FTIR and FTNIR multivariate analysis: detection of ethanol fuel adulteration

ANYL 142

Joel C. Rubim, jocrubim@unb.br1, Flavia C. C. Oliveira, flaviacco@unb.br1, Helena S. P. Carneiro, helena_carneiro@hotmail.com2, and Paulo A. Z. Suarez, psuarez@unb.br1. (1) Laboratório de Materiais e Combustíveis, University of Brasilia - Institute of Chemistry, CP 04478, Brasilia, 70904-970, Brazil, (2) Laboratório de Referência para Qualidade de Combustíveis (CEPAT), Agência Nacional de Petróleo, Gás Natural e Biocombustíveis (ANP), SGAN, Quadra 603 Módulo H sala 104 04478, Brasilia, 70900-000, Brazil
Ethanol fuel has been used in Brazil since 1975 and is known as AEHC (hydrated ethyl alcohol fuel). The Brazilian National Agency for Petroleum, Natural Gas and Biofuels (ANP) is the regulatory agency that establishes the fuels quality parameters. Since the use of ethanol fuel is very restricted, there are few research reports related to the determination of the quality parameters for this kind of fuel. In the last 5 years, the ANP has developed efforts in order to avoid fuel adulterations. Due to the lower costs of methanol, it is not rare that ANP finds ethanol fuel samples adulterated with methanol. In the present work we propose the use of partial least squares regression (PLS) calibration models based on Fourier transform-near infrared (FTNIR) and FTIR measurements of twenty five methanol/ethanol/water standards to detect these kind of adulteration. Ten independent samples were used as a group test. The results have shown that among the two investigated spectroscopic techniques PLS/FTIR presented the best performance with a root mean square of cross validation (RMSECV) of 0.077 as compared to the value of 0.125 obtained by the FTNIR model. Despite the difference in the RMSECV values, both techniques presented relative standard errors of prediction smaller than 1.23% for the test group. These models were further evaluated with ten commercial samples of ethanol fuel as obtained by the ANP whose methanol contents ware determined by CGFID. The PLS/FTNIR and PLS/FTIR models provided recovered methanol contents with relative errors better than 1%.
 

General Posters
7:00 PM-9:00 PM, Sunday, August 19, 2007 BCEC -- Exhibit Hall - B2, Poster

Division of Analytical Chemistry

The 234th ACS National Meeting, Boston, MA, August 19-23, 2007