Caracterização da qualidade e previsão dos parâmetros físico-químicos de gasolinas comerciais brasileiras através da aplicação de métodos quimiométricos em perfis (fingerprintings) espectroscópicos de ressonância magnética nuclear
Carregando...
Data
2010-04-20.
Autores
Flumignan, Danilo Luiz [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Uma das características mais interessantes da instrumentação analítica moderna é o número de variáveis que podem ser medidas simultâneamente em uma única amostra. Dentre as técnicas analíticas, a RMN 1 H e 13 C, destaca se como uma das mais importantes na análise de “fingerprintings” de misturas complexas. Por outro lado, problemas na qualidade de combustíveis comercializados, têm sido constantemente relatados, tornando relevante o desenvolvimento de novos métodos de análise alternativos que complementem os aplicados atualmente. No Brasil a qualidade da gasolina comercial é regulamentada pela Resolução ANP nº 309. A seleção das amostras mensais de gasolinas comerciais representativas com diferentes similaridades foi realizada por HCA, após a determinação de diversos parâmetros físico químicos estabelecidos pela Resolução ANP nº. 309, em 400 amostras coletadas durante abril a setembro de 2005 em postos revendedores da região Centro Oeste de São Paulo. A partir da amostra mais complexa, determinada por CG MS, foram otimizadas as condições de RMN 1 H (volume de amostra: 30 L, volume de solvente: 600 L e tempo de análise: 2 min), e RMN 13 C (volume de amostra: 200 L, volume de solvente: 600 L e tempo de análise: 10 min), que apresentaram razão sinal/ruído e tempo de análise adequados, contribuindo para a construção do banco de dados dos perfis espectrais. O banco de dados foi constituído das intensidades espectrais de cada perfil para cada deslocamento químico, bem como dos parâmetros físico químicos estabelecidos pela Resolução ANP nº. 309 e de suas respectivas características de qualidade. O algoritmo SIMCA RMN 1 H e SIMCA RMN 13 C mostrou que os perfis espectrais podem ser correlacionados de forma satisfatória e confiável com a qualidade das gasolinas comerciais brasileiras (sensibilidade superior a 90%)...
The most interesting features of modern instruments is the number of variables that can be measured in a single sample. The analytical techniques, one dimensional 1 H and 13 C NMR, stand out as the most important to fingerprintings analysis of complex mixtures. On the other hand, problems in Brazilian commercial fuel quality, including adulteration episodes, have been consistently reported, making relevant the development of news alternatives analytical methods in laboratories and governmental agencies. In Brazil, the gasoline quality control is regulated by ANP Resolution nº. 309. The samples selection of representative commercial gasoline was performed by HCA, after the determination of several physicochemical parameters, established by ANP Resolution nº. 309, in 400 gasoline samples collected during April to September 2005, in gas stations of the midwest region of Sao Paulo State. The most complex sample, determined by GC MS, promoted the optimization of 1 H NMR conditions (sample volume: 30 mL, solvent volume: 600 mL and analysis time: 2 min) and 13 C NMR conditions (volume Sample: 200 mL, solvent volume: 600 mL and analysis time: 10 min), whose showed adequated signal to noise ratio and time analysis, contributing for the construction of the spectral profile database. Database was composed by the intensities and chemical shift in the spectral profile, as well as by the physicochemical parameters and its respective quality characteristics. SIMCA 1 H NMR and SIMCA 13 C NMR showed satisfactory and confidential correlation between the spectral profiles and the quality of the Brazilian commercial gasoline (sensibility above 90%). Similarly, PLS 1 H NMR and PLS 13 C NMR provided satisfactory correlations in the prediction of several physicochemical parameters, in other words, showed standards erros results RMSEC, RMSEV and RMSEP with proportional intensities... (Complete abstract click electronic access below)
The most interesting features of modern instruments is the number of variables that can be measured in a single sample. The analytical techniques, one dimensional 1 H and 13 C NMR, stand out as the most important to fingerprintings analysis of complex mixtures. On the other hand, problems in Brazilian commercial fuel quality, including adulteration episodes, have been consistently reported, making relevant the development of news alternatives analytical methods in laboratories and governmental agencies. In Brazil, the gasoline quality control is regulated by ANP Resolution nº. 309. The samples selection of representative commercial gasoline was performed by HCA, after the determination of several physicochemical parameters, established by ANP Resolution nº. 309, in 400 gasoline samples collected during April to September 2005, in gas stations of the midwest region of Sao Paulo State. The most complex sample, determined by GC MS, promoted the optimization of 1 H NMR conditions (sample volume: 30 mL, solvent volume: 600 mL and analysis time: 2 min) and 13 C NMR conditions (volume Sample: 200 mL, solvent volume: 600 mL and analysis time: 10 min), whose showed adequated signal to noise ratio and time analysis, contributing for the construction of the spectral profile database. Database was composed by the intensities and chemical shift in the spectral profile, as well as by the physicochemical parameters and its respective quality characteristics. SIMCA 1 H NMR and SIMCA 13 C NMR showed satisfactory and confidential correlation between the spectral profiles and the quality of the Brazilian commercial gasoline (sensibility above 90%). Similarly, PLS 1 H NMR and PLS 13 C NMR provided satisfactory correlations in the prediction of several physicochemical parameters, in other words, showed standards erros results RMSEC, RMSEV and RMSEP with proportional intensities... (Complete abstract click electronic access below)
Descrição
Palavras-chave
Quimiometria, Controle de qualidade, Ressonancia magnetica nuclear, Gasollinas comerciais, Chemometrics, Commercial gasoline, Quality control, Nuclear magnetic resonance
Como citar
FLUMIGNAN, Danilo Luiz. Caracterização da qualidade e previsão dos parâmetros físico-químicos de gasolinas comerciais brasileiras através da aplicação de métodos quimiométricos em perfis (fingerprintings) espectroscópicos de ressonância magnética nuclear. 2010. 105 f. Tese (doutorado) - Universidade Estadual Paulista, Instituto de Química, 2010.