Immature and mature fruits of P. sapota were randomly collected from the producing region located in the municipality of Emiliano Zapata in the State of Veracruz, Mexico. Fruits from each batch were weighed and the anatomical parts of each of them, peel, pulp, stone, and seed, were removed in order to estimate the average percentage of each relative to the total mass of the fruit.

The peel, pulp and seed of the mamey were physically and chemically analyzed in terms of their contents in moisture, ash, crude fiber, proteins and fats by using the official analysis techniques (Horwitzs, 1995). Lots of pulp and seed were studied in the two stages of development considered and characterized further into other relevant components. For the pulp, the stage of its physiological development was characterized through titratable acidity (by neutralization with an alkaline solution), reducing sugars and total sugars (with the Lane and Enyon method), pH (potentiometric method) and tannins (with the method of Folhin-Denis). The seeds were further analyzed for their tannin and cyanide contents (with the Folhin-Denis method and alkali titration) according to official analysis techniques (Horwitzs, 1995).

The oil was extracted in a Soxhlet apparatus using hexane. The extraction was performed during periods of 6 hours from dehydrated and milled seeds (with about 15% humidity) until the samples were completely exhausted. The solvent was removed from the resulting micelle in a rotary laboratory (Büchi RE 111) at reduced pressure.

The lipid fraction was stored in a container protected from light under refrigerated conditions until processing and/or analysis.

The obtained oil was purified by an adaptation of the Wesson method (Mehlenbacher, 1970). In this procedure, the oil was dissolved in petroleum ether (1:5 mass/volume) and 1 mL of a 14% potassium hydroxide solution was added to each gram of oil, with vigorous agitation; subsequently, 50% ethyl alcohol was added at 1:2.5 (mass/volume) based on the mass of oil, with agitation and then resting, until separation of the phases. The ether layer containing the neutral oil was processed to recover the purified oil. It was then stored under refrigeration (-5°C) and protected from light until processing and/or analysis.

The refractive index, saponification value and iodine value of MSSO were measured using official analysis techniques (Horwitzs, 1995).

The oil was converted into the methyl esters corresponding to its fatty acids; then they were separated and analyzed through gas chromatography and mass spectroscopy in accordance with the methodologies described by Horwitzs (1995); Bannon et al. (1982) and Hendrikse et al. (1994).

The boron trifluoride method (Horwitzs, 1995) was used for oil esterification.

The analysis of fatty acid methyl esters was carried out in a Hewlett Packard model 5890 series II gas chromatograph (Hewlett Packard Co., Palo Alto, CA, USA), coupled to a JEOL SX-102A mass spectrometer with an electron ionization detector. The system had an AT Silar capillary column (Alltech Associates, Inc., Deerfield, IL, USA) 30 m long and 0.32 mm internal diameter and 0.25 μm film thickness. A database that contains about 75 000 compounds supplemented with the internal standards of methyl ester fatty acids (Sigma) was used for the identification of compounds. Helium was used at a flow of 0.4 mL·min⁻¹ and 38 kPa; the inlet temperature was 250 °C and the initial temperature of the oven was 100 °C. The temperature program was: 2 min at 100 °C and increments of 4 °C·min⁻¹ till 180 °C was attained. The mass analyzer operated in a mass range of 10:450 m/z.

MSSO fractionation was performed through a homogeneous crystallization by applying slow cooling and spontaneous nucleation (O'Brien, 1998; Timms, 2005). The procedure included: a) MSSO heating to 90 °C for 10 min; b) controlled cooling for crystal nuclei formation in a circulation and refrigeration bath (Polyscience Model 911), for which the temperature was gradually reduced from 30 °C until the appearance of crystalline nuclei (26.5 and 23.8 °C); c) crystalline aggregates were allowed to grow over a sufficient time period at each temperature for the MSSO to form a sufficient solid fraction; d) the separation of the solid and liquid phases at the established temperatures was done in a Hettich (Tuttlingen, Germany) refrigerated centrifuge, model Universal 32R. The recovered fractions were placed in amber bottles and stored under refrigeration (-5 °C).

The tender (immature) and mature fruits showed qualitative differences under observation. The immature fruit, for example, presented a dark-brown, hard and gritty epicarp like that of the ripe fruit, but not fully differentiated from the pulp (mesocarp), which was light-brown and greenish in color, with a characteristic odor and an astringent taste. The stone was not well shaped and had a light brown color. The seed of the tender mamey showed relatively defined tissues and the presence of a weak seed coat (integument) with a higher moisture content than the mature one.

The ripe sapote fruit presented a distinct shell of mesocarp. The pulp was orange-red in color with a characteristic taste and odor. It had a well-defined stone, dark brown in color with a white ventral line, and a well-defined seed showing the presence of integument; seed tissues were firmer than the immature ones, and it had a lower water content than that of the immature fruit.

Table 1 presents the percentage based on the mass of the main constituents of P. sapota fruit in its immature and mature states. The traditionally non-edible portions accounted for about 30% of the total mass of the fruit; the peel was between 16.21 and 14.54%, the whole stone about 14%, and the seed inside of the stone was 10.07 and 8.83% of the whole of the unripe and ripe fruits, respectively. The average mass of mature fruits was slightly more than twice the weight of the immature sapotes.

The peel, morphologically well formed only in mature fruits, was composed mainly of crude fiber (81.32% db) and had lower contents of protein (6.48%), lipids (6.3%) and ash (3.89% db) (Table 2).

The edible portion, the pulp, showed high moisture contents and carbohydrates, with simple sugars in ripe fruits and polysaccharides in immature ones.

The results in Table 2 for the pulp portion show the effects of the biological and structural changes due to maturation: an increase in pH and simple sugars at the expense of the starchy fraction; a lower content of moisture, organic acids and phenolic compounds (Table 2).

The almond-like portion or sapote seed was richer in lipids than the peel portion, with high values even in the earlier stage of physiological maturity (28.31%). In mature fruits lipids represented an average of 44.41% (db), with pieces with close to 50% of oil having been found in the sapote seeds. This is remarkable if one takes into account that some important traditional sources of vegetable oils have similar or lower contents: 3.5–5% in corn, 18–20% in soybeans, 22–24% in cotton, and safflower and peanuts with 25–37% and 45–50% of oil, respectively. Other notable components in the seed were protein and crude fiber (Table 2).

The purified MSSO had a refractive index of 1.463, a saponification value of 198 mg KOH/g and an iodine value of 60 gI₂·100 g⁻¹ of oil.

The seed oil composition in immature and mature mamey is shown in Table 3. This composition is relatively simple, with the main fatty acids being palmitic, stearic, oleic and linoleic (10.7, 26.5, 53.5 and 5.6 g of FA·100 g⁻¹ of oil, respectively, for immature fruits and 10.5, 28.6, 48.6, and 10.7 g of FA·100g⁻¹ of oil, respectively, for mature ones). The ratio between saturated fatty acids (SFA) monounsaturated (MUFA) and polyunsaturated ones (PUFA) was 39.9, 48.6 and 10.7, respectively, for g·100 g⁻¹ studied in mature fruits. These values are similar to those of some vegetable oils for industrial use such as palm oil, with the difference that, in the case of the mamey, stearic is its main saturated FA, similar to cocoa butter, one of the major industrial fats. However, in tropical zones at room temperature, sapote seed oil is a liquid and not a fat, due to the presence of high levels of oleic and linoleic acids in it.

Figure 1 shows the crystallization and melting curves of MSSO. The thermal behavior in both phase changes is generally simple. The change from liquid to solid has a well-defined peak at -2.3 °C and another in shoulder form at -26.2 °C. The crystallization onset temperature occurs at 4.5 °C and this process extends over a range of 41.5 °C. The crystallization enthalpy was 32 J/g. The melting curve of the oil solidified by rapid cooling began at -35.2 °C and it presented three well-defined maximum peaks (-2.7, 13.1 and 34.4 °C) and a small shoulder (-31.4 °C); the process comprises a melting range of 73.2 °C with a melting enthalpy of 28.6 J/g (Table 4). Also, the thermal behavior of MSSO resembles that of palm oil (Tan and Che-Man, 2000); but the sapote oil has a broader melting range and that explains why this oil is liquid at ambient temperatures in tropical regions. The calorimetric analysis shows the existence of three groups of acylglyceride fractions with melting ranges at low, intermediate and high temperatures. This enables the attainment by fractionation of fats and oils with more specific thermal characteristics and phase behaviors, usable from the perspective of the extensive use of natural vegetable fats derived without the intervention of chemical processing and/or severe heating.

The dry fractionation of MSSO oil yielded two solid fractions and a residual liquid fraction (RLF). The first two were obtained at 26.5 and 23.8 °C (F1 and F2, respectively) and constituted 21.0 and 19.6%, respectively, of the total oil subject to fractionation. The RLF remained liquid at temperatures of 23.8 °C, representing about 59% of the processed MSSO.

Figure 2 shows the melting curves of the fractions obtained by heating them (10 °C·min⁻¹) after their solidification by rapid cooling till -60 °C (F1A, F2A and RLF), and the melting curves of the samples of these solid fractions after they were tempered (F1B and F2B). This last circumstance allowed for recording the effect of the polymorphism of the MSSO solid fractions.

The fractionation efficiency can be evaluated through the melting curves in terms of their different phase behavior, that is, between the RLF and the two solid fractions; however, it is possible to notice some similarities in the curves for the solid fractions. It is known that the process of dry fractionation requires the formation of large crystals to ensure efficient separation of the fractions. Said crystals however, forming larger clusters, trap some of the remaining liquid oil or olein, which, added to the conditions prevailing during the separations of the phases, may result in fractions of “light stearin” or low yields of the liquid fraction (O'Brien, 1998).

The melting curves in Figure 2 show that RLF culminates by melting at 21 °C; the other fractions end up melting at 34.9 °C (F2A) and 39.5 °C (F1A). Molecular rearrangements favored by slow cooling of the fractions resulted in changes to more stable crystalline forms or polymorphs (β and β′) than α. As it is known, lipids (fatty acids, acylglycerides, and fats and oils) are meta-stable and exhibit polymorphism during processing and storage. Generally and most commonly they solidify in three forms, the α form being the polymorph with the lowest melting point (it is obtained by rapid cooling from after melting). The β′ polymorph, with a higher melting point than α, is generated under certain conditions of temperature or by a spontaneous transition from the α form. The β crystals are the most stable and are transitions of the above forms (Hagemann, 1988). In regard to such transitions, the molecular mobility, stability and their relationships are critical aspects for controlling the quality of the fatty products that are elaborated with these materials (Roos, 1995; Solís-Fuentes et al., 2005). The vast majority of phase transitions of lipids in foods, drugs and cosmetics associated with the behavior of processing, storage and use are established between the solid and liquid states and the different polymorphic forms of these materials (Roos, 1995; Sato, 1999).

Table 4 presents a comparison of the observable transition temperatures during the MSSO melting and its fractions. The stabilized fractions (F1B and F2B) exhibited higher end melting temperatures than their respective non-stabilized fractions (F1A and F2A), 44.8 and 49.0 °C, and 39.5 and 34.9 °C, respectively.

Figure 3 presents the solid-liquid relationship profiles of MSSO (measured as the percentage of solids) and its non-stabilized and stabilized fractions. As it is known, the ratio between the solid and liquid phases in the fat determines its consistency (i.e. hardness or firmness characteristics). These profiles include the four temperature zones in the fluency of triacylglycerols; this indirectly allows us to appreciate the lubrication properties and the primary structure of fats and oils in food products, that is, at refrigeration temperature (10 °C), environmental temperature (21.1 °C), body temperature (36.3–37.8 °C) and of the preparation of fatty products (40 °C) (O'Brien, 1998).

According to its phase behavior, as shown in Fig. 3, the mamey seed contains an oil with the characteristics of a liquid fat because at 20 °C it has 6% solids; at 10 °C 25% and at 0 °C almost 50%.

Except for MSSO (and RLF, which is not presented in Fig. 3) which has a solid profile that gives the oil a liquid consistency at ambient temperatures in tropical regions (20–22 °C), all fractions showed solid consistency with a solid index >35%, which means that over a third of their acylglycerides in the solid-crystalline phase are suspended in less than two-thirds of liquid triglycerides at that temperature. This solid/liquid ratio decreases differently in these fats as the temperature rises and is minimized for all at around 40 °C (Fig. 3).

The stabilized samples, with polymorphic effects, showed important shifts in their solid profiles towards higher temperatures.

The solid fractions of MSSO showed different degrees of plasticity than those required for the manufacture of confectionery products, bakery shortenings and frying fats.

Figure 4 shows the solid profiles of blends of MSSO and its F2 solid fraction, with better known fats, such as cocoa butter (Theobroma cacao L., with a broad spectrum of industrial applications in the fields of food, pharmaceuticals and cosmetics), and mango seed fat (Mangifera indica, L., widely investigated in recent years; Solís-Fuentes and Durán Bazúa, 2004; Jahurul et al., 2014).

Curve B shows the solids profile of tempered cocoa butter (CB) with its characteristic deep slope, a rapid decrease the percentage of solids at around human body temperature. This property is highly appreciated in the vast number of applications of this fat.

Curves C and D correspond to tempered mixtures whose weight contains 25% of the MSSO solid fraction (F2A), with 75% of CB in one case and 75% of mango seed fat (MSF) in the other, respectively. In both cases, a great similarity in the solids profile to that of the fat model CB is evident. In the first case, the approximation to the CB curve is very close, and represents the possibility of partial substitution with this fraction of MSSO in products that use CB as the fat ingredient. In the second case (curve D), it is shown that the mixture of two non-conventional natural fats such as MSF and the solid fraction of MSSO have a phase behavior close to that of CB.