringe pump at a flow rate of 20 μL min⁻¹ combining the mobile phases (methanol–water, 1 : 1, v/v) at a flow rate of 0.2 mL min⁻¹ via a T-fitting prior to liquid entering the ESI source. This provided a stable response during the optimization. Bromine (Br) and chlorine (Cl) both have two stable isotopes, which are Br⁻79 (50.69%), Br⁻81 (49.31%) and Cl⁻35 (75.77%), Cl⁻37 (24.23%), respectively. The corresponding isotopic peaks can be clearly observed in the mass spectrograms of them. Diagnostic fragment ions were selected and all mass-spectrometer parameters were optimized for maximum sensitivity. Table 1 lists the characteristic ions, declustering potential and collision energy for each compound. Selected product ions represented the most abundant fragments observed for each precursor at the collision energy noted. The typical MS/MS fragmentation pathways of seven halogenated salicylanilides were illustrated in Fig. 1.

3.2. Optimization of chromatographic separation conditions

Although baseline separation of UPLC chromatography is unnecessary with MS/MS detection, it will significantly improve the sensitivity and reduce the matrix effects. Therefore, chromatographic separation was performed on different columns, including Waters UPLC™ BEH C18 (50 mm × 2.1 mm, 1.7 μm), Waters UPLC™ HSS T3 (100 mm × 2.1 mm, 1.8 μm) and Phenomenex Kinetex PFP (50 mm × 2.0 mm, 2.6 μm) columns, and the peak shapes, resolutions and separation efficiencies were compared with each other. From the results, isomerides, metabromsalon and dibromsalon, were difficult to separate on the PFP column, moreover, same monitoring ion pairs of these two compounds could interfere the quantification of each other. Fortunately, metabromsalon and dibromsalon could be well separated on the C18 column, however, the interaction of silicon hydroxyl (not endcapping) and target analytes caused serious tailing of target chromatographic peaks. The silicon hydroxyl was endcapped on T3 column, which could reduce the interaction of target analytes, resulting in favorable and symmetrical peak shapes of seven compounds, the separation degree of metabromsalon and dibromsalon was bigger than 1.5 (see peak 4 and peak 5 in Fig. 2). Therefore, Waters UPLC™ HSS T3 column was chosen in this study to separate the halogenated salicylanilides.

Fig. 2 MRM chromatograms of 7 halogenated salicylanilides. (1) 5-Chlorosalicylanilide, (2) 4′-bromosalicylanilide, (3) 5-bromo-4′-chlorosalicylanilide, (4) dibromsalan, (5) metabromsalon, (6) tribromsalan, (7) tetrachlorosalicylanilide.

In this study, we also tested solvent systems to obtain the optimal separation of the target compounds. The effects of different organic phases (methanol and acetonitrile), aqueous phases (ammonium acetate buffer solution, 0.1% formic acid–water and ultrapure water) on the peak shapes and response values of seven target analytes were compared. Results indicated that effective separation of seven compounds could be performed by using both methanol and acetonitrile, however, higher separation degree of metabromsalon and dibromsalon was obtained with methanol, hence, methanol was used as organic solvent. As for aqueous phase, ionization of compounds was suppressed by adding ammonium acetate in 5 mmol L⁻¹ or 10 mmol L⁻¹ ammonium acetate solutions and thus the response values of compounds were significantly reduced. Under the pure water condition, phenolic hydroxyl groups of seven compounds were easy to ionize, leading to the trailing peaks. Although the 0.1% formic acid–water system would suppress the ionization of seven compounds under ESI in negative ion mode, the ionic condition provided by formic acid could also suppress the ionization of phenolic hydroxyl group of target analytes and effectively improve the peak shapes, thereby enhance the Signal to Noise Ratio and detection sensitivity. Therefore, the methanol and 0.1% formic acid–water system was chosen as mobile phases. The MRM chromatograms of seven halogenated salicylanilides standard solutions under the optimized UPLC-MS/MS conditions were shown in Fig. 2.

3.3. Optimization of extraction conditions

One of the major difficulties when analyzing organic compounds in cosmetic samples is the co-extraction of matrix components. And the fine dispersion of the sample prior to ultrasound-assisted extraction can improve the interactions between sample and extraction solvent, thus increasing the efficiency and reproducibility of the process. Therefore, suitable extraction solvents and dispersing and extraction means should be selected. In this study, the extraction effects of four extraction solvents (dichloromethane, methanol, acetone and acetonitrile) were compared by using complex cream cosmetics added with 20 μg kg⁻¹ mixed target compounds. From the results, with the extraction of methanol, acetone and acetonitrile, samples could be dissolved by vortexing, however, a mass of flocculent precipitate which can be explained as the co-extraction impurities were dissolved out by the concentration of extraction solution. Although the cream cosmetic samples could not be completely dissolved by dichloromethane, favorable dispersion of samples could be obtained by vortexing and demulsification of these samples could also be achieved by ultrasonic, thus could preferably extract target compounds. The relative recoveries (compared with the same concentration of matrix calibration standard solution) were calculated between 82% and 110%, in addition, less co-extraction impurities were dissolved out after the concentration of extraction solution, so dichloromethane was finally used as the extraction solvent. With respect to the emulsion and water kinds cosmetics which matrix are relatively simple, favorable extraction recoveries could also be obtained by vortexing and dispersion with dichloromethane and ultrasonic-assisted extraction.

3.4. Optimization of purification conditions

Cosmetic is a kind of complex mixture prepared and processed with various kinds of chemical raw materials. The presence of matrix interferents in sample extraction, such as lipids, can negatively affect both the chromatographic separation and the detection of the analytes. Therefore, a key issue was to minimize the lipid and other content of the extracts without reducing the analyte compounds. Therefore, a clean-up step by SPE was extremely needed. In the present work, four different SPE cartridge were tested, including mixed anion exchange SPE cartridge (Oasis MAX, 3 mL/60 mg), hydrophilic and lipotropy reverse phase adsorption SPE cartridge (Oasis HLB, 3 mL/60 mg), C18 SPE cartridge (6 mL/500 mg) based on reverse phase adsorption and amine SPE cartridge (3 mL/200 mg) based on positive phase adsorption and weak anion exchange mechanism. The experiment was performed with one of the most complex cream cosmetic sample spiked with all target compounds at the concentration of 20 μg kg⁻¹. After extraction following the previously discussed conditions, the extracts were concentrated to dryness under a gentle steam of nitrogen and then reconstituted in a suitable solvent prior to the SPE. For HLB and C18 cartridges, the residue was dissolved in 5 mL of 2% formic acid in water (v/v), which produced a very turbid dispersion likely due to the high lipid content. The dispersion completely clogged the SPE cartridge during the loading step, making this approach unsuitable for the clean-up of this kind of samples. With respect to MAX cartridge, the residue was dissolved in 5 mL of methanol–water–ammonia (50 : 48 : 2, v/v/v) solution and the reconstituted sample solution was performed on this kind of cartridge. Then it was washed by 3 mL water and 3 mL methanol, after which, 10 mL 5% formic acid–acetone was used to elute and the eluant was collected and concentrated to dryness. After that, the residue was dissolved by using 1.0 mL acetonitrile for the UPLC-MS/MS analysis. However, the target analytes were difficult to elute, even more unfortunately, the recovery of TCSA was only about 20% and those of the rest six analytes were no more than 70%. Therefore, MAX cartridge is not suitable for the purification of these compounds. As the aliphatic ammonia propyl was bonded to the surface of silica filler, amine cartridge (NH₂ cartridge) is often used as positive phase adsorbent in non-aqueous media, so the extract residue was reconstituted in 5 mL dichloromethane and n-hexane (1 : 9, v/v). Under these conditions, NH₂ cartridge could retain halogenated salicylanilides mainly through dipole–dipole interactions and hydrogen bonding with the hydroxyl moieties. After loading the sample extract, the cartridge was rinsed with 5 mL of n-hexane, without losing any of the halogenated salicylanilides. Elution of halogenated salicylanilides from the SPE cartridge was studied with different solvents with the goal of obtaining a high level of selectivity paired with good recoveries. Elution of halogenated salicylanilides was assessed with each 8 mL of the following eluents: methanol, dichloromethane, acetone and acetone/formic acid (95 : 5, v/v). As shown in Fig. 3, methanol, dichloromethane and acetone provided recoveries less than 75% for all halogenated salicylanilides, which were completely retained on the NH₂ sorbent. When elution was carried out with acetone containing 5% formic acid (v/v), favorable recoveries were obtained for all the analytes, indicating that the presence of this acid is able to disrupt the interactions of halogenated salicylanilides with the NH₂ phase. This behavior suggests that strong electron-withdrawing groups (halogen) of the seven halogenated salicylanilides made the electron cloud on the hydroxide radical shift to the benzene ring, which could facilitate the ionization of the hydroxide radical. In addition, the retention on the NH₂ sorbent is not only based on hydrogen bonding and dipole–dipole interactions, but also on ionic interactions. Although in non-aqueous media NH₂ normally acts as a normal-phase sorbent, it can also behave as a weak anion exchanger, which could be able to retain analytes holding a net negative charge. Thus, it was hypothesized that under the studied conditions halogenated salicylanilides are at least partially ionized. The addition of formic acid neutralizes them, allowing them to elute from the cartridge.

Fig. 3 Effect of different elution solvents on the recovery of halogenated salicylanilides during SPE with NH₂.

3.5. Method validation

Validation of the proposed method was performed based on the recommendations of the Eurachem guide on analytical method validation²⁰ and Commission Decision 2002/657/EC establishing criteria and procedures for the validation of analytical methods to ensure the quality and comparability of analytical results generated by official laboratories.²¹

3.5.1. Linearity and limits of detection (LOD). The linearity of the method was tested by using standard solutions of eight concentration levels evenly distributed over the range of 0.5–200 μg L⁻¹ (Table 2). Each concentration level was analyzed at least in triplicate. Linear regression analysis was performed by plotting the chromatographic peak area (on the ordinate (y)) of each target analyte versus the corresponding mass concentration (on the abscissa x (μg L⁻¹)), and thus the regression equations with calculated correlation coefficients (R²) greater than 0.997 were obtained, indicating that the linear relations were favorable within the seven halogenated salicylanilides.

Table 2 Linear regression equations, linear ranges, R² and the LOD and LOQ of seven target analytes

Analyte	Linear range (μg L^−1)	Regression equation	R^2	LODs (μg kg^−1)	LOQs (μg kg^−1)
TBS	1.0–200	y = 17500x + 26 400	0.9978	0.8	2.5
DBS	0.5–100	y = 15200x + 52 600	0.9988	0.5	1.5
MBS	0.5–100	y = 10200x + 30 200	0.9988	0.5	1.5
TCSA	0.5–100	y = 21500x + 41 000	0.9992	0.5	1.5
BCSA	1.0–200	y = 37900x + 155000	0.9986	0.8	2.5
BSA	1.0–200	y = 40300x + 137000	0.9992	0.8	2.5
CSA	1.0–200	y = 98300x + 468000	0.9974	0.8	2.5

---

Moreover, LODs were estimated from a composite cream sample spiked at low decreasing concentration levels. LODs were calculated as the average concentrations of compound producing a signal-to-noise (S/N) ratio of three using the less sensitive MS/MS transition (MRM2), i.e. the one permitting the unambiguous identification of the analytes. In addition, the limits of quantitations (LOQs) were estimated as the average concentrations of compound producing a S/N of 10 using the most sensitive MS/MS transition (MRM1), provided that the S/N for MRM2 was at least of three. The proposed method provided low LODs, which were in the range of 1.5–2.5 μg kg⁻¹.

3.5.2. Selectivity. The selectivity of the developed method was evaluated via the analysis of procedural blank samples, blank cosmetic samples and different cosmetic samples spiked at 5.0 μg kg⁻¹. MRM chromatograms obtained for quantifier and qualifier MS/MS transitions were checked for co-eluting interferents at the retention times of the corresponding analytes. No interferents were observed at the retention times of analytes ±0.1 min in any of the transitions, indicating the high selectivity of this method.

3.5.3. Matrix effects (MEs). It has been accepted that matrix effect (suppression or enhancement) of the analyte signal is a common phenomenon in ESI analysis and should be properly evaluated during method validation. In the present work, MEs were assessed by the post-extraction addition method, which is based on the comparison of the responses obtained for a spiked extract with those obtained for a standard solution at the same concentration. The percent matrix effect (%ME) was calculated as the following equation:

ME (%) = (R_se/R_std − 1) × 100

where R_se is the response of the analyte in the spiked extract and R_std is the corresponding response in the standard solution.^22,23

In this work, a negative result indicates ionization suppression, where as a positive result indicates signal enhancement. These experiments were performed using a very complex cosmetic sample extract spiked at three concentration levels equivalent to 2.5, 5.0 and 25.0 μg kg⁻¹. The results were presented in in Fig. 4. It had been well observed from Fig. 4 that MEs ranged from about −13.7% for TBS at 2.5 μg kg⁻¹ to 5.5% for CSA at 25.0 μg kg⁻¹. Among the studied compounds, four out of seven (TCSA, BCSA, BSA and CSA) presented MEs within ±10%, and only DBS showed absolute MEs higher than 10%. The observed %MEs can be considered satisfactory, especially when considering the complex nature of the analyzed samples.

Fig. 4 Matrix effects (ME%) of seven analytes in a composite cosmetic sample (n = 6).

3.5.4. Accuracy and precision. Negative samples, including moisturizer sample and skin water sample, at three spiked levels of analyte with low (2.5 μg kg⁻¹), medium (5.0 μg kg⁻¹) and high (25.0 μg kg⁻¹) of mixed standards were used to analysis the recoveries and intra-day precision of analytes according to the proposed method, with six identical samples determined at each concentration. In addition, the inter-day precision was also investigated by analyzing five spiked replicates for medium level (5.0 μg kg⁻¹). The results indicated that the recoveries of the seven target analytes were satisfactory with values in the range of 83.9–107% (Table 3). Moreover, relative standard deviations (RSDs) of 2.9–6.5% for intra-day precision (n = 6) and 4.9–7.0% for inter-day precision (n = 5) were observed, which means that the accuracy, precision and stability can meet the requirements for such an analysis.

Table 3 Summary of mean recoveries, relative standard deviations (RSDs, in parentheses) of seven analytes

Analytes	Added (μg kg^−1)	Intra-day (n = 6), recovery (%RSD)		Inter-day (n = 5), recovery^a (%RSD)
		Moisturizer	Skin water	Moisturizer	Skin water
a Spiked level was 5.0 μg kg^−1.
TBS	2.5, 5.0, 25.0	85.6(3.1), 101(4.1), 92.5(5.8)	83.9(3.8), 91.2(3.8), 104(4.8)	93.7(6.1)	94.1(4.9)
DBS		90.5(4.9), 91.1(3.2), 87.5(3.9)	85.0(4.3), 92.5(3.9), 93.4(4.2)	90.3(5.4)	91.1(5.0)
MBS		92.4(4.7), 93.5(5.2), 98.1(6.3)	82.1(3.9), 91.6(5.5), 95.9(4.6)	92.1(6.5)	90.7(5.7)
TCSA		88.2(3.8), 91.3(5.4), 101(4.2)	86.8(3.8), 92.1(5.6), 104(6.2)	89.6(5.9)	90.2(6.0)
BCSA		86.7(4.1), 93.6(4.7), 93.1(5.6)	86.0(4.0), 96.9(5.4), 106(3.9)	90.5(5.5)	93.8(5.9)
BSA		89.9(4.0), 93.4(3.9), 104(5.4)	85.5(4.1), 101(4.9), 99.4(4.5)	91.4(5.3)	97.0(6.2)
CSA		88.2(2.9), 91.3(4.2), 107(6.5)	85.5(5.1), 91.8(6.2), 102(3.8)	89.2(6.7)	89.9(7.0)

---

3.6. Analysis of practical samples

The method established in this work was adopted to determine a total of thirty samples collected from local markets, including cream samples, skin water samples and emulsion samples. TCSA (422 μg kg⁻¹) was detected in an emulsion sample and DBS (584 μg kg⁻¹) was found in a shampoo sample. The TCSA selected ion chromatogram of the emulsion sample and the DBS selected ion chromatogram of the shampoo sample were all shown in Fig. 5.

Fig. 5 The selected ion chromatogram of DBS and TCSA of two typical samples.

To test the accuracy of the developed method in real samples, two different cosmetic products (emulsion sample and shampoo sample) were fortified at two concentration levels (roughly 250 and 500 μg kg⁻¹) and analyzed by the optimized SPE-UPLC-MS/MS, taking into account the known analyte contents for these samples. The results showed a mean recovery ± SD of 93.8 ± 5.6% (n = 48).

4. Conclusions

A simple and rapid UPLC-MS/MS method was developed and validated for the determination of seven halogenated salicylanilides in water-class and cream-class cosmetics. The method enables these compounds to be simultaneously quantified in no more than 7 min. The extraction conditions and the UPLC-MS/MS conditions were optimized. In addition, the SPE technology was applied to purify the samples aimed at the complex nature of cosmetics. Finally, the target analytes were separated by a T3 column and detected through ESI source in a negative mode with MRM condition. The validation data exhibited good linearity, accuracy and precision. The LODs and LOQs of the method indicated that almost trace levels of halogenated salicylanilides could be detected. Most importantly, the matrix effect of the developed method is very low. These advantages can satisfy the determination requirements for the banned and restricted substances in cosmetics domestic and overseas.

Acknowledgements

This work was supported by the Science and Technology Project of General Administration of Quality Supervision, Inspection and Quarantine of China (2014QK046). At the same time, the authors would like to thank all the workers for sampling, sample preparation and measurement.

References

1. J. Vinšova and A. Imramovsky, Salicylanilides: still a potential antibacterially active group, Ceska Slov. Farm., 2004, 53(6), 294–299.
2. D. L. Fluente, R. Sonawane, N. D. Arumainayagam, D. Verman and A. S. British, Small molecules with antimicrobial activity against E. coli and P. aeruginosa identified by high-throughput screening, Br. J. Pharmacol., 2006, 149(5), 551–559.
3. M. K. Dahlgren, A. M. Kauppi and O. Ing-Marie, et al., Design, synthesis, and multivariate quantitative structure–activity relationship of salicylanilides-potent inhibitors of type III secretion in Yersinia, J. Med. Chem., 2007, 50(24), 6177–6188.
4. A. Imramovsky, J. Vinšova and J. M. Ferriz, et al., Salicylanilide esters of N-protected amino acids as novel antimicrobial agents, Bioorg. Med. Chem. Lett., 2009, 19(2), 348–351.
5. P. Cukor, C. Persiani, M. Woehler and N. Bischak, High-performance liquid chromatographic determination of halogen-substituted salicylanilides, J. Chromatogr. A, 1978, 10(266), 496–500.
6. Lever Bros Co, Halogenated phenoxy-substd salicylanilides useful as germicides particularly in soap, US3674850-A NL136091-B, 1972.
7. F. P. Jenkins, D. Welti and D. Baines, Photochemical reactions of tetrachlorosalicylanilide, Nature, 1964, 201(4921), 827–828.
8. L. C. Harber, H. Harris and R. L. Baer, Allergic photocontact dermatitis to salicylanilides and related compounds, J. Invest. Dermatol., 1966, 46, 303–305.
9. FDA/IAS, Cosmetic Handbook, U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition, 1992, http://www.fda.gov.
10. Regulation (EC) No 1223/2009 of the European Parliament and of the Council of 30 November 2009 on Cosmetic Products, Off. J. Eur. Commun., 2009.
11. Ministry of Health of the People’s Republic of China, Hygienic Standard for Cosmetics, 2007.
12. M. Kratky, J. Vinsova and V. Buchta, et al., New amino acid esters of salicylanilides active against MDR-TB and other microbes, Eur. J. Med. Chem., 2010, 45(12), 6106–6113.
13. M. Yoshimura, H. Hosokawa and T. Horio, et al., Persistent light reaction due to halogenated salicylaniledes, Arch. Dermatol., 1966, 94(4), 522–526.
14. M. Caldow, M. Sharman, M. Kelly, J. Day, S. Hird and J. A. Tarbin, Multi-residue determination of phenolic and salicylanilide anthelmintics and related compounds in bovine kidney by liquid chromatography-tandem mass spectrometry, J. Chromatogr. A, 2009, 1216, 8200–8205.
15. M. Sakamoto, K. Takeba, T. Sasamoto, T. Kusano, H. Hayashi, S. Kanai, M. Kanda and T. Nagayama, Determination of Bithionol, Bromophen, Nitroxynil, Oxyclozanide, and Tribromsalan in milk with liquid chromatography coupled with tandem mass spectrometry, J. AOAC Int., 2010, 93(4), 1340–1346.
16. Y. Xian, Y. Wu, X. Guo, Y. Lu, H. Luo, D. Luo and Y. Chen, Simultaneous determination of 11 restricted dyes in cosmetics by ultra high-performance liquid chromatography/tandem mass spectrometry, Anal. Methods, 2013, 5, 1965–1974.
17. X. Guo, Y. Xian, H. Luo, Y. Wu, D. Luo, Y. Chen, Y. Lu and D. Xu, Quantitative determination of seven fluorescent whitening agents in polystyrene and polyvinyl chloride plastics using ultra-high performance liquid chromatography/tandem mass spectrometry, Anal. Methods, 2013, 5(21), 6086–6093.
18. X. Guo, T. Mu, Y. Xian, D. Luo and C. Wang, Ultra-performance liquid chromatography tandem mass spectrometry for the rapid simultaneous analysis of nine organophosphate esters in milk powder, Food Chem., 2016, 196, 673–681.
19. H. Dong, X. Guo, Y. Xian, H. Luo, B. Wang and Y. Wu, A salting out-acetonitrile homogeneous extraction coupled with gas chromatography-mass spectrometry method for the simultaneous determination of thirteen N-nitrosamines in skin care cosmetics, J. Chromatogr. A, 2015, 1422, 82–88.
20. The Fitness for Purpose of Analytical Methods – A Laboratory Guide to Method Validation and Related Topics, ed. B. Magnusson and U. Örnemark, Eurachem Guide, 2nd edn, 2014, ISBN 978-91-87461-59-0, http://www.eurachem.org.
21. EU, Commission Decision 2002/657/EC of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results, Off. J. Eur. Communities: Legis., 2002, 221, 8–36.
22. E. Chambers, D. M. Wagrowski-Diehl, Z. Lu and J. R. Mazzeo, Systematic and comprehensive strategy for reducing matrix effects in LC/MS/MS analyses, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2007, 852, 22–34.
23. J. Regueiro, A. E. Rossignoli, G. Álvarez and J. Blanco, Automated on-line solid-phase extraction coupled to liquid chromatography-tandem mass spectrometry for determination of lipophilic marine toxins in shellfish, Food Chem., 2011, 129, 533–540.

---

Footnote

† Current address: No. 1-2, Zhujiang Road, Chaotian Industrial Zone, Panyu District, Guangzhou Guangdong, China, Fax: +86-20-82022322, Tel: +86-20-82022322.

---