Mondello, Casilli, Tranchida, Dugo, Costa, Festa, Dugo J. Sep. Sci. 2004, 27, 442 – 450 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Luigi Mondello 1 Alessandro Casilli 1 Peter Quinto Tranchida 1 Paola Dugo 2 Rosaria Costa 3 Saverio Festa 4 Giovanni Dugo 1 1 Dipartimento Farmaco-chimico, Università degli Studi di Messina, Viale Annunziata, 98168 Messina, Italy 2 Dipartimento di Chimica Organica e Biologica, Università degli Studi di Messina, Salita Papardo, 98166 Messina, Italy 3 Dipartimento Mo.Bi.Fi.P.A. – Sez. Zootecnica, Polo Universitario dell’ Annunziata, Università di Messina, viale Annunziata, 98168 Messina, Italy 4 Mauro Caff S.p.A., Zona Industriale, Villa San Giovanni, Reggio Calabria, Italy Comprehensive multidimensional GC for the characterization of roasted coffee beans The present investigation is based on the separation of one of the most complex food matrices: the roasted coffee bean volatile fraction. Analysis of the two main species of coffee (Arabica/Robusta) was achieved through an effective and simple sampling procedure, headspace solid-phase microextraction (SPME), and the unprecedented resolving power of comprehensive gas chromatography (GC6GC). The combination of these two techniques proved to be a powerful tool for the extraction and separation of coffee volatiles. In fact, thousands of compounds that play various roles in the con- stitution of coffee aroma profile were resolved in the 2-D contour plot, each occupying a specific position pinpointed by two retention time coordinates. The potential use of this method for the assessment of coffee quality and the detection of commercial fraud is discussed. The potential of GC6GC for identification and classification of unknowns was also demonstrated, as the formation of characteristic patterns for structurally related compounds was observed in the bidimensional chromatogram. Moreover, reproducibility results were supported by the use of an autosampler for SPME applications that allowed any inaccuracy arising from manual handling to be avoided. Key Words: Comprehensive Multidimensional Gas Chromatography GC6GC; Food analysis; SPME-GC6GC; Roasted coffee beans; Received: July 29, 2003; revised: December 8, 2003; accepted: December 9, 2003 DOI 10.1002/jssc.200301662 1 Introduction The coffee bean derives from the cherry of the coffee plant, family Rubiacae, genus coffea. The plant has its origin in north-east African countries where coffee was first used as a drink and from there it expanded to Arabia, then Constan- tinople, and Venice. By the middle of the seventeenth cen- tury it had spread all over Europe and today it is, after tea, the most popular beverage in the world with enormous social and commercial importance. The main producers of coffee beans are South American countries, in particular Brazil, while the principle consumers are, on the other hand, European countries. It is a very complex food matrix with an aroma profile due, mainly, to its volatile fraction. The two species of coffee that have acquired worldwide economic importance are Arabica (Coffea arabica) and the less expensive Robusta (Coffea canephora ex Froeh- ner). Green coffee beans cannot be consumed as such but need to undergo a series of chemical transformations induced by the process of roasting which is essential for the formation of coffee aroma. It is generally agreed that Arabica coffee has a finer flavour than the Robusta type. Commercial roasted coffee is usually available as Ara- bica, Robusta, or as a blend of both. In view of the higher price commanded by Arabica, the possibility of adultera- tion is high and so analytical differentiation of the two pro- ducts has always been an important issue. Although Ara- bica and Robusta raw beans can be easily identified because their sizes are quite different, this is not possible after the roasting process because this treatment causes approximately a 33% decrease in volume and from 18 to 22% loss in weight [1,2]. The detection of fraud in the com- mercial roasted products is achieved by the qualitative and quantitative determination of characteristic compo- nents predominant in one of the two species. Some of these are, for example, diterpene alcohols [3] or ster- ols [3,4]. Principle component analysis (PCA) methods have also been used for the differentiation of coffee sam- ples [5]. The coffee bean chemical composition depends upon several factors, such as species/variety of bean, geographical origin, soil conditions, storage, the time and roasting temperature. It is characterized by the presence of thousands of mainly volatile components, belonging to several classes of compounds and in a large range of con- centrations. Most of the volatiles are derived from non- volatile components of the green bean, which break down and react during roasting; they are numerous and varied in their aroma quality, potency, and concentration, and thus in the contribution that each makes to the unique cof- fee aroma. Some of the most common classes of com- Correspondence: Luigi Mondello, Dipartimento Farmaco-chimi- co, Facoltà di Farmacia, Università di Messina, viale Annunziata 98168 Messina, Italy. Phone: +39 090 0906766536. Fax: +39 090 0906766532. E-mail: lmondello@pharma.unime.it. 442