Fast almond drying is a fundamental operation from the commercial point of view (deficient drying reduces the operational profitability, and the shelf life of nuts susceptible to rancidity) and sanitary point of view (adequate drying prevents the growth and spread of imperfect fungus Aspergillus flavus, and therefore the aflatoxin production and accumulation). Almond drying can be done in different ways: direct sun exposure, hot air oven, using a fan, in a hot air dryer, etc (Piscopo et al., 2011).

Different extraction methods can be used for almond oil extraction, although, as with other seeds, solvent extraction will provide the highest industrial yield (Guerra and Zúñiga, 2003). Traditional equipment uses high temperatures and chemical products, reducing the quality of the oil due to the appearance of undesirable flavors and the inactivation of vitamins and active substances that appear in the raw material (Sineiro et al., 1995), forcing the posterior necessity of refining the oils, so they could not be defined as virgin oils. Matos and Acuña (2010) evaluated three main influence parameters: extraction temperature, size of almond particle and solid/solvent proportion regarding yield, and defined the optimal conditions as, 90 °C, 0.5 mm and 1:3 proportion, respectively, reaching an oil yield of 44.5%. This yield can be improved if samples are irradiated with ultrasounds of 42 kHz (Sharma and Gupta, 2006); observed data using electron microscopy indicate that ultrasound facilitates the development of micro-fractures and the cell walls rupture. It is likely that this effect was also associated with a reduction in particle size. The process consists of aggregating adequate quantities of ammonium sulphate and t-butanol to the vegetal material. The protein precipitates as an inter-phase between the superior organic layer and the aqueous layer underneath. The separation of proteins is simultaneously accompanied by oil liberation in the t-butanol phase (Sharma and Gupta, 2004).

In recent years, supercritical fluid extraction (CO₂) has improved its conception as a viable alternative to conventional solvent extraction methods (Mendiola et al., 2007); the use of lower temperatures and pressures results in higher quality products (Marzouki et al., 2008). Femenia et al. (2001), using pressures of 330 bar and temperatures of 50 °C, extracted the oil contained in raw almonds, raw peeled almonds, and roasted almonds, obtaining oil percentages of 15–16%, 27–33% y 49–64%, respectively. Leo et al. (2005) also extracted almond oil using this system, but using pressures of 350 to 550 bar, temperatures of 35 to 50 °C and solvent rates of 10 to 30 kg·h⁻¹, and observed that the increase in extraction pressure, and temperature caused an increase in oil yield. It was also observed that, to equal flow and pressure rate, the temperature increase caused an increase in yield of almost four times higher. An explanation to this phenomenon is that it produced an increase in oil solubility in CO₂. Experimental results were used to deduct that oil production, the initial stage of extraction, increased with an increase in the CO₂ flow rate of 10 to 30 kg·h⁻¹, with constant pressure and temperature. Thus, oil production increased with the increments in pressure, temperature and flow rate. Later on, Ma et al. (2007) studied the factors that influence the bitter almond oil extraction, finding optimal extraction conditions: extraction pressure = 35 MPa, extraction temperature = 50 °C, CO₂ flow rate = 24 l/h, almond particle size = 0.6 mm and extraction time = 2 h. The factor sequence that affects extraction is: almond particle size > extraction time > extraction pressure > CO₂ flow rate > extraction temperature. Under these conditions, almond oil yield reaches 53%.

An alternative to solvent use is the use of pressing, with both hydraulic and screw presses. As a general rule, the hydraulic press offers lower profitability due to the high demand of labor force; although the obtained oils would possibly find an increased acceptance by consumers because the oils maintain their physico-chemical and sensorial properties better (Da et al., 2014; Sena-Moreno et al., 2015). Extraction using a screw press demands a previous thermal conditioning of the material, resulting in a decrease in the quality of the final product (Rohne, 1971). By comparing the almond oils extracted using hydraulic press with the ones extracted using solvents, several authors found that the first ones had a higher tendency toward auto-oxidation (Dugo et al., 1979; Salvo et al., 1980). Prats (2000), obtained yields of 40–45%, using sweet almonds and the hydraulic press. The oil obtained using this method was quite fluid and had a pleasing and light, sweet flavor, and were more stable than others to oxidation processes, which is not in agreement with the findings of Dugo et al. (1979) and Salvo et al. (1980).

Regarding the screw press, although it is generally defined as a cold pressing system, it requires a temperature increase (preheating) to obtain better results. This increase produces a better oil separation, which can affect extraction yield. In some screw presses, temperature is hard to control and, therefore, the yield presents higher variations (Álvarez-Ortí et al., 2012). Recently Martínez et al. (2013), evaluated the combined effects of almond humidity (4, 6, 8, 10 y 12%) and preheating press temperature (20, 40 y 60 °C) on the yield and the quality of obtained oils, and observed that the largest quantity of oil (79,3%) was obtained with a seed humidity of 8% and a preheating temperature of 40 °C. The increase in humidity from 4 to 8% resulted in a higher yield but a posterior increase from 8 to 12%, caused a light decrease. All extraction conditions used were compatible with an acceptable physicochemical quality.

Almond oil extraction using supercritical fluids (CO₂) and pressing (both hydraulic and screw press) provides a product that is fit for human consumption with the pleasant sensorial characteristics that belong the initial product (the almond), consequently no refining is needed, making the product a virgin oil.

In the almond lipid fraction five fatty acids are predominant: oleic acid (C18:1), linoleic acid (C18:2), palmitic acid (C16:0), palmitoleic acid (C16:1) and stearic acid (C18:0). These fatty acids appear in descending order of composition and constitute 95% of the total (Table 2). This fraction is complemented by eight less common fatty acids (Saura et al., 1988; García et al., 1996; Abdallah et al., 1998). Regarding the most abundant fatty acids, the unsaturated fatty acids represent about 90% of the fatty acid content. Regarding the unsaturated fatty acids, monounsaturated represents the highest proportion with respect to polyunsaturated (Saura et al., 1988; Abdallah et al., 1998; Kodad, 2006).

The greatest variability in fatty acid contenta has been observed by comparing different genotypes. Palmitoleic and stearic acids show greater variability (coefficient of variation of more than 10%), followed by linoleic and palmitic acids (coefficients of variation of more than 4% and 7%, respectively) and, finally, the oleic acid, with a coefficient of variation smaller than 2% (Kodad, 2006). Yada et al. (2013), also discovered significant cultivar differences by comparing saturated fatty acids (values ranged from 3.2 to 4.7 g·100 g⁻¹ fresh almond) and the polyunsaturated fatty acids (values ranged from 9.4 to 15.1 g·100 g⁻¹ fresh almond). Comparing harvests (in two consecutive years), highest variability was found in palmitoleic and stearic acids (more than 10%); while oleic acid showed low coefficients of variation, with contents ranging from 63.1% to 78.7% of the fatty acid content. Linoleic acid values ranged from 12 to 27% of the fatty acid content (Kodad, 2006). Kodad et al. (2014), found that for traditional Spanish cultivars, the effect of genotype, year and genotype x year interaction, was significant in relation to variability in the contents of stearic, palmitoleic, linoleic and oleic acids. However, apart from palmitoleic acid, the magnitude of difference was insignificant in comparison with the variability among genotypes.

García et al. (1996), discovered that the lowest value for oleic acid was found in Californian cultivar Texas (57.5%) and the highest in the Spanish cultivar Ramillete (74.2%). Regarding linoleic acid, the opposite was observed. A possible explanation for the presence of the highest oleic acid concentrations in European cultivars, compared to the American ones, is the positive correlation observed between the shell thickness and the oleic acid with r = 0.44, which matches the lower oleic acid content in the American cultivars than in the ones from the Mediterranean basin (Askin et al., 2007). Nevertheless, the most important part of García et al. (1996) research was the statistical analysis used to classify almond tree cultivars form different origins, based on their similarities in the fatty acid profile. This research was completed by incorporating classification criteria, minority fatty acid concentrations (Martín et al., 1998) and triglyceride compositions (Martín et al., 1999).

Sathe et al. (2008), analyzing Californian cultivars, found that palmitic, oleic and linoleic fatty acids were predominant in the lipid composition regardless of the cultivar, crop origin and harvest year. The ranges of these fatty acids were from 5.15% to 6.65%, from 59.52% to 73.80%, and from 19.49% to 33.29%, respectively; mean values for oleic and linoleic acids were 65.77% and 27.18%. Palmitoleic acid varied from 0.31% to 0.57%, stearic acid from 0.24% to 1.66%, α-linolenic acid from 0.05% to 0.09%, and the araquidic acid from 0.03% to 0.07%. All these acids showed presence in all the studied samples. Similar values for the three main fatty acids were previously found in Iranian almonds. 75.4% of the total fatty acid content was oleic acid, 19.4% was linoleic acid and 5.3% palmitic acid (Safari and Alizadeh, 2007).

Another important aspect is the existing relationship among the different fatty acids. The most frequently studied relationship is the oleic/linoleic ratio because of its nutritional importance, which shows a wide range of variation depending on the considered author; from 1.96 to 3.05 (Saura et al., 1988), from 2.35 to 6.56 (Kodad, 2006) or from 1.8 to 3.8 (Sathe et al., 2008). The correlation observed in both fatty acids is very high and negative, with r values from −0.92 (Askin et al., 2007) to −0.99 (Sathe et al., 2008).

In ripe almonds, fatty acids appear mainly in the form of triglycerides (Nassar et al., 1977). In fact, almond oil is the conventional nut oil with the highest content in triglycerides, about 98% (Miraliakbari and Shahidi, 2008), which results in a low acidity index (Saura et al., 1988). However, its composition has not been the issue of many research works.

The kind and quantity of oil triglycerides determine the final physical and functional properties of them (Jahaniaval et al., 2000). Prats (2000) detected nine triglycerides in ten studied cultivars from seven different production zones, observing this order, from high to low contents:

OOO > OLO > OLL > POO > PLO > SOO ≅ LLL ≅ PLL > PLP

where O represents oleic acid, L linoleic acid, P palmitic acid and S stearic acid.

The first five are predominant triglycerides and the last four are less common. Due to the fact that oleic acid is the most abundant fatty acid, triglycerides containing this fatty acid are also more common. OOO have the highest content with values around 11.3 to 31.3 mg·100 g⁻¹ of dry sample, and the content in OLO varies from 15.7 to 6.9 mg·100 g⁻¹ of dry sample. Hereafter OLL and POO triglycerides appear. The PLP triglyceride is the one that appears in the smallest quantities. The OOO triglyceride has the highest correlation with the rest of triglycerides and moreover it has a negative sign, except for POO and SOO, which makes sense because they appear in the largest proportions in almond oil.

With a similar approach, Martín (1999) conducted a previous study on 19 representative cultivars from the main almond production zones in the world, finding the same triglycerides in the same concentration order. Triglycerides OOO and OLO together amount to 60% of the triglyceride total. The main aim of this research was the use of the concentration data of different triglycerides to determine the cultivar oil origin with the two-fold purpose of characterizing the product to avoid industrial frauds and to assign genotypes into groups with homogeneous behavior concerning cultivar selection. In general, triglyceride composition in almond oil was very similar in the different cultivars.

Holcapek et al. (2003) found 18 triglycerides in almond oil, probably due to the better resolution in the analysis carried out with the proposed working method. Nevertheless, the triglyceride order by content has not been the same as in other revised research: OLO (28.0%), OLL (27.6%), OOO (13.3%), LOP (11.3%) LLL (8.7%), LLP (4.8%), OOP (2.7%), SLO (1.8%), SOO (0.6%), PLP (0.5%), OOMo (0.5%), OLLn (0.1%), LLMo (0.1%), OLMo (0.1%), GLO (0.1%), POP (0.1%), OOMa (0.05 %) and GOO (0.05%), where Mo is heptadecanoic acid, Ln is linolenic acid, Ma is margaric acid and G is gadoleic acid.

Cherif et al. (2004), found 10 triglycerides with LOO that had not been identified by other authors before. They did not, however, find PLL in an interesting research about the fatty acid evolution and triglycerides throughout the development and ripeness of the almond seed. They conclude that differences in the triglyceride profile can be useful to distinguish cultivars. However, the triglyceride composition found in almond oil was similar for all the studied cultivars. Triglyceride total contents for major components (OOO, LOO, LLO, LOP y POO) were very similar in the three studied varieties, and the OOO and LOO added 60–70% of the total of triglycerides, which is in accordance with Martín (1999). Triglyceride differences suggest that OOO triglyceride can be used to discriminate among cultivars, for example, to detect adulterations in almond oil.

Later, by using a reverse non aqueous HPLC method with an acetonitrile -2-propanol gradient, Holcapek et al. (2005), identified unequivocally the highest number of triglycerides ever reported in analyzed oils. This allowed them to detect 24 triglycerides in almond oil. The obtained values are in line with those obtained in the consulted literature.

Barreira et al. (2012), made a study about almonds that had been collected over three months in Trás-os-Montes (Portugal), with the aim of discovering differences between the Protected Designation of Origin (DOP) Amêndoa Douro and the non-DOP commercial cultivars. The comparative test provided general conclusions for almost all the evaluated cases. The DOP Amêndoa Douro cultivars had the highest content in OLL and LLP triglycerides. But apart from previous considerations, results confirmed the prevalence of OOO and OLO triglycerides. In general, detected profiles were similar to previous research (Martín, 1999; Prats, 2000).

Recent studies carried out using wild almonds in Iran, have resulted in the triglyceride composition: OOO (47,27%), POO+SOL (26,25%), OOL+ PLnP (10,67%). The triglyceride profile found was very similar to the olive oil profile (Givianrad et al., 2013).

The advantage of using a triglyceride analysis, in comparison to the fatty acid profile, is that a stereo-specific distribution of fatty acids in the glycerol molecule is genetically controlled and, as a result, the information content of intact triglycerides is generally higher (Aparicio and Aparicio-Ruiz, 2000; Bail et al., 2009).

Pasini et al. (2013), established a lipid classification, in which unsaponifiable fraction and complex lipids (phospholipids and glycolipids) appear. These fractions are quantitatively small but with huge importance from the biological and nutritional point of view. The almond composition in these compounds, denominated fat-soluble bioactives, consists of tocopherols, tocotrienols, phospholipids, sterols, phytosterols, phytostanols, sphingolipids, squalene and terpenoids. This group of compounds has been included by several authors inside the phytochemical concept (Alasalvar and Pelvan, 2011).

The unsaponifiable fraction is basically formed by sterols, methylsterols, aliphatic alcohols, and fat-soluble vitamins (Prats, 2000). More representative sterols are sitosterol (80–86%), campesterol (2–4%) and traces of stigmasterol (Garcia et al., 1978). Generic studies about nuts have reported an unsaponifiable substance content of 0.44 g·100 g⁻¹ of extracted oil, with a range from 0.35 to 0.53 (Kornsteiner et al., 2006).