Pattern recognition applied to mineral characterization of Brazilian coffees and sugar-cane spirits

https://doi.org/10.1016/j.sab.2005.02.013Get rights and content

Abstract

Aluminium, Ca, Cu, Fe, K, Mg, Mn, Na, Pb, S, Se, Si, Sn, Sr, and Zn were determined in coffee and sugar-cane spirit (cachaça) samples by axial viewing inductively coupled plasma optical emission spectrometry (ICP OES). Pattern recognition techniques such as principal component analysis and cluster analysis were applied to data sets in order to characterize samples with relation to their geographical origin and production mode (industrial or homemade and organically or conventionally produced). Attempts to correlate metal ion content with the geographical origin of coffee and the production mode (organic or conventional) of cachaça were not successful. Some differentiation was suggested for the geographical origin of cachaça of three regions (Northeast, Central, and South), and for coffee samples, related to the production mode. Clear separations were only obtained for differentiation between industrial and homemade cachaças, and between instant soluble and roasted coffees.

Introduction

Globalization has caused a revolution in the consumption habits all over the world. As well as in the internal and external markets, the improvement of the products quality, the decisive aspect of commercial barriers, and expansion of markets accentuate the pressures. The quality certification, based on standards, patterns, and technical specifications, will be the pre-requirement of any product. The search for superior levels of quality, time, and competitiveness is a constant concern of economic agencies, and in the agribusiness sector, it could not be different. The national producer has to be attentive on not losing market to other countries.

Among the representative products of the Brazilian agribusiness, two are prominent in a differentiated way: coffee and sugar-cane spirit, denominated Brazilian cachaça. The former for still highlighting Brazil as a major world producer, consumer, and exporter, and the latter, for waking up recently the interest on the increase of its export front to the growing search by external markets.

Coffee is the world most popular beverage after water, with over 400 billion cups consumed annually [1]. It is one of the most important agricultural products in the international trade, putting into motion approximately US$35 billion per year and being supplanted only by petroleum [2]. In 2004, the coffee industries estimate total sales–internal market and exports–of approximately US$1.5 billion [3]. In the case of cachaça, official and exclusive denomination of the spirit distilled of sugar cane produced in Brazil, it comes conquering new markets. With a production of 1.5 billion liters/year, it generates profits of US$700 million and an expected growth of 27% in the exports in this year, with the European countries being the largest importers [4].

Therefore, it is easy to understand the huge relevance of the availability of suitable analytical methods to characterize these products of great consumption for millions of people worldwide. Additionally, the determination of geographical origin of commodities and food products is becoming an increasingly active research area, focused on both geographical authenticity and adulteration of foods [5]. Chemical analyses in conjunction with pattern recognition techniques provide interesting tools for the study of the quality and origin of food products [6].

The objective of this study was to evaluate the feasibility of using a multielement analysis combined with pattern recognition tools, in order to contribute to the development of the Brazilian agribusiness, supplying important information regarding the identity and quality of the national product. Thus, 48 samples of coffee and 156 samples of cachaça from different geographical sources were selected and analyzed with relation to their mineral content. The determination of Al, Ca, Cu, Fe, K, Mg, Mn, Na, Pb, S, Se, Si, Sn, Sr, and Zn was carried out by axial viewing inductively coupled plasma optical emission spectrometry (ICP OES). Afterwards, pattern recognition techniques were applied to characterize samples with relation to the geographical origin and production mode (industrial or homemade and organically or conventionally produced).

Section snippets

Instrumentation

An inductively coupled plasma optical emission spectrometer with axially viewed configuration (VISTA AX, Varian, Mulgrave, Australia) was used for Al, Ca, Cu, Fe, K, Mg, Mn, Na, Pb, S, Se, Si, Sn, Sr, and Zn determinations. This equipment involves a simultaneous charge coupled device (CCD) detector that allows readings from 167 to 785 nm. The pre-optical system was purged with Ar in order to enable readings below 190 nm. The polychromator was thermostatized at 34 °C and purged with argon.

Coffee analysis

Although the class of the samples was a priori known, a preliminary study based on an unsupervised pattern recognition method was applied to observe the structure of the data sets. Thus, initially an HCA was applied on the digested coffee samples data set to observe any natural grouping feature. The resulting dendrogram is shown in Fig. 1. Two separated clusters appear, a bigger one containing the roasted coffee samples and a smaller, at the bottom with dots, containing the six instant soluble

Conclusions

According to the results, pattern recognition analyses were able to differentiate some important features in coffee and cachaça samples based on their metal contents. For coffee samples it was possible the differentiation between instant soluble and roasted coffee, the latter being characterized by greater amounts of Cu and Zn, while the instant soluble presented greater contents of Na, Mg, K, and S. It was also possible for the classification of samples according to the production mode, being

Acknowledgements

The authors wish to express their appreciation to Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for financial support and for the research scholarships provided.

References (14)

There are more references available in the full text version of this article.

Cited by (75)

  • Simple ICP-OES based method for determination of selected elements in brewed ground and soluble coffees prior to evaluation of their intake and chemical fractionation

    2018, Food Chemistry
    Citation Excerpt :

    SCs were dissolved in water and also acidified with HCl at 1% (Oliveira et al., 2012), HNO3 at 1.0 mol L−1 (Pohl et al., 2014a) or without it (Stelmach et al., 2016). The use of simplified sample preparation prior to ICP-OES analysis of infused coffees on the total content of different elements was rarely reported and included only acidification with HNO3 and at lower concentrations, i.e., 0.014 mol L−1 in case of GCs (Fernandes et al., 2005), or 0.25 mol L−1 in case of GCs and SCs (Szymczycha-Madeja et al., 2016). In present work, to eliminate eventual memory effects and reduce washout time between the samples, the final concentration of HNO3 in brews of GCs and SCs before analysis was increased from 0.25 mol L−1 to 1.3 mol L−1.

View all citing articles on Scopus

This paper was presented at the 8th Rio Symposium on Atomic Spectrometry, held in Paraty, RJ, Brazil, 1–6 August 2004, and is published in the special issue of Spectrochimica Acta Part B, dedicated to that conference.

View full text