Volatile methylsiloxanes in personal care products – Using QuEChERS as a “green” analytical approach

Talanta 155 (2016) 94–100

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Short communication

Volatile methylsiloxanes in personal care products – Using QuEChERS as a “green” analytical approach Daniela Capela, Vera Homem n, Arminda Alves, Lúcia Santos LEPABE – Laboratory for Process Engineering, Environment, Biotechnology and Energy, Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

art ic l e i nf o

a b s t r a c t

Article history: Received 7 March 2016 Received in revised form 11 April 2016 Accepted 14 April 2016 Available online 16 April 2016

Organosiloxanes, namely volatile methylsiloxanes (VMSs) are one of the most relevant classes of ingredients incorporated in personal care products (PCPs), such as creams and lotions, bath soaps and hair care products. Their use has caused concern among the scientific community due to their potential toxic behaviour to human health and environment. This manuscript reports the first application of QuEChERS (“Quick, Easy, Cheap, Effective, Rugged and Safe”) extraction followed by gas chromatography – mass spectrometry analysis to determine VMSs in cosmetics and personal care products. Eight VMSs, four linear (L2–L5) and four cyclic (D3–D6) were investigated in 36 samples. The validated method was able to remove the interfering matrix components, conducting to high recovery percentages (74–104%) and low relative standard deviations (o 18%). A linear behaviour was observed in the range of 0.005–2.50 mg L
1 (correlation coefficient, R2 40.996) and limits of detection ranged from 0.17 ng g
1 (L2) to 3.75 ng g
1 (L5). Matrix effects were also investigated for all analysed compounds and matrices and showed not to be significant. Global uncertainty of the proposed methodology was also estimated using a bottom-up approach being between 5% and 35% (on average). Finally, the method was satisfactorily applied to the analysis of 36 personal care products. As expected, results showed the existence of VMSs in all analysed samples in concentrations up to 754 mg g
1. D4 and D5 were more frequently detected while body moisturizers, facial creams and shampoos showed the highest levels of VMSs. & 2016 Elsevier B.V. All rights reserved.

Keywords: Volatile methylsiloxanes QuEChERS Gas chromatography-mass spectrometry Personal care products Cosmetics

1. Introduction Volatile methylsiloxanes (VMSs) are one of the most relevant classes of ingredients incorporated in personal care products (PCPs), due to their unique properties, such as high thermal stability and smooth texture [1,2]. They are extensively used as ingredients in the formulation of a wide range of cosmetic and PCPs, including creams and lotions, bath soaps, shampoos and hair care products, to soften, smooth and moisten. When siloxanes are presented in cosmetic and toiletry labels, they are seldom listed by their chemical names, but rather by their generic names “dimethicone” or “cyclomethicone” , which represents the volatile linear and cyclic methylsiloxanes, respectively [3]. Although highly insoluble in water due to their lipophilic nature, they are enough hydrolytically stable to be easily emulsified into most cosmetic preparations [4]. n

Corresponding author. E-mail address: [email protected] (V. Homem).

http://dx.doi.org/10.1016/j.talanta.2016.04.029 0039-9140/& 2016 Elsevier B.V. All rights reserved.

VMSs are incorporated in products with a high utilization rate and so, they are continuously introduced into the environment through sewer systems [5,6] or solid waste deposition in landfills [6]. Due to their physicochemical properties, they are considered lipophilic, bioaccumulative and are only partially biodegradable [7]. Therefore, when they reach the conventional wastewater treatment plants WWTPs, they are not completely removed [8]. Studies have shown potential toxic behaviour of VMSs towards human and environment causing some concern among the scientific community [1]. In fact, the European Chemical Agency considered that D4 meets the criteria for both ‘persistent, bioaccumulative and toxic’ (PBT) and ‘very persistent and very bioaccumulative’ (vPvB) substance in the environment, while D5 is considered a vPvB substance [9]. Reports also suggest that some VMSs may have oestrogenic activity, cause impairment of fertility, reproductive failures [10–12] and hepatic or lung injury [13,14], but also some immune system disorders [15]. Although in recent years more attention has been given to this class of compounds, they still constitute an emerging issue due to

D. Capela et al. / Talanta 155 (2016) 94–100

the lack of sufficient information concerning their occurrence, fate and (eco)toxicological effects [16,17]. Actually, they have been poorly studied in cosmetics and toiletries, the primary route of human exposure and simultaneously one of the most important environmental contamination sources. Only a few studies have been developed to quantify and determine siloxanes in cosmetics and toiletries (Table S1, Supporting information). The more frequently used analytical methodologies to determine VMSs in cosmetics and toiletries matrices are liquid-liquid extraction (LLE) [1,2] and ultrasound (USE) followed by solidphase extraction (SPE) [18,19] coupled to gas chromatography– mass spectrometry (GC–MS) analysis. Those two extraction methodologies have the disadvantage of using large quantities of organic solvents, being time-consuming and requiring a higher sample handling. To overcome these issues, new approaches have been developed, such as the QuEChERS (“Quick, Easy, Cheap, Effective, Rugged and Safe”) methodology, which combines extraction with clean-up. In fact, the QuEChERS technique is considered a “green” analytical approach because it combines low utilization of consumables (namely solvent and sorbents) with a quick procedure (all-inclusive methodology consisting of dispersion, extraction, and clean-up steps) [20], with low waste production. In this methodology, the homogenized samples are extracted with an organic solvent in an ultrasound bath to reduce the extraction time and increase the extraction yield. In this step, different salts, acids and buffers (e.g. MgSO4, Na2SO4, NaCl, NaCH3COO, Na3C6H5O7, etc.) may be added to enhance the separation/extraction efficiency of the analytes to the organic phase and protect sensitive analytes, preventing their degradation by maintaining an optimal pH. Maintaining the pH throughout the process is also important in order to ensure method reproducibility. Then, the extract suffers a clean-up process through a dispersive solid-phase extraction (dSPE) using suitable sorbents (e.g. C18, PSA, graphitized carbon, etc.) to remove undesired matrix components [21]. The last step consists in the sample analysis, usually employing a chromatographic technique. This method also presents as main advantages the speed (short extraction/clean-up time), low cost and minimum handling of extracts, when compared with the previously techniques and versatility, being adapted to a wide range of substances and matrices [20]. In fact, the efficiency of QuEChERS methodology mainly arises from the replacement of complex analytical procedures by simpler and expeditious protocols with less sample handling, which can minimize the possibility of cross-contamination. Therefore, the aim of this work was to develop and validate a methodology based on QuEChERS extraction followed by GC–MS analysis for the determination of VMSs in different toiletries, which, to the authors' best knowledge, has never been tested before.

95

purchased from VWR (Fontenay-sous-Bois, France). Helium (99.999%), used in the GC–MS system, and nitrogen (99.999%) for solvent evaporation, were supplied by Air Liquide (Maia, Portugal). 2.2. Standards preparation For each individual VMS, including the internal standard M4Q, stock solutions were prepared in hexane at 1.0 g L
1. A 15.0 mg L
1 final mix stock solution comprising all cyclic and linear siloxanes was prepared by diluting appropriate amounts in hexane. An intermediate solution of M4Q was prepared at 150 mg L
1 in hexane and from this, a final stock solution with a concentration level of 75.0 mg L
1. The calibration standards (0.005–2.50 mg L
1) were prepared also in hexane from the final mix stock solution of siloxanes and M4Q (final concentration of 5 mg L
1). A standard solution of M4Q at 5.0 mg L
1 in hexane was also prepared from the initially mentioned stock solution (1.0 g L
1), and was used during samples extraction. All solutions were preserved at
20 °C and protected from the light. 2.3. QuEChERS preparation For the analysis of each sample, two different QuEChERS were prepared based on the work performed by Homem et al. [22]. The first one (QuEChERS 1) contained 800 mg of anhydrous MgSO4, a drying agent used to decrease the water amount, promoting the migration of VMSs to the organic phase, and 750 mg of NaCH3COO, a salt used to maintain the pH value (pH 5.0–5.5). It is also used to increase even more the aqueous phase polarity, enhancing the extraction of siloxanes. The second (QuEChERS 2) contained 60 mg of MgSO4 to remove the remaining water, 60 mg of PSA and 30 mg of C18. The PSA is used to remove polar interferences such as sugars, fatty acids, organic acids, and some pigments existing in the extract, while C18 is used to remove long chain fatty acid compounds, sterols and other non-polar interferences [23]. 2.4. Samples Cosmetics and toiletries were purchased from retail stores in Oporto (Portugal), chosen among the best-selling brands in this region. These products (n¼ 36) were selected according to the information of marketing sales data provided by supermarkets. The samples were divided into different categories according to their overall composition: moisturizers, toothpastes, toilet soaps, shower gels, deodorants/antiperspirants, shaving products and hair care products. Samples were kept in their original containers at room temperature until analysis. 2.5. Sample extraction

2. Materials and methods 2.1. Chemicals and materials Eight volatile methylsiloxanes (four cyclic and four linear) were investigated. Individual linear (L2–L5) and cyclic (D3–D6) volatile siloxanes and also the internal standard used, tetrakis(trimethylsiloxy)silane (M4Q), were purchased from Sigma-Aldrich (St. Louis, MO, USA) with a purity 497%. For the QuEChERS preparation, anhydrous magnesium sulphate (MgSO4) and sodium acetate (NaCH3COO) were also obtained from Sigma–Aldrich (St. Louis, MO, USA), while primary and secondary amine exchange bonded silica sorbent (PSA) and octadecyl-silica (C18) from Supelco (Bellefonte, PA, USA). The MgSO4 was baked at 450 °C overnight before being used. n-Hexane (analytical grade) was

500 mg of each sample (duplicate) was weighed into disposable polypropylene conical tubes and 100 mL of a 5.0 mg L
1 in hexane of M4Q (internal standard) was added. 3 mL of extraction solvent (hexane) was added and the samples were vortexed and sonicated for 3 and 10 min, respectively. Then, the QuEChERS 1 was added to the sample and the mixture was vortexed for 3 min and centrifuged for 10 min at 3700 rpm (2143  g). The supernatant was removed from QuEChERS 1 and transferred to a tube containing the QuEChERS 2. The mixture was vortexed and centrifuged again under the same conditions as before, and the supernatant was transferred to an amber glass vial. The extract was dried under a gentle stream of nitrogen, reconstituted with 1.0 mL of hexane, and then analysed by GC–MS. Whenever necessary, extracts were further diluted to an appropriate volume and reanalysed.

96

D. Capela et al. / Talanta 155 (2016) 94–100

which the GC column was heated at 20 °C min
1 to 300 °C and kept for 3 min.

2.6. Instrumental analysis The extracted samples were analysed using a Varian Ion Trap GC–MS system. The mass spectrometer was operated in electron ionization (EI) mode (70 eV). Separation was achieved at a constant flow of helium (1.0 mL min
1), using a Varian CP-SIL 8-CB capillary column (50 m  0.25 mm, 0.12 mm). The oven temperature was programmed as follows: 35 °C hold for 5 min, raised at 6 °C min
1 to 155 °C and then 20 °C min
1 to 300 °C (hold for 2.75 min) – total time of analysis of 35 min. Injection (1 mL) was in split mode, with a split ratio of 5. Temperatures of manifold, ion trap, transfer line and injector were maintained at 50, 200, 250 and 200 °C, respectively. The filament emission current was 50 mA. For quantitative analysis of target compounds, selected ion storage (SIS) mode was applied. Table S2 (Supporting information) shows the retention time and the quantifier and qualifier ions used for the SIS detection. Quantification was performed, using the internal standard M4Q. 2.7. Validation procedure The analytical methodology was validated based on the procedure described by EuraChem [24,25]. Linearity, limits of detection (LOD) and quantification (LOQ), precision, accuracy and the global uncertainty were evaluated. Linearity was assessed by the direct injection of standards containing all the VMSs at concentrations ranging 0.005 and 2.50 mg L
1. LODs and LOQs were estimated based on a signal-to-noise ratio (S/N) of 3 and 10, respectively. Precision (repeatability) was evaluated by the relative standard deviation (%RSD) of three replicates at different spiking levels (0.1 mg L
1, 0.5 mg L
1 and 1.0 mg L
1) and across-samples (duplicates). Similarly, the accuracy was determined by recovery tests, using three replicate spiked samples at the same concentration levels as before. Due to the differences in the formulations of distinct PCPs, tests were applied to all classes of products investigated. Global uncertainty was assessed using the bottom-up approach proposed by EURACHEM [25]. Stability of the standards solutions and extracts were also investigated, as well as matrix effects (ME). They were calculated for each analyte and type of formulation (concentration level of 0.5 mg L
1), according to Eq. (1):

(peak area of spiked extract − peak area of extract) ME=

−peak area of standard solution peak area of standard solution

× 100

(1)

2.8. Quality assurance/quality control Due to the extensive use of VMSs in daily life products, some precautions were taken into account in order to minimize samples contamination. During this study, analysts avoided the use of personal care products, such as hand creams and lotions and switched gloves whenever they changed samples. Glassware was also subject to a special cleaning and decontamination procedure. After soaking in a phosphates free detergent solution (Derquim LM03, Panreac, Barcelona, Spain) and rinsing with distilled water and acetone, non-calibrated material was baked-out at 400 °C for at least one hour. Finally, the containers were rinsed with pure hexane before use. All plastic material was discarded after use. Procedural blanks were analysed with every extraction batch and samples concentration was corrected accordingly, whenever needed. Chromatographic blanks (injection of pure n-hexane) were also performed, but no memory effects were verified. To remove any residual material from the injections, the chromatographic runs were programmed with a final clean-up step, in

3. Results and discussion 3.1. Extraction and clean-up method selection The QuEChERS methodology applied in this work was based on that previously developed by Homem et al. [22] for extraction of synthetic musks from personal care products. As mentioned before, VMSs are low-polar compounds and due to the physicochemical similarity between these two classes of compounds, the authors decided to use the previously optimized method as a starting point. Initially, the same conditions used in that study (500 mg sample, 3 mL acetonitrile, 10 min USE, Q1: 800 mg MgSO4 þ750 mg NaCH3COO, Q2: 60 mg MgSO4 þ 60 mg of PSA þ30 mg of C18) were tested in the extraction of 500 mg L
1 spiked skin waterin-oil moisturizer samples, because it was considered to be the more complex and challenging matrix within the range of the chosen products. Low recoveries were achieved for all compounds, including the internal standard M4Q (o30%). In fact, VMSs are less polar than synthetic musks and therefore, a non-polar solvent should be used, such as hexane. The use of this solvent conducted to better results, with recovery values ranging from 74 to 104% for all compounds, except L2 (30%) and D3 (60%). In addition, no significant matrix effects were found analysing the chromatograms, which indicates that the applied dispersive solid-phase extraction (dSPE) clean-up methodology is effective. Although, almost all of the obtained recoveries were within an acceptable range (70–120%), the authors decided to improve this step to enhance further lower recoveries. Two sorbents were used in this stage, primary and secondary amine exchange bonded silica (PSA) and octadecyl-silica (C18). As mentioned before, PSA is considered a weak anion exchange sorbent, being used as a polar sorbent to extract polar analytes from non-aqueous matrices, including sugars, fatty acids and some pigments which are definitely present in toiletries products and may behave as interferences. Furthermore, in this procedure C18 was also used with the purpose of removing interfering non-polar compounds, such as lipids or fats. However, VMSs are also considered low polar compounds, being prone to be also removed by such sorbent. Thus, a clean-up step solely using PSA was also evaluated (500 mg sample, 3 mL hexane, 10 min USE, Q1: 800 mg MgSO4 þ750 mg NaCH3COO, Q2: 60 mg MgSO4 þ60 mg of PSA). Lower recoveries (21–42%) were achieved, proving the strong influence of non-polar co-extracted compounds. Thus, based on these results, the authors decided to use the hexane as extracting solvent and PSA þ C18 as clean-up sorbents since higher recoveries and precision values were achieved as well as lower matrix effects. 3.2. Method validation As previously mentioned, only few authors have studied this topic and most of them did not focus their attention in the method validation for such complex and diversified matrices. To evaluate the quality of the implemented analytical method (QuEChERS-GC– MS), validation tests were performed. Additionally, the global uncertainty was estimated and the stability and matrix effects were also assessed. 3.2.1. Linearity range, limits of detection and quantifications Calibration curves were established by direct injection of ten calibration standards in hexane containing all siloxanes at different levels (0.005–2.50 mg L
1). In fact, concentrations were

D. Capela et al. / Talanta 155 (2016) 94–100

Table 1 Linearity range, detection and quantification limits for each compound studied. Compound

Linearity range (mg L
1)

R2

LOD (ng g
1)

LOQ (ng g
1)

L2 D3 L3 D4 L4 D5 L5 D6

0.005–2.50

0.998 0.999 0.999 0.998 0.998 0.999 0.998 0.996

0.17 0.23 0.50 0.38 1.43 0.86 3.75 1.20

0.57 0.76 1.67 1.25 4.76 2.86 12.50 4.00

correlated with the response factors (RF ¼Asiloxane/Ainternal standard), using M4Q as internal standard (500 μg L
1). All examined compounds showed a linear behaviour within the studied range, with R2 between 0.996 and 0.999. The LODs ranged between 0.17 ng g
1 (L2) and 3.75 ng g
1 (L5) (Table 1). Comparing these values with those found in literature for the determination of VMSs in cosmetics and personal care products (Table S1, Supporting information) it is possible to verify that the values obtained in this study are generally lower than the ones already published. Only Lu et al. [21], who applied SPE

97

followed by GC-FID analysis, obtained values in the same order of magnitude. Therefore, the developed methodology (QuEChERSGC–MS) is an improvement as it allows to distinguish lower levels of target substances. 3.2.2. Precision and accuracy The precision was evaluated by the relative standard deviation (%RSD) of three replicates at different levels of spike (0.10 mg L
1, 0.50 mg L
1 and 1.00 mg L
1). The repeatability values indicate that the method is precise once the values obtained are mostly below 10% (average of 5%) (Fig. 1). In fact, RSD values varied between 1% and 18% and the highest values were obtained for L2 and D3. This behaviour may be explained by the higher volatility displayed by these compounds, which may lead to greater losses during the extraction process and therefore, to higher relative standard deviations. Analysing the product categories, higher RSD values were verified for toothpastes (average of 7%) and the lowest values for shampoo, shower gel/gel toilet soap and deodorant (average of 3%). A satisfactory across-samples precision was also obtained, with RSD values usually below 15%. Recovery tests were performed using spiked samples at three different levels (0.10 mg L
1, 0.50 mg L
1 and 1.00 mg L
1). Fig. 1 shows the relation between products and compounds for the three

Fig. 1. Recovery and repeatability (error bars) values for triplicate spiked samples: (A) 0.1 mg L
1, (B) 0.5 mg L
1 and (C) 1.0 mg L
1.

98

D. Capela et al. / Talanta 155 (2016) 94–100

Fig. 2. Matrix effects (%) for each kind of product analysed (sample spiked at 0.5 mg L
1).

different spike levels regarding recovery values. The average recovery obtained for these tests was 84%, which is well acceptable for this type of analysis. Lower recoveries were achieved for L2, which may be also explained for its high volatility. No significant relationship between the type of matrix and the recoveries values was found, but it seems that solid toilet soaps and aftershaves conducts to lower recoveries (around 70%). Mean recoveries of the internal standard, M4Q, in samples was 80%. These recovery values are similar to those determined in literature, using more complicate and time-consuming techniques [18,2,19]. 3.2.3. Matrix effects Although gas chromatography-mass spectrometry (GC–MS) is a powerful instrumental technique, it is also susceptible to matrix effects, which may negatively affect the quantification of the target analytes. Therefore, in this study the percentage of matrix effect (% ME) was calculated for each compound and type of product (Eq. (1)) and the main results are shown in Fig. 2. Generally, values between
20% and 20% indicates that there is no significant matrix effect [26]. In this specific case, the %ME values were between
23% (L2) in aftershaves and 22% (L3) in solid toilet soap, showing no relevant matrix effects. Hence, calibration standards in hexane were used for quantification of VMSs without the need for matrix-match calibration. 3.2.4. Global uncertainty The global uncertainty is determined by identifying, estimating and combining all sources of uncertainty associated with the result, enabling the interpretation and evaluation of individual contributions and the detection of the most significant ones. Four main sources of uncertainty were considered: the uncertainty

associated to the standards preparation (U1), calibration curve (U2), precision (U3) and accuracy (U4) [27]. Fig. 3A shows the global uncertainty for moisturizers, in this case for D5, the compound referred in the literature as the most used one in toiletries. A percentage of global uncertainty was achieved for the upper and intermediate levels of the calibration range (about 5%). However, when concentrations decrease, approaching the limits of detection, the global uncertainty rises significantly. In Fig. 3B the variation of the relative weight of each individual source may also be seen for D5 in moisturizers. The relative contribution of the uncertainty of standard preparation (U1) decreases when the concentration decreases too. Clearly, the importance of the calibration curve uncertainty (U2) increases as it reaches towards the lower concentrations. In fact, for the concentrations between 0.005 and 0.05 mg L
1, U2 accounts for more than 80% of the global uncertainty. The contribution of the uncertainty related to the precision (U3) decreases as the lower concentrations are reached. Regarding the accuracy (U4), it has an important relative contribution for the global uncertainty at the highest concentrations (in some cases more than 50%), decreasing as it reaches the lowest ones. A similar behaviour was verified for the other investigated type of products and compounds. 3.2.5. Stability The stability of the target analytes and internal standard in the stock and working solutions under the period of storage was assessed to ensure that those conditions do not affect the concentration of the analytes. The stability of the solutions maintained at
20 °C was tested as well as the freeze and thaw stability of the analytes in the standards from freezer storage conditions (
20 °C) to room temperature. In the first case, the results

Fig. 3. Global uncertainty of D5 in a moisturizer (A) and variation of the relative weight of each individual source of uncertainty for D5 in a moisturizer (B).

D. Capela et al. / Talanta 155 (2016) 94–100

99

Table 2 Concentrations (mg g
1; mean and range) and frequency of detection (%) of VMSs in different personal care products purchased from local supermarkets. Category

Product type

Moisturizers (n¼ 3)

Body moisturizer

Hand cream

Facial cream

Toilet soaps (n ¼ 3)

Gel soap

Solid soap

Body and hair wash (n ¼3)

Shower gel

Shampoo

Conditioner

Deodorants (n ¼3)

Deodorant

Dentifrice products (n¼ 3)

Toothpaste

Shaving products (n¼ 3)

Shaving foam/gel

Aftershave

Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency Mean Range Frequency

L2

D3

L3

D4

L4

D5

L5

D6

nd nd 0 nd nd-0.16 67 nd nd 0 oLOQ nd-LOQ 67 nd nd-LOQ 33 nd nd 0 0.19 0.16–0.21 100 nd nd 0 nd nd 0 nd nd-LOQ 67 nd nd 0 0.01 nd-0.01 33

4.6 0.04–10.8 100 nd nd-LOQ 33 0.47 nd-0.85 67 12.6 10.4–15.3 100 3.5 nd-3.5 33 130 2.6–310 100 296 44–584 100 23 2.1–40 100 10 0.05–30 100 0.21 0.00–0.59 100 1.42 nd-1.68 67 0.91 0.19–1.67 100

nd nd-LOQ 33 0.04 nd-0.04 67 nd nd-LOQ 33 nd nd 0 nd nd 0 nd nd 0 nd 0.15–0.25 100 0.14 0.08–0.25 100 nd nd 0 nd nd 0 nd nd 0 0.78 nd-0.78 33

13.7 4.1–26 100 4.8 0.83–12.4 100 13.8 LOQ-23.7 67 0.99 0.61–1.46 100 0.92 0.09–2.4 100 36 0.60–93 100 77 20–129 100 61 23–117 100 5.1 2.2–11 100 0.10 0.02–0.28 100 nd nd 0 3.0 0.14–8.1 100

o LOQ nd-LOQ 67 nd nd 0 nd nd-LOQ 33 nd nd 0 nd nd 0 nd nd 0 0.85 0.66–1.00 100 0.61 0.33–0.98 100 nd nd-LOQ 33 nd nd 0 nd nd 0 1.23 nd-1.23 33

356 4.6–754 100 3.3 0.79–1.22 100 250 3.1–408 100 0.10 0.01–0.27 100 5.7 4.6–6.9 100 4.0 0.70–0.62 100 21 7.4–37 100 31 20–49 100 0.59 0.18–1.34 100 0.13 nd-0.24 67 nd nd 0 210 nd-210 33

0.02 nd-0.02 33 0.47 0.10–1.10 100 nd nd 0 nd nd 0 nd nd 0 nd nd 0 0.77 0.62–0.87 100 0.51 0.30–0.78 100 nd nd 0 nd nd 0 nd nd 0 7.9 nd-7.9 33

146 2.8–38 100 0.87 nd-1.32 67 243 0.68–594 100 0.21 nd-0.21 67 5.4 4.0–6.3 100 2.1 0.20–5.4 100 22 4.4–42 100 34 19–62 100 nd nd-0.01 33 0.20 0.02–0.55 100 4.0 nd-4.0 33 50 nd-100 67

LOQ – limit of quantification, nd – not detected.

confirmed that solutions were stable at least 3 months. However, when subjected to regular procedures of freeze/thaw (2 in 2 days), the solutions stability diminishes from 3 months to about 10 days. The stability of processed samples and standard solutions in the autosampler (maximum time of exposure: 12 h) were also checked. A maximum of 10% variation was found.

3.3. Application of the QuEChERS methodology to toiletries To evaluate the suitability of the developed methodology, the occurrence of these compounds in 36 personal care products were examined. The frequency of detection, range and mean concentrations of several cosmetics products are listed in Table 2. Analyses were performed in duplicate. In general, D4 and D5 were more frequently detected in all products. Highest mean concentrations were obtained for D5 in body moisturizers (356 mg g
1), followed by D3 in shampoo (296 m g g
1). Analysing different kind of products, body moisturizers, facial creams and shampoos contained the highest average concentrations of VMSs (ΣVMSs ¼520 mg g
1; 507 mg g
1and 417 m g g
1, respectively). For all the analysed compounds, only D3 and D6 were detected simultaneously in all samples. Globally, these results indicate that QuEChERS extraction is a suitable and effective technique for this kind of matrices once it allows the detection and quantification of VMSs in different cosmetics and personal care products.

3.4. Time consumption and cost estimation The time consumption and cost analysis for this study and for the studies presented in the literature were estimated. For the cost estimation only consumables such as solvents and sorbents were taken into account. These cost values were based on the information from catalogues provided by chemical suppliers. Similarly, the time consumption was estimated according to the details provided in literature [18,2,19]. Once there is no information available related to the number of samples that could be extracted simultaneously, only the extraction time was compared. As mentioned previously, so far only four studies on the literature are related to the determination of VMSs in PCPs (Table S1, Supporting information). Those methodologies without a clean-up step are cheaper (0.12–0.42 €/sample), but present as main disadvantage a decrease in the lifetime of the chromatographic columns. On the other hand, the studies with clean-up step used higher solvent quantities, being more expensive (0.46–1.00 €/ sample) and more time consuming (90–845 min extraction time þclean-up time). The proposed QuEChERS methodology seems to be more appealing than those conventional techniques due to the fast procedure and lower amounts of solvents employed (approx. 40 min and 3 mL of solvent). In the present study, lab-made QuEChERS were used instead of commercial ones, contributing to a decrease of the total cost (0.79 €/sample vs 5.00 €/sample). This, combined with good analytical performance, makes QuEChERS methodology an interesting environmental and cost friendly alternative to conventional approaches.

100

D. Capela et al. / Talanta 155 (2016) 94–100

4. Conclusion In this study, an alternative methodology based on QuEChERS extraction followed by GC–MS analysis was successfully applied to detect and quantify eight VMSs in several personal care products. The proposed methodology was validated, providing a good linearity range (0.005–2.50 mg L
1) and low detection limits (0.17–3.75 ng g
1). Considering these complex matrices, good precision (average RSD values of 5%) and accuracy (average recovery of 84%) were achieved. Negligible matrix effects were found. Subsequently, the whole procedure was successfully applied to real samples. The concentrations of VMSs ranged from not detected to 754 mg g
1 (D5). D4 and D5 were more frequently detected and body moisturizers, facial creams and shampoos showed the highest levels. Overall, QuEChERS proved to be an adequate, simple and fast extraction method that uses small amounts of organic solvents/ sorbents and seems to be a good alternative to other extraction methodologies, being cheaper and more environmentally friendly.

[8]

[9]

[10]

[11]

[12]

[13]

[14]

Acknowledgments [15]

This work was financially supported by: Project POCI-01–0145FEDER-006939 (Laboratory for Process Engineering, Environment, Biotechnology and Energy – EQU/00511/2013) by FEDER funds through Programa Operacional Competitividade e Internacionalização – COMPETE2020 and by national funds through FCT – Fundação para a Ciência e a Tecnologia (Vera Homem postdoctoral grant SFRH/BPD/76974/2011).

[16]

[17]

[18]

Appendix A. Supplementary material [19]

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2016.04.029.

References [1] Y. Horii, K. Kannan, Survey of organosilicone compounds, including cyclic and linear siloxanes, in personal-care and household products, Arch. Environ. Contam. Toxicol. 55 (2008) 701–710. [2] R. Wang, R.P. Moody, D. Koniecki, J. Zhu, Low molecular weight cyclic volatile methylsiloxanes in cosmetic products sold in Canada: implication for dermal exposure, Environ. Int. 35 (2009) 900–904. [3] C. Lassen, C. Hansen, S. Mikkelsen, J. Maag, Siloxanes-Consumption, toxicity and Alternatives, 1st ed., COWI A/S, Denmark, 2005. [4] W. Johnson Jr., W.F. Bergfeld, D.V. Belsito, R.A. Hill, C.D. Klaassen, D.C. Liebler, J. G. Marks Jr., R.C. Shank, T.J. Slaga, P.W. Snyder, F.A. Andersen, Safety assessment of cyclomethicone, cyclotetrasiloxane, cyclopentasiloxane, cyclohexasiloxane, and cycloheptasiloxane, Int. J. Toxicol. 30 (2011) 149s–227s. [5] R. Dewil, L. Appels, J. Baeyens, A. Buczynska, L. Van Vaeck, The analysis of volatile siloxanes in waste activated sludge, Talanta 74 (2007) 14–19. [6] J. Pierce, Siloxanes in Landfill and Digester Gas Update (accessed 20.05.15), 2004 http://www.scsengineers.com/scs-white-papers/siloxanes-in-landfilland-digester-gas-update. [7] D.G. Wang, M. Alaee, H. Steer, T. Tait, Z. Williams, S. Brimble, L. Svoboda, E. Barresi, M. Dejong, J. Schachtschneider, E. Kaminski, W. Norwood, E. Sverko, Determination of cyclic volatile methylsiloxanes in water, sediment, soil, biota,

[20]

[21]

[22]

[23] [24]

[25] [26]

[27]

and biosolid using large-volume injection-gas chromatography-mass spectrometry, Chemosphere 93 (2013) 741–748. A.A. Bletsou, A.G. Asimakopoulos, A.S. Stasinakis, N.S. Thomaidis, K. Kannan, Mass loading and fate of linear and cyclic siloxanes in a wastewater treatment plant in Greece, Environ. Sci. Technol. 47 (2013) 1824–1832. European Chemical Agency (ECHA), Identification of PBT and vPvB substance. Results of evaluation of PBT/vPvB properties for decamethylcyclopentasiloxane (accessed 16.02.16), 2012 http://echa.europa.eu/documents/10162/13628/ decamethyl_pbtsheet_en.pdf. J.M. McKim Jr., P.C. Wilga, W.J. Breslin, K.P. Plotzke, R.H. Gallavan, R.G. Meeks, Potential estrogenic and antiestrogenic activity of the cyclic siloxane octamethylcyclotetrasiloxane (D4) and the linear siloxane hexamethyldisiloxane (HMDS) in immature rats using the uterotrophic assay, Toxicol. Sci. 63 (2001) 37–46. R.G. Meeks, D.G. Stump, W.H. Siddiqui, J.F. Holson, K.P. Plotzke, V.L. Reynolds, An inhalation reproductive toxicity study of octamethylcyclotetrasiloxane (D4) in female rats using multiple and single day exposure regimens, Reprod. Toxicol. 23 (2007) 192–201. A.L. Quinn, J.M. Regan, J.M. Tobin, B.J. Marinik, J.M. McMahon, D.A. McNett, C. M. Sushynski, S.D. Crofoot, P.A. Jean, K.P. Plotzke, In vitro and in vivo evaluation of the estrogenic, androgenic, and progestagenic potential of two cyclic siloxanes, Toxicol. Sci. 96 (2007) 145–153. M.W. Lieberman, E.D. Lykissa, R. Barrios, C.N. Ou, G. Kala, S.V. Kala, Cyclosiloxanes produce fatal liver and lung damage in mice, Environ. Health Perspect. 107 (1999) 161–165. J.M. McKim Jr., G.B. Kolesar, P.A. Jean, L.S. Meeker, P.C. Wilga, R. Schoonhoven, J.A. Swenberg, J.I. Goodman, R.H. Gallavan, R.G. Meeks, Repeated inhalation exposure to octamethylcyclotetrasiloxane produces hepatomegaly, transient hepatic hyperplasia, and sustained hypertrophy in female Fischer 344 rats in a manner similar to phenobarbital, Toxicol. Appl. Pharmacol. 172 (2001) 83–92. B. He, S. Rhodes-Brower, M.R. Miller, A.E. Munson, D.R. Germolec, V.R. Walker, K.S. Korach, B.J. Meade, Octamethylcyclotetrasiloxane exhibits estrogenic activity in mice via ERalpha, Toxicol. Appl. Pharmacol. 192 (2003) 254–261. T. Alvarino, S. Suarez, E. Katsou, J. Vazquez-Padin, J.M. Lema, F. Omil, Removal of PPCPs from the sludge supernatant in a one stage nitritation/anammox process, Water Res. 68 (2015) 701–709. M. Oosterhuis, F. Sacher, T.L. Ter Laak, Prediction of concentration levels of metformin and other high consumption pharmaceuticals in wastewater and regional surface water based on sales data, Sci. Total Environ. 442 (2013) 380–388. T. Dudzina, N. von Goetz, C. Bogdal, J.W. Biesterbos, K. Hungerbuhler, Concentrations of cyclic volatile methylsiloxanes in European cosmetics and personal care products: prerequisite for human and environmental exposure assessment, Environ. Int. 62 (2014) 86–94. Y. Lu, T. Yuan, W. Wang, K. Kannan, Concentrations and assessment of exposure to siloxanes and synthetic musks in personal care products from China, Environ. Pollut. 159 (2011) 3522–3528. C. Ribeiro, A.R. Ribeiro, A.S. Maia, V.M. Goncalves, M.E. Tiritan, New trends in sample preparation techniques for environmental analysis, Crit. Rev. Anal. Chem. 44 (2014) 142–185. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, F.J. Schenck, Fast and easy multiresidue method employing acetonitrile extraction/partitioning and “dispersive solid-phase extraction” for the determination of pesticide residues in produce, J. AOAC Int. 86 (2003) 412–431. V. Homem, J.A. Silva, C. Cunha, A. Alves, L. Santos, New analytical method for the determination of musks in personal care products by Quick, Easy, Cheap, Effective, Rugged, and Safe extraction followed by GC–MS, J. Sep. Sci. 36 (2013) 2176–2184. H. Hubschmann, Handbook of GC–MS: Fundamentals and Application 3rd ed., Wiley-VCH Verlag GmbH & Co. kGaA, 2015. B. Magnusson, U. Örnemark, Eurachem Guide: The Fitness for Purpose of Analytical Methods – A Laboratory Guide to Method Validation and Related Topics, 2nd ed., Eurachem, 2014. S. Ellison, M. Rosslein, A. Williams, EURACHEM/CITAC Guide, Quantifying Uncertainty in Analytical Measurement, Teddington, 3rd ed., 2012. P. Kacznski, B. Lozowicha, M. Jankowska, I. Hrynko, Rapid determination of acid herbicides in soil by liquid chromatography with tandem mass spectrometric detection based on dispersive solid phase extraction, Talanta 152 (2016) 127–136. N. Ratola, L. Santos, P. Herbert, A. Alves, Uncertainty associated to the analysis of organochlorine pesticides in water by solid-phase microextraction/gas chromatography-electron capture detection – evaluation using two different approaches, Anal. Chim. Acta 573–574 (2006) 202–208.