Contents lists available at ScienceDirect Vibrational Spectroscopy journal homepage: www.elsevier.com/locate/vibspec A preliminary study on traceability of biodiesel mixtures based on the raw materials profiles from Brazilian regions and fourier transform infrared spectroscopy (FTIR) Victor Hugo J.M. dos Santos, Víctor Z. Pestana, Jean S. de Freitas, Luiz F. Rodrigues ⁎ Institute of Petroleum and Natural Resources, Pontifical Catholic University of Rio Grande do Sul, Av. Ipiranga, 6681 – Building 96J, 90619-900, Porto Alegre, Brazil ARTICLE INFO Keywords: Infrared spectroscopy Multivariate analysis Chemometrics Biodiesel Fuel traceability Fuel blends ABSTRACT There are still few initiatives studying applications for the classification of biodiesel derived from a mixture of raw materials, whether pure (B100) or mixed with diesel. The present study aims to conduct a preliminary assessment on the classification of pure biodiesel blends, and their mixtures with diesel at a proportion of 10%, based on FTIR spectroscopy and multivariate analysis. The work is contextualized around the monthly raw material consumption of Brazilian regions along a year. From this work, it is possible to verify that FTIR analysis, combined with multivariate methods, can be applied to classify both pure biodiesel blends and those mixed with diesel. The PCA showed great potential for recognizing the data patterns, while the HCA are able to discriminate the B100 blends from different Brazilian regions and cluster the samples according to the region biodiesel profile. For the B10 fuel blends, the OSC-PLS-DA achieved 100% sensitivity and specificity and can be applied for the classification procedure of biodiesel/diesel blends. 1. Introduction One of the current environmental concerns is the use of fossil re- sources as the main source of raw materials and energy [1]. Because of this issue, academic research has been dedicated to finding renewable alternatives to reduce dependence on mineral resources. Within this context, biofuels have been presented as viable renewable alternatives to petroleum derivatives [2]. Biodiesel is a biomass-derived fuel, obtained through the transes- terification of triglycerides with short-chain alcohols, resulting in a mixture of alkyl esters. Even though biodiesels can completely sub- stitute for diesel oil, they are almost exclusively consumed blended with conventional mineral diesel [3–6]. Brazil has a plentiful supply of renewable raw materials, resulting from extensive agricultural production, abundant natural resources and a favorable climate, which favors investment in biofuel research [7–9]. The National Program of Production and Use of Biodiesel (PNPB), promotes Brazilian biodiesel production due the need to reduce diesel imports, meet increasing energy demand, and reduce the environmental footprint of energy generation [10,11]. The publication of Law No.11,097/2005 officially introduced bio- diesel, blended with fossil diesel, into the Brazilian energy matrix, creating a mandatory and growing demand for biofuel production [12]. The initial target of 2% (B2) was reached in the early years of the program, and the current Brazilian regulatory framework (Law No. 13,263/2016) establishes a mandatory target of 10% (B10) to be reached by 2018 [11,12]. In order to meet this impending demand, the application of fast analytical methods, combined with multivariate data analysis, seems to be the best approach to perform high throughput data analysis and ensure fuel quality standards [13]. Several analytical techniques have already been applied to biodiesel assessment such as: UV–vis spectro- scopy [14], digital imaging [15,16], mid-infrared spectroscopy (MIR) [17], near-infrared spectroscopy (NIR) [18], X-ray spectrometry [19], spectrofluorimetry [20], liquid scintillation counter radiocarbon ana- lysis [21], gas chromatography–mass spectrometry (GC–MS) [22–24], nuclear magnetic resonance spectroscopy (NMR) [25,26] and isotope ratio analysis (IRMS) [27,28]. https://doi.org/10.1016/j.vibspec.2018.09.005 Received 4 April 2018; Received in revised form 30 August 2018; Accepted 10 September 2018 Abbreviations: FTIR, Fourier transform infrared spectroscopy; HATR, horizontal attenuated total reflectance; ASTM, American Society for Testing and Materials; PLS-R, partial least squares regression; PLS-DA, partial least squares discriminant analysis; PCA, principal component analysis; SVM, support vector machine; SIMCA, soft independent modeling of class analogy; HCA, hierarchical clustering analysis; C-SVM, C-support vector classification; OSC, orthogonal signal correction; MIR, mid-infrared spectroscopy; NIR, near infrared spectroscopy; NMR, nuclear magnetic resonance; ANP, national petroleum, gas and biofuels agency ⁎ Corresponding author. E-mail address: frederico.rodrigues@pucrs.br (L.F. Rodrigues). Vibrational Spectroscopy 99 (2018) 113–123 Available online 13 September 2018 0924-2031/ © 2018 Elsevier B.V. All rights reserved. T