Argan oil, a typical Moroccan product, is prepared from the fruits of argan trees (Argania spinosa L. Skeels). It is extracted by pressing the argan kernels through a multistep process (Charrouf et al., 2002; Matthaus et al., 2010). Edible argan oil is obtained when the kernels are slightly roasted prior to grinding (Gharby et al., 2011). Unroasted kernels are used to prepare a cosmetic oil (Faez et al., 2013; Guillaume and Charrouf, 2011).

For years, argan oil has been prepared exclusively by Berber women through a traditional multistep process (Charrouf et al., 2002).

Unfortunately, this method is time consuming for producing oil batches with different organoleptic properties due to the non-reproducible roasting (Charrouf et al., 2006) and chemical composition of oil batches (Matthaus et al., 2010; Gharby et al., 2011). In addition, bacteriological concerns are frequently raised (Gharby et al., 2011).

Recently, in the women's cooperatives, argan oil has been extracted using mechanical presses producing a high quality of oil on a large scale (Charrouf et al., 2002). At present, the production of high quality argan oil is particularly important since it has been reported to promote several benefits for human health (Charrouf and Guillaume 2008). However, as a consequence of its high degree of unsaturation, argan oil is very susceptible to oxidation (Gharby et al., 2011).

Lipid oxidation in vegetable oils is one of the most important reactions that can cause the deterioration of their quality. The oxidation process of edible oil results in the development of undesirable flavors which affect the sensory quality and considerably reduce the pharmaceutical and food use of these oils (Jacobsen and Nielsen 2008; Matthaus et al., 2010). Therefore, the monitoring of the oxidative stability of argan oil is of great importance for producers and manufacturers in order to control and optimize the production process and to predict the shelf-life of the final oil (Gharby et al., 2011).

A number of accelerated methods have been developed to test the resistance of edible fats and oils to oxidation. All these accelerated methods involve the use of elevated temperatures because it is known that the rate of the reaction is exponentially related to temperature (Reynhout, 1991). Among them, the Rancimat test, due to its ease of use and reproducibility, has become frequently used and reviewed (Hasenhuettl and Wan 1992; Matthäus 1996; Mendez et al., 1996; Anwar et al., 2003; Kowalski et al., 2004; Gonzaga et al., 2007; Farhoosh 2007; Gharby et al., 2012).

This test allows for the determination of the induction period (IP) or oil/oxidative stability index (ISO-6886), which is the time before rapid deterioration of the oil occurs. The oxidative stability index can be determined at various temperatures and, therefore, the relative oxidative stability of edible fats and oils can be evaluated in a range of temperatures. In addition, previous studies have demonstrated the correlation between the stability data obtained by the Rancimat test and those determined by other sensory and/or analytical methods (Anwar et al., 2003; Adhvaryu et al., 2000; Kowalski et al., 2004; Gordon and Mursi, 1994).

A number of kinetic parameters can be determined under the Rancimat test conditions. Kinetic data can be used to characterize the differences or similarities in the oils. These data are very useful for predicting the oxidative stability of vegetable oils under various heat processing, storage, and distribution conditions (Tan et al., 2001). Also, a whole series of experiments and calculations of kinetic data for each argan oil under the Rancimat test conditions can be performed relatively quickly.

The aim of the present study was to determine the relative oxidative stabilities of argan oil obtained from roasted and unroasted kernels and the kinetic parameters of their oxidation under the Rancimat test conditions.

The initial compositions of the argan oils studied are summarized in Table 1. All samples fulfil the requirements of Moroccan Regulations for extra virgin argan oil (EVAO): acidity, PV and K270 (SNIMA 2003). The amount of free FAs is an indicator of the quality and is traditionally used as an indicator for the classification of the different commercial types from EVAO (Gharby et al., 2012). It can be seen that at the beginning of this study, all EVOO varieties showed an acidity index of 0.3%, which is much lower than the regulated 0.8% as the maximum for any EVOO. Concerning PV, this index is considered to be an indicator of primary oxidation (Gharby et al., 2011). The two argan oils considered here showed low PV values, with the cosmetic argan oil yielding the highest value at 0.9 meq O₂·kg⁻¹. In any case, none of the analyzed oils surpassed 15 meq O₂·kg⁻¹ oil, which is the limit established for EVAO (SNIMA 2003). Another quality index specified in Moroccan Regulations is K270 (Rahmani, 2005). An increase indicates that oil oxidation has begun, which could be due to some factors affecting storage conditions or to an inadequate EVAO processing (Cayuela et al., 2008). In this study, K270 coefficients were less than 0.2, compared to a maximum established by Moroccan Regulations for EVAO of 0.35.

We also found that there was no statistically significant difference among the percentages of saturated fatty acids SFA and unsaturated fatty acids (MUFA+PUFA) of the roasted and unroasted argan oil.

The initial content of tocopherol, in the oil prepared from roasted seeds was 791.2 ppm; it differs significantly from the corresponding values in oils from unroasted seeds of 667.04 ppm. The beta-carotene and the phosphorus/phospholipid contents were significantly lower in the beauty oil than in the edible oil. Beta-Carotene is a very lipophilic molecule that is known to present antioxidant properties in synergy with α-tocopherol (Palozza and Krinsky, 1992). However, no precise results can be given concerning the influence ofbeta-carotene onthe oxidative stability of edible argan (Gharby et al., 2011). The phospholipid content of the edible argan oil was found to be ten-times higher than that of the beauty oil (0.3 vs 0.006 mg·100 mg⁻¹, respectively). Interestingly, the phosphorus/phospholipid content of the beauty oil was found to be as low as that of artisanally-prepared argan oil (Gharby et al., 2011), the extraction of which is completed by a water-assisted stage. Such a hypothesis is fully consistent with the known fact that the simultaneous presence of phospholipids, β-carotene and tocopherols, which are abundant in argan oil, act synergistically as antioxidants in oils (Weng and Gordon, 1993; Lee et al., 2004; Steel et al., 2005). The k values for lipid oxidation of each argan oil at each temperature are presented in Table 2. By studying the rates of lipid oxidation as a function of temperature, an increasing rate of oxidation can be observed as temperature increases (Gharby et al., 2011). As shown in Fig. 1, the semi-logarithmic relationship between k and T values in all argan oils showed a linear dependency with good correlation of determination (R²>0.99). The lipid oxidation at low and high temperatures may go through different steps or reaction pathways, depending on the reactivity of metal ions and antioxidants at different temperatures (Tan et al., 2001). Moreover, the oil temperature affects the degree of oxygen solubility in vegetable oils which decreases by almost 25% for each 10 °C rise in temperature (Robertson, 2000). The T _Coeff values calculated from the linear functions in Fig. 1 for the argan oils ranged from 7.2×10⁻² to 7.5×10⁻² K^–1. These values could be interpreted as quantities representative of the argan oils studied. Our values are in accordance with Farhoosh and Moosavi (2007), who reported such values between 6.5×10^–2 to 7.4×10⁻² K⁻¹ (mean value: 7.17×10^–2 K^–1) in the case of other oils (canola, soybean, sunflower, olive and corn oils) (Farhoosh et al., 2008).

Table 3 provides the regression parameters for Arrhenius relationships between the reaction rate constant and the temperature for the argan oils studied. Using these regression parameters, the frequency factors, activation energies, and Q₁₀ numbers for the formation reaction of the secondary oxidation products under the Rancimat test conditions (volatile acids, mostly formic acids, with lesser amounts of acetic, propionic, and other acids) were calculated (de Man et al., 1987). These quantities for the lipid oxidation of the argan oils under the Rancimat test conditions differed significantly. This implies that the production of volatile acids under these conditions was dependent on the oil source, so it affected the assessment of the relative stability of the argan oils (Mendez et al., 1996).

The results obtained showed that a high content of β-Carotene, Phospholipid and Phosphorus would improve resistance to lipid oxidation (increase the Ea value). These would result in delaying the onset of the initial oxidation process where bond scission takes place to form primary oxidation products. The B Carotene, Phospholipid and Phosphorus contents of the argan oils (Table 1) explain the observed trends in various activation energies (Table 3) to a certain extent. However, it was observed that several other factors affecting the oxidative stability, e.g. the tocopherol content influenced the susceptibility to oxidative degradation to varying extents. The frequency factors with a trend similar to that of the Ea values for the argan oils studied increased from 7.89 10¹³ for RAO to 5.66 10¹⁴ for UAO.

The magnitude of the temperature effect on the oxidation rate of the argan oils was evidenced by the Q₁₀ numbers. In general, a higher Q₁₀ number implies that a smaller temperature change is needed to induce a certain change in the rate of lipid oxidation. As can be seen in Table 3, the Q₁₀ number increased from 2.11 for RAO to 2.16 for UAO.

The ΔH and ΔS values estimated based on the activated complex theory and the corresponding regression parameters are summarized in Table 4. The high correlation of determination (R²>0.99) indicated an adequate fit and characterization of the temperature dependence of lipid oxidation using the activated complex theory.

The ΔH and ΔS values for the argan oils studied ranged from 87.47 kJ·mol⁻¹ and −46.81 J·mol⁻¹·K⁻¹ for roasted argan oil to 90.87 kJ·mol⁻¹ and −30.35 J·mol⁻¹·K⁻¹ for unroasted Argan oil, respectively. In their study on the determination of the oxidative stability of rapeseed, sunflower and soybean oils by the Rancimat test, Kowalski et al., calculated the amounts of 82 kJ·mol⁻¹ and −52.7 J·mol⁻¹·K⁻¹, 84 kJ·mol⁻¹ and −42.8 J·mol⁻¹·K⁻¹, and 74.9 kJ·mol⁻¹ and −70.2 J·mol⁻¹·K⁻¹ for their ΔH and ΔS values, respectively (Kowalski et al. 2004). The negative values for ΔS indicate that the activated complexes are more ordered than the molecules of the reactants, and its greater negative values indicate fewer numbers of species in the activated complex state, and hence a lower probability of the activated complex toward lipid oxidation and therefore a slower rate.

The experimental results allowed us to draw the following conclusions: An increasing rate of oxidation could be observed as temperature increases and, edible argan oil can be stored much better than cosmetic oil.