Available online at www.sciencedirect.com Journal of Chromatography A, 1174 (2007) 95–103 Sensitive determination of free benzophenone-3 in human urine samples based on an ionic liquid as extractant phase in single-drop microextraction prior to liquid chromatography analysis Lorena Vidal a , Alberto Chisvert a,∗ , Antonio Canals a,∗ , Amparo Salvador b a Departamento de Qu´ ımica Anal´ ıtica, Nutrici ´ on y Bromatolog´ ıa, Facultad de Ciencias, Universidad de Alicante, P.O. Box 99, E-03080 Alicante, Spain b Departamento de Qu´ ımica Anal´ ıtica, Facultad de Qu´ ımica, Universidad de Valencia, 46100 Burjassot, Valencia, Spain Available online 6 August 2007 Abstract Benzophenone-3 (BZ3), one of the compounds most commonly used as UV filter in cosmetic products, can be absorbed through the skin into the human body, since it can be found at trace levels in urine from users of cosmetic products that contain BZ3. Moreover, different undesirable effects have been attributed to this compound. Thus, sensitive analytical methods to monitor urinary excretion of this compound should be developed. This paper presents a selective and sensitive methodology for BZ3 determination at ultratrace levels in human urine samples. The methodology is based on a novel microextraction technique, known as single-drop microextraction (SDME). An ionic liquid (IL) has been used as extractant phase instead of an organic solvent. After the microextraction process, the extractant phase was injected into a liquid chromatography system. The variables of interest in the SDME process were optimized using a multivariate optimization approach. A Plackett–Burman design for screening and a circumscribed central composite design for optimizing the significant variables were applied. Ionic strength, extraction time, stirring speed, pH, ionic liquid type, drop volume and sample volume were the variables studied. The optimum experimental conditions found were: sodium chloride concentration, 13% (w/v); extraction time, 25 min; stirring speed, 900 rpm; pH, 2; ionic liquid type, 1-hexyl-3-methylimidazolium hexafluorophosphate ([C 6 MIM][PF 6 ]); drop volume, 5 L; and sample volume, 10 mL. The proposed method requires a standard addition calibration approach, and it has been successfully employed to determine free BZ3 in urine samples coming from human volunteers who applied a sunscreen cosmetic containing this UV filter. The limit of detection was in the order of 1.3 ng mL -1 and repeatability of the method, expressed as relative standard deviation, was 6% (n = 8). © 2007 Elsevier B.V. All rights reserved. Keywords: Benzophenone-3; UV filter; Sunscreen; Cosmetics; Urine; Single-drop microextraction (SDME); Ionic liquid; Multivariate optimization; Liquid chro- matography 1. Introduction The damage that the ozone layer has suffered in recent years has increased the solar UV radiation reaching Earth. It is largely reported elsewhere that exposure to this UV radiation in small amounts has a therapeutic effect on human health; however, it is also well-documented that over-exposure to sunlight can pro- mote different harmful effects on human health, including skin cancer. So, the use of sunscreen cosmetic products incorporat- ∗ Corresponding authors. Tel.: +34 96 590 97 90; fax: +34 96 590 97 90. E-mail addresses: alberto.chisvert@ua.es (A. Chisvert), a.canals@ua.es (A. Canals). ing UV filters, which quench the harmful solar radiation, may prevent or minimize these negative effects of sunlight. Nowadays, in order to achieve greater protection to solar radi- ation, UV filters are added not only to cosmetics to be used for sunbathing but also to daily cosmetic products, such as face day- creams, after-shave products, makeup formulations, lipsticks, shampoos, etc. The number and maximum concentration of UV filters to be used in cosmetics is regulated by the legislation in force in each country [1], since different dermatological side-effects have been attributed to these cosmetic ingredients [2,3]. On the other hand, different studies carried out in vitro or in vivo (using animals, e.g., fishes or rats) seem to indicate that some organic UV filters have significant estrogenic and/or 0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.07.077