Solvent extraction and gas chromatography–mass spectrometric determination of probable carcinogen 1,4-dioxane in cosmetic products

In the present work, a method based on solvent extraction and gas chromatography–mass spectrometry (GC-MS) has been validated for the determination of 1,4-dioxane in cosmetics. Various solvents including ethyl acetate, hexane, methanol, dichloromethane and acetone have been used for the extraction of 1,4-dioxane, among them the ethyl acetate was found to be the most efficient extracting solvent. This method has offered excellent quality parameters for instance linearity (R2 > 0.9991), limit of detection (LOD, 0.00065-0.00091 µg/mL), limit of quantification (LOQ, 0.00217–0.00304 µg/mL) and, precision intra-day (1.65–2.60%, n = 5) and inter-day (0.16–0.32%, n = 5) in terms of relative standard deviation (RSD%). A total of thirty-nine cosmetic samples of different brands and origin have been studied. Among them, the 1,4-dioxane was found in twenty-three samples (FB1-FB7, MC1-MC4, MC6-MC8, HS3, HS5, BL1-BL3, BL5 and PLD1-PLD3) at the levels between 0.15 µg/mL and 9.92 µg/mL, whereas in sixteen samples (MC5, HS1, HS2, SG1-SG5, BL4 and HP1- HP6) was found to be not detected. The recovery values were achieved between 93% and 99% in both low and high level of spiked samples. In comparison to the traditional analytical techniques, the proposed method was found to be very sensitive and cost-effective for the routine analysis of 1,4-dioxane at low concentration in cosmetics.

The objective of the present work was to develop an analytical method based on low-cost solvent extraction and GC-MS for the analysis of 1,4-dioxane in a broad range of commercially available cosmetics. The method has been demonstrated to be proficient of measuring 1,4-dioxane at trace level lower than the recommendations value 28,29 . The overall high throughput including sensitivity, recovery and precision presented by the proposed technique can be an advantage for this kind of investigation.

Results and discussion
Optimization of solvent extraction method. The assay of 1,4-dioxane in cosmetics was carried out by solvent extraction and GC-MS methods, respectively. Relating to the optimization of extracting solvents, initially, water was chosen as extracting solvent but 1,4-dioxane was found to have higher solubility due to the low octanol/ water partition coefficient, and thus it was very challenging to extract 1,4-dioxane in aqueous medium using solvent extraction method. To sort out this problem, we have selected various pure extracting solvents for instance ethyl acetate, hexane, methanol, dichloromethane and acetone, and attempted to vary many parameters to extract 1,4-dioxane from cosmetics. The extraction procedures have been described as a sample extraction procedure section, the identical extraction procedures were applied for each extracting solvent. Among the tested extracting solvents, the ethyl acetate, hexane, methanol, dichloromethane and acetone have offered 1,4-dioxane recovery values of 99%, 47%, 70%, 89% and 35%, respectively. Among them, the ethyl acetate was found to be the most efficient extracting solvent where the 1,4-dioxane recovery value reaches up to 99%, nonetheless, the acetone offered the least 1,4-dioxane recovery value (35%). In previously reported study, the authors had applied combined liquid-liquid extraction (dichloromethane as extracting solvent) and GC-MS method 30 for the identification of 1,4-dioxane in water samples, obtained excellent recovery up to 95%. Comparatively, the sample matrix was somewhat dissimilar but the recovery values were found to be in good agreement with the values obtained in current study. Nevertheless, the chlorinated solvents should be avoided to use as extracting solvent due to the environmental concern. Thus, the ethyl acetate was selected as an extracting solvent for the extraction of 1,4-dioxane in cosmetics. The enhancement, in the extraction of the target compound, can be because of the diverse composition of cosmetics. The effect of extracting solvents on recovery values has been demonstrated in Fig. 1. The Fig. 1 comprises a standard error bar of the mean which is the estimated values of the standard deviation (s) of means of analyzed samples, this is calculated using formula s/√(n), where n corresponds to the number of replicates (n = 5).

Method validation.
In order to assess the accurateness of the proposed solvent extraction-GC/MS procedure for the identification of 1,4-dioxane in cosmetics, the linearity, limits of detection (LOD, signal-to-noise ratio of 3:1) and limits of quantification (LOQ, signal-to-noise ratio of 10:1), precision (intra-day and inter-day) and accuracy were studied. The linearity of the method was investigated at six concentrations between 0.005 µg/ mL and 50 µg/mL. This higher range was chosen due to 1,4-dioxane which is expected to be detected at low level (ng/L) to higher level µg/mL 2 . The linearity of the system was investigated by constructing calibration plot of the ratio of the peak area value of target compound 1,4-dioxane to the peak area value of cyclohexanone vs. ratio of the amounts of 1,4-dioxane and cyclohexanone. The linear range, slope, intercept and correlation coefficient (R 2 ) of 1,4-dioxane have been illustrated in Fig. 2.
Sensitivity in terms of LOD and LOQ of the method was measured from cosmetic samples (MC 5 , HS 1 , HS 2 , HS 4 , SG 1 , SG 2 , SG 3 , SG 4 , SG 5 , BL 4 , HP 1 , HP 2 , HP 3 , HP 4 , HP 5 ) free from 1,4-dioxane. These samples were spiked with low amounts (0.001 µg/mL) of 1,4-dioxane at the beginning of the extraction process. The achieved results have been illustrated in the Table 1. LOD of the solvent extraction-GC/MS in these cosmetic samples were achieved between 0.00065 µg/mL to 0.00091 µg/mL, whereas the LOQ obtained was between 0.00217 µg/mL and 0.00304 µg/mL ( Table 1). The achieved outcomes show that the offered solvent extraction-GC/MS technique is appropriate for the determination of 1,4-dioxane in different kinds of cosmetic samples even at very low concentrations. The system precision was assessed by the intra-day and inter-day analysis. The intra-day analysis discusses to the determination of five samples injection in the same day whereas inter-day analysis discusses the determination of five samples injection over three consecutive days at low (0.3 µg/mL) and high (1.0 µg/mL) concentrations. The excellent intra-day (1.65-2.60%, n = 5) and inter-day (0.16-0.32%, n = 5) values in terms of RSD% were achieved. The precision of the system has been presented in Table 2.   Table 1. The obtained LOD and LOQ values of 1,4-dioxane in cosmetic samples. LOD, limit of detection (signal-to-noise, 3:1); LOQ, limit of quantification (signal-to-noise, 10:1); *1,4dioxane was not found in these samples.
The outcomes have revealed that the present method can be proposed for the routine analysis of 1,4-dioxane in cosmetic samples at very low concentrations. Until now, a few of literatures relating to the determination of 1,4-dioxane in cosmetics are available. In one of the study, the author has applied combined method based on headspace-GC-MS for the analysis of 1,4-dioxane in cosmetic products comprising polyethoxylated surfactants, found higher LOD value (0.3 µg/mL) 27 . Nevertheless, in another studies, the authors have applied combined headspace-SPME-GC-MS method for the analysis of 1,4-dioxane in nonionic surfactants and cosmetics, and SPE-GC-flame ionization detector (FID) detecting identification of 1,4-dioxane in cosmetic raw materials and finished cosmetic products, respectively, where the authors also detected higher LOD values ranged from 0.06 to 0.51 µg/mL 5 . Alternative method based on SPE-HPLC-UV detector using LiChrospher CH-8 reversed-phase column was also applied for the analysis of 1,4-dioxane in cosmetic samples, where the lowest measurable concentration was found to be 6.5 µg/g 24 . In comparison to the LOD value (0.00065 µg/mL) obtained in our work, the available traditional method offered lower sensitivity including solvent consumption, expensive materials and it was time consuming. As a result, the proposed method have offered advantages over sensitivity, rapidness, faster and inexpensive, and therefore the method could be applied as an advance technique for the routine analysis of 1,4-dioxane at trace level in cosmetics. Furthermore, there are other literatures available about 1,4-dioxane identification in water samples by means of numerous analytical methods for instance SPE-GC-MS 6,10,21 , and headspace-SPME-GC-MS 22 . Relatively, the outcomes achieved in the present work were also found in good agreement with the results obtained in water samples 6,10,22 , and thus the proposed method could also be applied for the analysis of 1,4-dioxane in water samples.
Application to cosmetic samples. The developed procedure in the present work was practically applied to examine the1,4-dioxane in commercially available cosmetic samples in Saudi Arabia. A total of thirty-nine cosmetic samples of different brand and origin have been investigated. The achieved results including concentrations and recovery values of 1,4-dioxane in cosmetic samples have been presented in Table 3. The 1,4-dioxane was found in most of the analyzed samples at levels between 0.15 to 9.92 µg/mL whereas in some samples (MC 5 , HS 1 , HS 2 , SG 1 -SG 5 , BL 4 and HP 1 -HP 6 ) the 1,4-dioxane found to be not detected. Recovery values were achieved from 93-98% at low level spiked concentrations and 95 to 99% at high level spiked concentrations. Relatively lower recovery values were obtained in low level spiked concentrations, the cause may be due to the samples complexity and higher loss during the sample extraction procedure. The amounts of 1,4-dioxane in facial and body scrub were obtained between 0.29 µg/mL and 9.92 µg/mL whereas in moisturizing cream (not detected-0.88 µg/ mL), hair shampoo (not detected-0.23 µg/mL), shower gel (not detected), body lotion (not detected-0.16 µg/mL), hand soap (not detected) and powder laundry detergent (0.15 µg/mL and 0.21 µg/mL) were obtained. Among the analyzed samples, facial and body scrub, and powder laundry detergent produced 1,4-dioxane in all brand with highest concentration up to 9.92 µg/mL particularly in FB 4 sample, however, the lowest concentration was obtained in PLD 1 , PLD 2 and PLD 5 samples up to 0.15 µg/mL. In shower gel and hand soap sample the 1,4-dioxane has not been identified in any brands. To demonstrate the outcomes, the GC-MS chromatogram of 1,4-dioxane (retention time (RT), 4.78) and cyclohexanone (RT, 6.03) of facial and body scrub (FB 3 , Perfect Cosmetics) sample have been displayed in Fig. 3.
It can be seen that the excellent resolution, symmetry and sensitivity was attained in the analysis, no tailing and interfering peaks were perceived at the same retention time of the studied analyte in cosmetic samples. Fuh and his coworkers have studied the presence of 1,4-dioxane in shampoo, liquid soap and dish washing detergent samples. The 1,4-dioxane was detected in shampoo samples at concentrations ranged from 11.5 to 41.1 µg/mL, whereas in liquid soap and dish washing detergent, the 1,4-dioxane was detected at concentrations ranged from 7.8 to 6.5 µg/ mL 5 . These values were found in good agreements with the values obtained in the current study. In another work, Rastogi (1990) identified 1,4-dioxane in 82% of the studied cosmetic samples at amounts ranged from 0.3 to 96 µg/ mL, and 85% of the dish washing samples contained 1,4-dioxane between 1.8 µg/mL and 65 µg/mL 27 . The variation in the concentrations was might be due to the different composition of the products, brand and origin. The results obtained from the current study are the substantial data source related to the presence of 1,4-dioxane in cosmetic samples of various brand and origin that are commercially available in the Saudi Arabian markets. The 1,4-dioxane could be primarily originated from the contaminations of nonionic surfactants through production. The cosmetic products are being worthy of further concerns, as they are extensively applied in our everyday existence. Table 4 displays the comparison of analyzed 1,4-dioxane in cosmetics from Saudi Arabia and with the earlier reported data. The LOD value in the present study was significantly enhanced as compared to those LOD values obtained in traditional extraction methods. Furthermore, the current method has also offered shorter analysis time, excellent precision and recovery values. According to the European Union directives on the safety of cosmetics 31,32 , 1,4-dioxane must not be applied in their preparations. Nevertheless, 1,4-dioxane was remains identified in cosmetic products in Saudi Arabia.

conclusions
A low cost, sensitive and reliable procedure based on solvent extraction and GC/MS has been developed for the determination of 1,4-dioxane in various cosmetic samples. Many solvents (methanol, hexane, dichloromethane and ethyl acetate) have been applied for the extraction of 1,4-dioxane, among them ethyl acetate was found to be the most efficient extracting solvent. A total of thirty-nine cosmetic samples of different brands and origin have www.nature.com/scientificreports www.nature.com/scientificreports/ been investigated, 1,4-dioxane was detected in most of the analyzed samples at concentrations ranged from 0.15 to 9.92 µg/mL, while in some samples (MC 5 , HS 1 , HS 2 , SG 1 -SG 5 , BL 4 and HP 1 -HP 6 ) 1,4-dioxane found to be not detected. The excellent quality parameters including recovery values (up to 99%) were achieved. The applied technique illustrates insignificant matrix effect and the lowest sample pretreatment needed deprived of losing target analyte. The inexpensive solvent extraction procedure implemented through the experimentation is a substantial benefit since during the sampling procedure it reduces the analysis period and the possibility of negligible losses of the studied compound and thus assists to offer enhanced method performance. The obtained results including excellent quality parameters and negligible matrix influences have made possible the applicability of the offered procedure as a new method for the routine examination 1,4-dioxane in cosmetic products.

Materials and Methods
Chemicals and reagents. The solvents and chemicals used in the current investigation were of analytical or liquid chromatography grade, obtained from Merck (Darmstadt, Germany). 1,4-dioxane and cyclohexanone were supplied from Sigma Aldrich (St. Louis, USA). Sodium sulfate anhydrous was purchased from Merck (Darmstadt, Germany). For sampling procedure, ultrapure water was produced using water filtration system (Milli-Q, Millipore Corporation, Bedford, USA). The stock standard solution of 1,4-dioxane was prepared in methanol at level 1000 µg/mL and applied for additional dilution methods. The IS solution was prepared in methanol at concentration 5 µg/mL and added to calibration solutions and samples during analysis. To verify the linearity (R 2 ) of the procedure, 1,4-dioxane standard at various levels (0.005-50 µg/mL) were made using weight by weight. 1,4-dioxane standard solutions and cosmetics were filtered through syringe filter (0.22 μm) (PTFE, Macherey-Nagel GmbH, Düren, Germany) before being injected into the GC-MS system. Sample extraction procedure. Cosmetic samples of various brands and origin were purchased from cosmetic and pharmacy shops located in Saudi Arabia. Cosmetic samples were refrigerated between 2-4 °C and investigated within one week to evade any microbial contamination. For the extraction of 1,4-dioxane from  www.nature.com/scientificreports www.nature.com/scientificreports/ cosmetic samples, 1 g of cosmetic sample was accurately weighed in a 40 mL glass tube followed by the addition of 5 mL ethyl acetate. The sample solution was thoroughly mixed by means of vortex mixer (Jeio Tech, Seoul, South Korea) for 1 min. Afterward, the sample was moved to ultrasonic bath for sonication (10 min) at room temperature, sample centrifugation was carried out by HERMLE, model Z32 HK centrifuge system (Wehingen, Germany) at 6000 rpm for 8 min. The sample supernatant was taken out and moved to separating funnel followed by the addition of sodium sulfate to extract water contents. Finally, the sample was collected in a new glass tube and solvent was evaporated by means of nitrogen gas. Then, the extract of the sample was redissolved with 1 mL methanol (containing IS, 5 µg/mL) standard followed by mixing using vortex system. The 1 mL sample solution was filtered by means of PTFE syringe filter (0.22 μm) and injected into the GC-MS system. The sample injection volume was 1 µL to the splitless mode.
To evaluate the effectiveness of the extraction method and preventing the effect of matrix influences on retention time, peak intensity and shape, the quantification of 1,4-dioxane was carried out by a standard addition method comprising two non-spiked samples at zero levels and three spiked samples at 50% (0.1 µg/mL, values representing the increase of 1,4-dioxane in the sample after spiking), 100% (0.2 µg/mL) and 500% (1.0 µg/mL). The samples spiking was performed at the beginning of each extraction procedure. Samples were examined in triplicates, and the statistical analysis was carried out using ANOVA method.
The LOD and LOQ of the procedure were assessed in cosmetic samples (MC 5 , HS 1 , HS 2 , HS 4 , SG 1 , SG 2 , SG 3 , SG 4 , SG 5 , BL 4 , HP 1 , HP 2 , HP 3 , HP 4 , HP 5 ) free from 1,4-dioxane. These samples were spiked with low amounts (0.001 µg/mL) of 1,4-dioxane before being analyzed by the optimized method. The recovery values of 1,4-dioxane were estimated at low and high concentrations in cosmetic samples, achieved from the added and found 1,4-dioxane concentrations. The quality control parameters of the system were also studied in each sample lot to confirm that the contamination of the samples did not come out and the sensitivity of the system was steady throughout the study.
Gas chromatography-mass spectrometric conditions. GC-MS analysis was carried out on a gas chromatograph (TRACETM 1310 GC) equipped with single quadrupole mass spectrometer (ISQLT) and auto-sampler unit model AI/AS1310 (Thermo Scientific, Waltham, USA). Separation of 1,4-dioxane was carried out on analytical Rxi-624Sil MS column with dimension 60 m, 0.53 mm I.D., 3.0 µm thickness (Restek, USA). The applied temperature program was started at 40 °C for 3 min, then rising to 240 °C at a rate of 50 °C/min (10 min). The flow-rate of helium (carrier gas) was controlled to get 100 kPa. The temperatures of injector, transfer line and ion source were adjusted to 250 °C. Ion source was established in electron ionization mode with 70 eV. The sample injection volume was 1 µL and operated to the splitless mode. The identification of the target analyte was performed using a mass spectrometer system. The data acquisition, data reporting, method setup and data processing were monitored using Xcalibur 3.1 software (Thermo Scientific, Waltham, USA).
To identify the target analytes, mass spectral database search methods was applied from National Institute of Standards and Technology (NIST, Gaithersburg, MD, USA). Gas chromatography-mass spectrometric analysis was carried out on a gas chromatograph (TRACETM 1310 GC) equipped with single quadrupole mass spectrometer (ISQLT) and auto-sampler unit model AI/AS1310 (Thermo Scientific, Waltham, USA). To identify the target analytes, mass spectral database search methods was applied from NIST. Prior to the analysis, the ionization mass spectra scan of 1,4-dioxane and cyclohexanone (IS) was carried out at concentration of 5 μg/mL. The electron ionization mass spectra scan was ranged from m/z, 20-300, and quantification was based on selected ion monitoring