Fresenius J Anal Chem (2001) 369 : 638–641
© Springer-Verlag 2001
S P E C I A L I S S U E PA P E R
Alberto Chisvert · Maria-Carmen Pascual-Martí · Amparo Salvador
Determination of UV-filters in sunscreens by HPLC
Received: 26 September 2000 / Revised: 14 December 2000 / Accepted: 19 December 2000
Abstract Simultaneous determination of six internationally authorised organic UV-filters in sunscreen formulations was performed by HPLC with UV spectrophotometric detection. The filters determined were: sulisobenzone, oxybenzone, octyl dimethyl PABA, octyl methoxycinnamate, octyl salicylate and homosalate. A C18 stationary phase and a mobile phase of ethanol water acetic acid (70 : 29.5 : 0.5) were used with a flow rate of 0.5 mL/min. UV measurements were carried out at 313 nm. The time required for the analysis was 25 min and the limits of detection were between 0.2 and 2 mg/L, except for sulisobenzone, which gave a limit of detection of 20 mg/L. The procedure proposed provides an accurate, fast and green analytical method, that does not involve toxic organic solvents.
chromatography (GC) [8, 9], high performance thin layer chromatography (HPTLC) [10, 11], and especially high performance liquid chromatography (HPLC) [12, 13]. Although usually sunscreen cosmetics were analysed, UVfilters have also been determined in urine [8] and human milk [9]. The HPLC methods generally use reverse phase C-18 or C-8 as stationary phases and acetonitrile-water [14], tetrahydrofuran-water [15], methanol-water [13] or mixtures of three solvents acetonitrile-tetrahydrofuran-water [13], methanol-tetrahydrofuran-water [16] or methanolacetonitrile-water [17] as mobile phases. In this study, a simple and fast method using a commercially available liquid chromatograph and low toxicity solvents is proposed for simultaneous determination of six internationally authorised UV-filters.
1 Introduction 2 Experimental Chemical sunscreen filters are compounds that absorb deleterious UV light, thereby decreasing the amount of solar radiation reaching the skin. The use of sunscreens could help to prevent or to minimise the harmful effects of solar radiation on the skin [1]. The efficacy of the sunscreens can be estimated by the Sun Protection Factor (SPF), which depends on the UV-filters content of the formulation. Moreover, in the case of organic filters, some dermatological reactions have been described [2]. Therefore, the maximum content of such filters in sunscreens has been legislated, and it is this that makes it vital to analyse their composition. Diverse methods have been used to determine UV-filters in sunscreen products. FAAS and ICP-AES have been used for inorganic filters [3], while organic filters have been determined by NMR spectroscopy [4], Raman spectroscopy [5], UV-VIS absorption spectroscopy [6, 7], gas
2.1 Instrumentation A Hitachi L-7100 liquid chromatograph equipped with a 20 µL loop injector and a Hitachi L-7420 UV-VIS detector were employed, using a LiChrospher® 100 RP-18 (12.5 cm length, 4 mm i.d., 5 µm particle size) (Merck) column. 2.2 Reagents Sulisobenzone (SUL), octyl dimethyl PABA (ODP) and octyl methoxycinnamate (OMC) were from ROIG FARMA S.A., Tarrasa (Spain); oxybenzone (OX) and octyl salicylate (OS) from ALDRICH, Barcelona (Spain); homosalate (HS) from CHEMIR S.A., Barcelona (Spain); 4-methylbencylidene camphor (MBC) from GUINAMA S.L., Valencia (Spain); ethanol (EtOH) HPLC grade from SCHARLAB, Barcelona (Spain); acetic acid (AcH) analytical grade from PANREAC, Barcelona (Spain). Other reagents were of analytical grade except the reagents used to prepare the homemade sunscreen. 2.3 Samples
A. Chisvert · M. C. Pascual-Martí () · A. Salvador Departamento de Química Analítica, Facultad de Química, Universitat de València, Doctor Moliner 50, 46100-Burjassot, Valencia, Spain e-mail:
[email protected]
Samples were purchased in local markets: a solar milk from Laboratorios VIGMAR S.L., Valencia (Spain), a sun lotion from HAWAIIAN TROPIC, Hawaii (USA), two oil-free lotions from CLINIQUE Laboratories, London (UK). A homemade sunscreen
639
Fig. 1 (a) Resolution of octyl dimethyl PABA (ODP) and octyl methoxycinnamate (OMC) peaks versus % of ethanol at different flow rates (detection at 313 nm). (b) Resolution of octyl methoxycinnamate (OMC) and homosalate (HS) peaks versus % of ethanol at different flow rates (detection at 313 nm)
sample containing known concentrations of sulisobenzone, octyl dimethyl PABA, octyl methoxycinnamate, oxybenzone, octyl salicylate and homosalate was prepared in the laboratory according to a common procedure followed in the cosmetic industry (provided by GUINAMA Laboratories, Valencia, Spain) and using GUINAMA S.L. reagents. This formulation also contained other ingredients such as: a base PFC o/w (base cream with myristyl myristate, cetyl alcohol, glyceryl laurate, cetearyl octanoate, isopropyl myristate and others lipophilic components), aguacate oil, dimethicone 350, propylen glycol, vitamin E, hydroviton, and phenonip.
Fig. 2 (a) Chromatogam of a standard solution containing 50 mg/L of each UV-filter, (b) chromatogam of a homemade sunscreen sample, (c) chromatogam of a commercial sample C. (1) sulisobenzone, 1.4 min; (2) oxybenzone, 4.9 min; (3) 4-methylbenzylidene camphor (internal standard), 8.2 min and 10.8 min; (4) octyl dimethyl PABA, 13.7 min; (5) octyl methoxycinnamate, 15.9 min; (6) homosalate, 17.6 min; (7) octyl salicylate, 20.2 min; (8) homosalate, 22.7 min. Chromatographic conditions: mobile phase EtOH/H2O/AcH (70:29.5:0.5); flow rate 0.5 mL/min; injection volume 20 µL; detection at 313 nm
640 2.4 Method 0.3–0.5 g of sample were dissolved with 25 mL ethanol, and 1 mL of this solution (previously filtered when the sample contained TiO2) was transferred to a 10 mL volumetric flask and diluted with ethanol. 20 µL of standard and sample solutions were injected into the liquid chromatograph and eluted, using H2O/HAc/EtOH 29.5 : 0.5 : 70 (v/v/v) as mobile phase, at a flow rate of 0.5 mL/min. The UV-VIS detection was carried out at 313 nm. Multicomponent solutions of the six UV-filters in ethanol were used as standards (25–150 mg/L).
3 Results and discussion Ethanol water acetic acid was chosen as mobile phase because of the good solubility of the samples in ethanol and the low toxicity and cost of this solvent. Acetic acid was used to decrease the peak tailing of oxybenzone. Table 1 Analytical parameters. (Sulisobenzone (SUL), oxybenzone (OX), octyl dimethyl PABA (ODP), octyl methoxycinnamate (OMC), octyl salicylate (OS) and homosalate (HS)) UVfilter
Retention time (min)
Sensitivitya (mg/L)
LODb Variation (mg/L) coefficient (%)
SUL OX ODP OMC OS HS
1.44 ± 0.01 4.90 ± 0.01 13.71 ± 0.07 15.9 ± 0.1 20.2 ± 0.1 22.7 ± 0.1
(751 ± 9) 102 (1105 ± 3) 102 (2497 ± 3) 102 (2175 ± 3) 102 (336 ± 3) 102 (218 ± 3) 102
20 0.6 0.2 0.4 0.9 2.0
0.72 0.58 0.59 0.44 0.48 0.52
a slope
of the calibration curve. as 3 Sx/y/b (Sx/y standard deviation and b slope of the straight line calibration). For SUL the LOD was calculated as three times the blank signal at the dead time column (t0 ) b Calculated
Table 2 Analysis of a homemade sunscreen cream.(Sulisobenzone (SUL), oxybenzone (OX), octyl dimethyl PABA (ODP), octyl methoxycinnamate (OMC), octyl salicylate (OS) and homosalate (HS))
Table 3 Analysis of commercial sunscreen samples.(Sulisobenzone (SUL), oxybenzone (OX), octyl dimethyl PABA (ODP), octyl methoxycinnamate (OMC), octyl salicylate (OS) and homosalate (HS))
SUL Found Content (percentage m/m) % Recovery Real Content (percentage m/m)
UV-filter
ODP
OMC
OS
HS
2.7 ± 0.1 5.42 ± 0.05 4.17 ± 0.03 4.75 ± 0.02 2.63 ± 0.05 4.91 ± 0.03 96 ± 2 2.64
98 ± 1 5.17
98 ± 1 4.03
98 ± 1 4.58
98 ± 1 2.51
99 ± 1 4.68
Contents ± Sa (percentage m/m) % Recovery ± S Sample B
Sample C
Sample D
2.57 ± 0.02 98 ± 3 –
5.24 ± 0.03 99 ± 2 –
–
OS
1.52 ± 0.08 97 ± 1 2.26 ± 0.04 97 ± 1 0.58 ± 0.05 96 ± 1 –
7.72 ± 0.08 100 ± 2 –
HS
–
–
7.78 ± 0.04 96 ± 3 5.14 ± 0.04 101 ± 2 3.42 ± 0.03 100 ± 2
4.22 ± 0.03 99 ± 2 4.09 ± 0.03 100 ± 2 –
OX
OMC
are average value of three replicates
OX
Sample A
ODP
a Results
In order to establish suitable conditions for good chromatographic resolution, different ethanol/water ratios and flow rates were tested. Figure 1 shows the resolution of the peaks obtained at different % ethanol and flow rates for ODP-OMC and OMC-HS filters. These compounds were chosen to carry out the optimisation due to their poor chromatographic resolution (see Fig. 2). As can be seen (Fig. 1), when the ethanol content increases the resolution of the ODP-OMC peaks decreases, but the resolution of the OMC-HS peaks increases, consequently 70% ethanol was chosen. Figure 1 also shows that in all instances the resolution increased when the flow rate decreased, therefore 0.5 mL/min was chosen. Ethanolic solutions containing the six UV-filters in the range of 0.1–150 mg/L and 4-methylbencylidene camphor (MBC) (15 mg/L), as internal standard, were injected into the chromatograph and eluted using ethanol water acetic acid (70 : 29.5 : 0.5, v/v/v) as mobile phase. The chromatography was performed in isocratic conditions at a flow rate of 0.5 mL/min. Detection was accomplished at 313 nm, which is a compromise absorption wavelength that permits satisfactory responses for all analytes. The calibration curves obtained (using external and internal standard methods) were linear in the assayed range of concentrations except for sulisobenzone, which eluted at t0, and showed a linear behaviour up to 25 mg/L. The results obtained show that the use of an internal standard can be avoided. Table 1 shows the chromatographic and analytical parameters obtained under the optimised experimental conditions. As can be seen, the sensitivity, reproducibility and limits of detection (LOD) are satisfactory. The high LOD obtained for sulisobenzone is due to its being an unretained compound which is eluted at dead time of the col-
–
641
umn (to) with a high baseline disturbance. Repeatability was < 1% (n = 5). Figure 2 shows the chromatograms obtained for a standard solution and the homemade sunscreen cream. The homosalate results in two peaks (at 17.6 and 22.7 min) corresponding to two stereoisomeric forms with similar sensitivity but very different proportions and, as can be seen in Fig. 2 c, the two isomers are also present in the samples. The main peak (at 22.7 min) was used for the determination of homosalate. A homemade sample was employed to check the accuracy of the method. The results obtained are shown in Table 2. As can be seen, the contents obtained were in agreement with the real contents, with a maximal relative error below 5%. Four commercial sunscreen products were analysed by the proposed method. (see Table 3). The recoveries (obtained by spiking commercial samples with standards of the filters) were between 96 and 101%, which indicates that no proportional errors occur during the analysis of the samples. From the study performed it can be concluded that the proposed method shows suitable accuracy, precision, sensitivity and sampling rate for the determination of the UVfilter in sunscreens. Acknowledgements The authors wish to thank the Spanish Ministerio de Educacion y Cultura (PM-98–0210) for financing this study. A. Chisvert expresses his gratitude to Consellería d’Educació i Cultura (Generalitat Valenciana, Valencia, Spain) and to Ministerio de Educación y Cultura (Spain) for a predoctoral grant.
References 1. Gasparro FP, Mitchnick M, Nash JF (1998) Photochem Photobiol 68:243–256 2. Schmidt T, Ring J, Abeck D (1998) Dermatology 196:354–357 3. Salvador A, Pascual-Marti MC, Adell JR, Requeni A, March JG (2000) J Pharm Biomed Anal 22:301–306 4. Mori K, Itoh K, Susuki S, Nakamura H (1996) Jpn J Toxicol Environ Health 42:60–66 5. Cheng J, Li YS, Roberts RL, Walker G (1997) Talanta 44:1807–1813 6. Chisvert A, Salvador A, Pascual-Marti MC (2000) Anal Chim Acta 428:183–190 7. Chisvert A, Salvador A, Pascual-Marti MC, March JG (2001) Fresenius J Anal Chem 369 (this issue) 8. Felix T, Hall BJ, Brodbelt JS (1998) Anal Chim Acta 371:195– 203 9. Hany J, Nagel G (1995) Dtsch Lebensm Rundsch 91:341–345 10. Musial B, Sherma J (1998) Acta Chromatogr 8:5–12 11. Musial B, Sherma J (1997) J Planar Chromatogr Mod 10:368– 371 12. Vaquerp V, Rodriguez C, Coiffard C, Coiffard LJM, De RoeckHoltzhauer Y (1999) J Chromatogr 832:273–277 13. Rastogi SC, Jensen GH (1998) J Chromatogr 828:311–316 14. Gagliardi L, Cavazzutti G, Montaranella L, Tonelli D (1989) J Chromatogr 464:428–433 15. DiNunzio JE, Gadde RR (1990) J Chromatogr 519:117–124 16. Ikeda K, Suzuki S, Watanabe Y (1989) J Chromatogr 482:240– 245 17. Gagliardi L, Amato A, Turchetto L, Cavazzutti G, Tonelli D, (1990) Anal Lett 23:2123–2133
