Desenvolvimento de metodologia espectrofotométrica multivariada para o controle de qualidade da associação acido kojico e hidroquinona em dermocosméticos

• Ano de defesa: 2010
• Autor(a) principal: Calaça, Giselle Nathaly
• Orientador(a): Nagata, Noemi
• Banca de defesa: Pessoa, Christiana Andrade , Peralta-zamora, Patricio Guillermo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: UNIVERSIDADE ESTADUAL DE PONTA GROSSA
• Programa de Pós-Graduação: Programa de Pós-Graduação em Química Aplicada
• Departamento: Química
• País: BR
• Palavras-chave em Português: ácido kójico; hidroquinona; espectrofotometria UV-Vis; calibração multivariada; PLSR; validação
• Palavras-chave em Inglês: kojic acid; hydroquinone; UV-Vis spectrophotometry; multivariate calibration; PLSR, validation
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA
• Link de acesso: http://tede2.uepg.br/jspui/handle/prefix/2078

Kojic Acid (KA) and Hydroquinone (HQ) are depigmenting agents used as skinwhitening cosmetics to treat dyschromias. The need for low cost analytical methodologies that enable quality control of this association is evident. Due to the characteristics such as high sensitivity, low cost and operational simplicity, this work aims to use spectrophotometry in the visible region and multivariate calibration tools, mainly Partial Least Squares Regression (PLSR) for quantitative etermination of KA and HQ. The method is based on the complexation of kojic acid with Fe3+ ion (λ = 494nm), while hydroquinone reduces Fe3+ ions to Fe2+, which complex with 1,10-phenanthroline (λ = 510nm). The quantitative determination of KA and HQ by conventional methods was performed by analytical curve, spectrophotometric additivity principle and first-derivative pectrophotometry, and the average errors were: 20,65% (HQ) and 469,91% (AK), 41,44% (HQ) and 47,11% (AK), 57,59% (HQ) and 156,72% (AK), respectively. The results of conventional methods were compared to PLSR, and the multivariate models with better predictive capacity used: spectral range from 350 to 800 nm, mean center data, 03 and 02 latent variables for kojic acid and hydroquinone, respectively. The mean relative errors obtained from these models were: 4,20% for AK and 6,05% for HQ, indicating that PLSR method showed better results for the quantification of analytes, when compared to conventional methods. The multivariate methodology developed was validated according to ANVISA criteria and then used for quantification of analytes in real samples, obtained from manipulation pharmacies of the region.