Quantification of Sunscreen Ethylhexyl Triazone in Topical Skin-Care Products by Normal-Phase TLC/Densitometry

Ethylhexyl triazone (ET) was separated from other sunscreens such as avobenzone, octocrylene, octyl methoxycinnamate, and diethylamino hydroxybenzoyl hexyl benzoate and from parabens by normal-phase HPTLC on silica gel 60 as stationary phase. Two mobile phases were particularly effective: (A) cyclohexane-diethyl ether 1 : 1 (v/v) and (B) cyclohexane-diethyl ether-acetone 15 : 1 : 2 (v/v/v) since apart from ET analysis they facilitated separation and quantification of other sunscreens present in the formulations. Densitometric scanning was performed at 300 nm. Calibration curves for ET were nonlinear (second-degree polynomials), with R > 0.998. For both mobile phases limits of detection (LOD) were 0.03 and limits of quantification (LOQ) 0.1 μg spot−1. Both methods were validated.


Introduction
Ethylhexyl triazone (ET, Figure 1) is an oil-soluble UVB filter (λ max in ethanol 314 nm) manufactured by BASF under the trade mark Uvinul T150 and used in cosmetic formulation at concentrations up to 5%. Due to its insolubility in water and affinity to the skin keratin, it is particularly suitable for water-resistant products. Its excellent photostability and high absorption coefficient make it a valuable ingredient when a high SPF (sun protection factor) value is required [1].
The objective of this study was to develop a simple and cost-effective method of analysis of ethylhexyl triazone in complex sunscreen preparations by normal-phase thin-layer chromatography followed by densitometry.  imately 70 mL methanol was added to each sample, and the flasks were vigorously shaken by use of a Premed (Poland) type 327 Universal Shaker for 60 min. Methanol was then added to volume, and the flasks were wrapped with aluminum foil and left to stand for 60 min.

Thin-Layer Chromatography.
Thin-layer chromatography was performed on 10 × 10 cm HP quality silica gel 60 plates (layer thickness 0.2 mm) from Merck or on 10 × 20 cm standard quality silica gel 60 plates (layer thickness 0.25 mm), also from Merck. Plates were spotted with the Desaga AS 30 sampler equipped with a 10 μL syringe (1 μL spot −1 ), 15 mm from the bottom edge and at 8 mm intervals, starting 10 mm from the plate edge and developed with either cyclohexanediethyl ether 1 : 1 (v/v), Method A, or cyclohexane-diethyl ether-acetone 15 : 1 : 2 (v/v/v), Method B. Plates were developed in a vertical chromatographic chamber lined with filter paper and previously saturated with the appropriate mobile phase vapor for 20 min. Development distance was 75 mm from the plate bottom edge. After development, plates were dried at room temperature (20 • C), scanned, and analyzed in reflectance mode with the Desaga CD 60 densitometer at 300 nm.

Analysis of Ethylhexyl Triazone in Sunscreen
Creams or Lotions. The sunscreen products solutions in methanol, prepared as described above, were spotted on silica gel 60 HP TLC plates (2 μL). The plates were then chromatographed as described above for ET standards (Section 2.3).

Method Development.
The sun-care preparations analyzed in this study contained, apart from ET, other UV filters, that is, avobenzone (AVO) and octocrylene (OCR) (Sample A) or octyl methoxycinnamate (OMC) and diethylamino hydroxybenzoyl hexyl benzoate (DHHB) (Sample B), and preservatives absorbing within the UV range (parabens). In the course of our earlier research [16][17][18], three stationary phases (silicagel 60, RP-2 and RP-18) and several mobile phases were investigated. ET is a relatively lipophilic compound with strong affinity to RP-18 stationary phase [16][17][18]. On the other hand, its separation from AVO, OCR, OMC, and DHHB on RP-2 stationary phase is poor [16]. For these reasons, it was decided that silica gel 60 is the stationary phase of choice. Mobile phases capable of effective ET separation from other UV filters listed above included cyclohexane-diethyl ether 1 : 1 (v/v), cyclohexanediethyl ether-isopropanol 15 : 1 :  [16,17]. Both mobile phases gave ET spots of sufficient quality for densitometric analysis, although in the case of mobile phase B the spots were of slightly better quality. Analytical wavelength suitable for ET analysis (300 nm) was selected on the basis of multiwavelength scans obtained for this sunscreen.     to our earlier studies [16], from the majority of other sunscreens used in contemporary sun-care preparations. The efficiency of Method B is slightly lower since the separation of ET from parabens is incomplete (R f values for ET and ethylparaben are 0.10 and 0.15, resp.); this is, however, not a problem, since the analytical wavelength for ET is 300 nm (Section 3.1.), and, as it can be seen in Figure 4, parabens do not absorb at 300 nm. Purity of ET peaks obtained during the analysis of Sample A was confirmed by UV/VIS spectra of sunscreens acquired directly from chromatographic plates in reflectance mode. Spectra collected at three different points of particular peaks obtained for the sample solution were compared with spectra acquired for the standard ( Figure 5).

Calibration. Calibration plots for Methods A and B
were obtained by plotting peak areas against amount of ET in the range 0.1-2.0 μg spot −1 . In both cases linear regression coefficients were relatively high (R = 0.9905 and 0.9851,  resp.) but since this should not be used as the sole proof of linearity, nonnumerical analysis of residues according to [19] was performed. Residues (differences between experimental values and those calculated on the basis of appropriate equations) for linear calibration plots proposed for methods A and B showed strong tendencies which suggested that linear fit is inappropriate ( Figure 6). Two possibilities were considered at this stage: selecting a narrower, pseudolinear range or using a different type of equation. Calibration plots were finally generated in the form of second-degree polynomials (Table 1), and their quality was assessed again by means of R values and non-numerical analysis of residues ( Figure 7). Residues plots for quadratic calibrations A and B (Figure 7) showed the lack of tendency that combined with very high R values confirmed the correctness of curves fitting. It should be mentioned in this point that densitometric detection in Methods A and B was performed in reflectance   Table 1).
mode. Lambert-Beer's law cannot be applied to diffuse reflectance so calibration in TLC/densitometry is seldom perfectly linear [19]; if this is the case, quadratic equations are often used [19].

Precision.
Repeatability of the method was tested according to [19][20][21] by replicating the entire method on the same day, using the same cosmetic preparations, batches of solvents, and chromatographic plates, by the same analyst (Day 1, Analysis I and II). Intermediate precision was verified according to [19][20][21] by repeating the procedure on the same cosmetic preparations but on a different day, by a different analyst, using other batches of solvents and chromatographic plates (Day 2). The results of these experiments (Table 2) prove that the methods' precision is sufficient for routine product analysis.

Limits of Detection and Quantification.
The limits of detection and quantification for ET determined experimentally on the basis of signal-to-noise ratio according to [22] are given in Table 1. The results of these changes are summarized in Table 2. 3.2.6. Accuracy. Blank cosmetic creams were spiked with ET, AVO, and OCR (A) or ET, OMC and DHHB (B) at three concentrations 1, 3, and 5% (w/w) of each sunscreen corresponding to 0.2, 0.6, and 1.0 μg spot −1 (2 μL spot −1 of the cream solution prepared according to Section 2.2.). The analytical procedures A and B described in Section 2 were performed on the samples, and the recoveries are presented in Table 3.

Storage and Stability of Standard Solutions.
Standard solutions of ET as well as solutions of other sunscreens and preservatives used in this investigation were refrigerated between the experiments and not exposed to light except for time needed for plate spotting. The stability of all solutions was in these conditions excellent as tested by UV/VIS spectroscopy over the period of 2 weeks.

Conclusions
Ethylhexyl triazone may be quickly and effectively separated from other oil-soluble UV filters and preservatives by normal-phase HPTLC on silica gel 60. Separation can be achieved by a variety of mobile phases, of which two, cyclohexane-diethyl ether 1 : 1 (v/v) or cyclohexane-diethyl ether-acetone 15 : 1 : 2 (v/v/v), were found superior. The methods of ethylhexyl triazone separation and quantification presented in this paper are based on one of the cheapest stationary phases (silica gel 60, compared e.g., to RP-18 or RP-8 layers) and do not require toxic solvents. The analyses may be performed with analytical-grade solvents (HPLC purity solvents are not required), and, although HPTLC plates and automatic spotting are preferred, relatively good results may be achieved on standard-quality TLC plates spotted with a microsyringe. Fast, reliable and cost-effective densitometric quantification of ET proposed in this paper may, therefore, be recommended for routine analysis of cosmetic products.