Thermal and chemical characterization of fractions from Syagrus romanzoffiana kernel oil

2.1. Oil extraction

The jerivá fruits (Figure 1) were collected from the south-east region of Brazil (Lavras, Minas Gerais state) directly from the ground in the mature stage, broken in a hydraulic press (MPH-15, Marcon) and the kernels were manually separated. The JKO was obtained using an expeller press (Home Up Yoda), with an oil yield of 42.6%, and centrifuged for 5 minutes at 4000 rpm (relative centrifugal force of 2150 g) to remove fine particles. Figure 1 illustrates the treatment process of treatment and characterization of the samples.

2.2. Fractionation

2.3. Fatty acid composition

A fatty acid composition analysis was performed according to the AOCS Ce 2-66 method (AOCS, 1993AOCS (1993). Official Methods and Recommended Practices of the American Oil Chemists Society. Champaing.). Samples were transesterified in methyl esters with potassium hydroxide in methanol and n-hexane. All reagents and solvents (analytical grade) were from Vetec. The methyl esters were analyzed by gas chromatography (GC-2010 - Shimadzu) equipped with a flame ionization detector and an SPTM-2330 capillary column (30 m x 0.25 mm x 0.20 µm). Chromatographic conditions were split ratio 1:100, initial column temperature 140 °C, heating from 140 °C to 250 °C for 3 min at a rate of 5 °C/min for 25 min. Helium was used as carrier gas at a flow rate of 1 mL/min and detector/injector at a temperature of 260 °C. For identification of the peaks, the retention times of the fatty acid methyl ester (Supelco 37 Component FAME Mix) standards were compared with the observed peak retention times using the Openchrom software. Quantification was done by normalizing the peak area and expressed in percentage. The saturated and unsaturated percentages were compared by the Tukey test (p < 0.05) using the online tool available at https://statpages.info.

2.4. Oxidative stability

Oxidative stability was expressed with the induction period in hours and was determined using 3 g of sample in a Rancimat apparatus (873 - Biodiesel Rancimat) at 110 °C with airflow of 10 L/h and 50 mL of distilled water in vials containing electrodes, according to the AOCS Cd 12b-92 method (AOCS, 1993AOCS (1993). Official Methods and Recommended Practices of the American Oil Chemists Society. Champaing.). The oxidative stability indices were compared by the Tukey test (p < 0.05) using the online tool available at https://statpages.info.

2.5. Thermal behavior and solid fat content

Previously established conditions (Márquez et al., 2013Márquez AL, Pérez MP, Wagner JR. 2013. Solid fat content estimation by differential scanning calorimetry: Prior treatment and proposed correction. J. Am. Oil Chem. Soc. 90, 467-473. https://doi.org/10.1007/s11746-012-2190-z ) were used with modifications. Samples were heated at 60 °C for 15 min and left at 4 °C for 24 h. The thermograms of crystallization and melting were obtained using a Differential Scanning Calorimeter (DSC-60A - Shimadzu) coupled to the flow controller (FC-60A). The samples (5 mg) were placed in capped aluminum crucibles and cooled to -50 °C at 10 °C/min, maintaining this temperature for 10 min. The samples were then heated to 80 °C at 5 °C/min using a modulation amplitude of 1 °C every 60 s.

The solid fat content (SFC) was determined based on the area percentages of the integrated melting thermogram at 10, 15, 20, 25, 30, 35, and 40 C° (Cobo et al., 2017Cobo M, Deublein E, Haber A, Kwamen R, Nimbalkar M, Decker F. 2017. TD-NMR in Quality Control: Standard Applications, in Webb G (Ed.) Modern Magnetic Resonance. Springer, Cham, pp. 1819-1836. https://doi.org/10.1007/978-3-319-28388-3_12.; Gunstone et al., 1994Gunstone FD, Harwood JL, Padley FB. 1994. The Lipid Handbook. 2nd ed, CRC Press, London, UK.). Measurements for onset, peak, and final temperatures for crystallization and melting, as well as enthalpies, were obtained using the SciDAVis program. The deconvolution peak and integration were determined using the MagicPlot software.

2.6. Qualitative phase diagram

Experimental phase diagrams were obtained by storing the samples in capped 5-ml glass tubes and kept inverted at temperatures of 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, and 20 °C for 24 hours. The phase change described as solid, thick liquid, or liquid was visually evaluated. Non-flowing materials were named as solids, slowly-flowing materials were named as thick liquid, and immediately-flowing materials were named as liquids (Rocha et al., 2013 Rocha JCB, Lopes JD, Mascarenhas MCN, Arallano DB, Guerreiro LMR, Cunha RL. 2013. Thermal and rheological properties of organogels formed by sugarcane or candelilla wax in soybean oil. Food Res. Int. 50, 318-323. https://doi.org/10.1016/j.foodres.2012.10.043 ).

3.1. Fractionation process

The dry fractionation yielded 63.2% olein and 36.8% stearin, while the solvent fractionation yielded 71.8% olein and 28.2% stearin. In dry fractionation, crystals grew agglomerated on the vessel wall, unlike solvent fractionation, in which crystals grew dispersed in the medium. This higher amount of stearin in dry fractionation may be due to entrapment of the liquid fraction in the crystals. Slowly cooling at high temperatures typically results in larger crystals that can become trapped within the liquid fraction, while rapid cooling at low temperatures forms smaller crystals in larger quantities (Chaleepa et al., 2010Chaleepa K, Szepes A, Ulrich J. 2010. Effect of additives on isothermal crystallization kinetics and physical characteristics of coconut oil. Chem. Phys. Lipids 163, 390-396. https://doi.org/10.1016/j.chemphyslip.2010.03.005 ; Rodrigues-Ract et al., 2010Rodrigues-Ract JN, Cotting LN, Poltronieri TP, Silva RC, Gioielli LA. 2010. Crystallization behavior of structured lipids by chemical interesterification of milkfat and sunflower oil. Food Sci. Technol. 30, 258-267. https://doi.org/10.1590/S0101-20612010000100038 ; Silva et al., 2008Silva RC da, Escobedo JP, Gioielli LA. 2008. Crystallization behavior of structured lipids by chemical interesterification of milkfat and sunflower oil. Quim. Nova 31, 330-335. https://doi.org/10.1590/S0101-20612010000100038 ).

In solvent fractionation, there is a lower viscosity medium due to the presence of solvent and better heat transfer compared to the dry process, which results in a high nucleation rate and a rapid crystal growth which facilitate dispersion of the crystals. These effects were evaluated by Grall and Hartell (1992)Grall DS, Hartel RW. 1992. Kinetics of butterfat crystallization. J. Am. Oil Chem. Soc. 69 (8), 741-747. https://doi.org/10.1007/BF02635909 , who reported that the rate of nucleation and crystal growth increased when crystallization times decreased when using agitation. By increasing the agitation, the crystal mass increased, but with a greater effect on solid fractions at lower temperatures and less effect at elevated temperatures. Thus, with the use of solvents, crystals can grow more stable and faster than in the dry process, which could contribute to better phase separation (Kellens et al., 2007Kellens M, Gibon V, Hendrix M, De Greyt W. 2007. Palm oil fractionation. Eur. J. Lipid Sci. Technol. 109, 336-349. https://doi.org/10.1002/ejlt.200600309 ; Rossell, 1985Rossell JB. 1985. Fractionation of lauric oils. J. Am. Oil Chem. Soc. 62, 385-390. https://doi.org/10.1007/BF02541409 ).

3.2. Fatty acid composition and oxidative stability

Nine fatty acids were identified, including seven saturated (caproic; caprylic; capric; lauric; myristic; palmitic and stearic) one monounsaturated (oleic), and one polyunsaturated (linoleic) (Table 1). JKO had 73.4% saturated fatty acids and 26.6% unsaturated fatty acids, mainly lauric acid (37.4%), myristic acid (10.0%) and oleic acid (21.3%), which is in agreement with what is reported in the literature (Coimbra and Jorge, 2011Coimbra MC, Jorge N. 2011. Characterization of the Pulp and Kernel Oils from Syagrus oleracea, Syagrus romanzoffiana, and Acrocomia aculeata. J. Food Sci. 76, 1156-1161. https://doi.org/10.1111/j.1750-3841.2011.02358.x ; Moreira et al., 2013Moreira MAC, Payret Arrúa ME, Antunes AC, Fiuza TER, Costa BJ, Weirich Neto PH, Antunes SRM. 2013. Characterization of Syagrus romanzoffiana oil aiming at biodiesel production. Ind. Crops Prod. 48, 57-60. https://doi.org/10.1016/j.indcrop.2013.04.006 ).

In general, there was a slight numerical increase in saturated fatty acids in stearin (Es), with a consequent decrease in unsaturated ones, as opposed to olein (Os) fractions, although without statistically significant differences. Olein compositions were more similar to oil than stearins, which may indicate the presence of crystals in the oleins. However, it should be considered that the melting (or crystallizing) temperature of an oil or fat depends on the triacylglycerol (TAG) configuration, not just on the fatty acid composition. Previous work reported slight significant variations in the fatty acid compositions of coconut oil fractions (Marikkar et al., 2013Marikkar JMN, Saraf D, Dzulkifly MH. 2013. Effect of Fractional Crystallyzation on Composition and Thermal Behavior of Coconut Oil. Int. J. Food Prop. 16, 1284-1292. https://doi.org/10.1080/10942912.2011.585728 ) and palm oil fractions (Mo et al., 2016Mo S-Y, Teng K-T, Nesaretnam K, Lai O-M. 2016. Similar physical characteristics but distinguishable sn-2 palmitic acid content and reduced solid fat content of chemically interesterified palm olein compared with native palm olein by dry fractionation: A lab-scale study. Eur. J. Lipid Sci. Technol. 118, 1389-1398. https://doi.org/10.1002/ejlt.201500305 ), but these fractions still had very distinct characteristics for the composition of TAG and thermal properties.

JKO presented oxidative stability of 19.4 h, similar to that previously reported (Moreira et al., 2013Moreira MAC, Payret Arrúa ME, Antunes AC, Fiuza TER, Costa BJ, Weirich Neto PH, Antunes SRM. 2013. Characterization of Syagrus romanzoffiana oil aiming at biodiesel production. Ind. Crops Prod. 48, 57-60. https://doi.org/10.1016/j.indcrop.2013.04.006 ). Despite the fact that stearins showed numerically greater oxidative stability than oleins, no statistically significant difference was found. In general, the oxidative stability of oils is related to their composition and distribution of the fatty acids in TAG, as well as the presence of constituents with antioxidant properties. As previously reported (Ullah et al., 2016Ullah R, Nadeem M, Ayaz M, Muhammad I, Muhammad T. 2016. Fractionation of chia oil for enrichment of omega 3 and 6 fatty acids and oxidative stability of fractions. Food Sci. Biotechnol. 25, 41-47. https://doi.org/10.1007/s10068-016-0006-x ), approximately 70% of the oxidative stability of chia oil fractions depended on the fatty acid profile, but 30% depended on antioxidant substances such as chlorogenic and caffeic acids, quercetin, and phenolic glycoside. Thus, the similarities in the fatty acid composition of the fractions may explain the similarities in their oxidative stabilities.

3.3. Crystallization profile

The analysis of DSC peaks serves to determine the transition temperature of a fat or oil (Tan and Che Man, 2002Tan CP, Che Man YB. 2002. Differential scanning calorimetric analysis of palm oil, palm oil based products and coconut oil: effects of scanning rate variation. Food Chem. 76, 89-102. https://doi.org/10.1016/S0308-8146(01)00241-2 ). The stearins obtained started and ended crystallization at higher temperatures than oleins (Table 2, Figure 2), i.e., stearins should be more abundant in high melting TAGs than oleins. The olein obtained from dry fractionation started crystallization before oil, presenting higher T_onset. This may indicate that this olein dragged a reasonable amount of high melting (crystalline phase) TAGs into the liquid matrix during the fractionation process. Crystals are known to tend to form clusters due to attractive interactions between them, which can drag the olein into the crystals, decreasing the efficiency of separation (Kellens et al., 2007Kellens M, Gibon V, Hendrix M, De Greyt W. 2007. Palm oil fractionation. Eur. J. Lipid Sci. Technol. 109, 336-349. https://doi.org/10.1002/ejlt.200600309 ). The largest crystals that can be observed during crystallization are often composed of small crystals held together by weak interactions. Thus, this agglomeration may lead to lower separation efficiency due to a high drag of liquid in the agglomerates.

The crystallization curves (Figure 2) presented different profiles in different temperature ranges. After deconvolution, the curves revealed some sub-events associated with the crystallization of different TAG species. The thermograms were divided into three temperature ranges that were associated with the sub-events involving high, medium and low melting temperature TAGs. A first sub-event was associated with TAGs with crystallization temperature above 14 °C; a second sub-event was associated with TAGs with crystallization temperatures between 6 and 14 °C; and a third TAG-associated sub-event with crystallization temperatures below 6 °C. The enthalpies associated with these sub-events were calculated and are presented in Table 3.

When the temperature was reduced and the JKO started to crystallize (Figure 2), a small crystallization peak was observed near 15 °C with 0.8 J/g enthalpy (7.4% total enthalpy) associated with high melting TAGs, overlapping the larger enthalpy of crystallization peaks of 9.5 J/g (88% total enthalpy) associated with medium melting TAGs (Table 3). This peak overlap, as well as the presence of a broad peak, is due to the crystallization of more than one TAG species in the same temperature range due to the complexity of the structures of these compounds. The sharpness of a DSC peak indicates the cooperative nature of the transition from TAG composition. If the transition occurs in a narrow temperature range, it is highly cooperative. Therefore, the crystallization (or melting) of two or more TAG structures could take place simultaneously in a certain temperature range, resulting in a broad or overlapping DSC peak (Tan and Che Man, 2002Tan CP, Che Man YB. 2002. Differential scanning calorimetric analysis of palm oil, palm oil based products and coconut oil: effects of scanning rate variation. Food Chem. 76, 89-102. https://doi.org/10.1016/S0308-8146(01)00241-2 ).

The olein crystallization curve from dry fractionation was similar to the JKO curve, but with slight differences in the enthalpies in each temperature range, i.e., a decrease in the enthalpy fraction associated with high melting TAGs (0.2 J/g, 2.6% total enthalpy) and an increase associated to medium melting TAGs (7.3 J/g, 93.6% total enthalpy) (Table 3). The crystallization olein peaks of the solvent fractionation occurred at lower temperatures compared to JKO and stearins. A relevant increase in enthalpy associated with low melting TAGs (5.5 J/g, 70.5% total enthalpy) was also observed, as well as a decrease in enthalpy associated with medium melting TAGs (2.3 J/g, 29.5% total enthalpy) and absence of those with a high melting point (Table 3).

The stearins began to crystallize at higher temperatures compared to oil and oleins, with an increase in the enthalpy fraction associated with high melting TAGs. A slighter variation was observed for dry fraction stearin, with a relative area of 11.1% (0.5 J/g) associated with high melting TAGs, while in solvent fractionation this variation was more relevant, yielding an 18.5% enthalpy stearin associated with high melting TAGs, as well as a decrease in the enthalpy fraction associated with medium melting TAGs (Table 3).

The solvent fractionation revealed a greater ability to modify the crystallization characteristics of JKO when compared to dry fractionation. Moreover, even with the similarity in fatty acid composition (Table 1), differences in oil thermograms and fractions were observed. This reinforces the notion that the thermal characteristics of the fractions depend not only on the fatty acid composition, but also on the disposition of these fatty acids in the TAGs, as well observed for the solvent fractionation of coconut oil (Sonwai et al., 2017Sonwai S, Rungprasertphol P, Nantipipat N, Tungvongcharoan S, Laiyangkoon N. 2017. Characterization of Coconut Oil Fractions Obtained from Solvent Fractionation Using Acetone. J. Oleo Sci. 66, 951-961. https://doi.org/10.5650/jos.ess16224 ).

3.4. Melting profile

The JKO melting process started at 17.2 °C (T_onset) and continued at 36.0 °C (T_f), with only one broad endothermic peak at 30.9 °C (Table 2). The presence of only one endothermic transition can be characteristic of the constitution of the JKO, which should have TAGs with a less diversified structure when compared to other oils, such as palm oil, cocoa butter and some hydrogenated fats, which have fusion thermograms with a higher number of endothermic events (Tan and Che Man, 2002Tan CP, Che Man YB. 2002. Differential scanning calorimetric analysis of palm oil, palm oil based products and coconut oil: effects of scanning rate variation. Food Chem. 76, 89-102. https://doi.org/10.1016/S0308-8146(01)00241-2 ; Zaliha et al., 2004Zaliha O, Chong C, Cheow C, Norizzah AR, Kellens MJ. 2004. Crystallization properties of palm oil by dry fractionation. Food Chem. 86, 245-250. https://doi.org/10.1016/j.foodchem.2003.09.032 ; Hinrichsen, 2016Hinrichsen N. 2016. Commercially available alternatives to palm oil. Lipid Technol. 28 (3), 65-67. https://doi.org/10.1002/lite.201600018 ). The melting curve profile of the fractions was similar to that of oil, but with differences in initial, final and peak melt temperatures (Table 2). Similar results were found by Yantya et al. (2013Yantya NAM, Marikkar JMN, Shuhaimia M. 2013. Effect of fractional crystallization on the composition and thermal properties of engkabang (Shorea macrophylla) seed fat and cocoa butter. Grasas Aceites 64 (5), 546-553. https://doi.org/10.3989/gya.023213 ).

After deconvolutions, the melting curves (Figure 2) also presented distinct sub-events which were divided into three temperature ranges associated with TAGs with different melting temperatures. In a first temperature range, low melting TAGs melted at temperatures below 23 °C, predominantly in oleins; in a second range, medium melting TAGs melted at temperatures between approximately 23 and 33 °C; and in a temperature range above 33 °C, high melting TAGs were melted, predominantly in stearins. The enthalpies associated with these sub-events were calculated and are presented in Table 3.

All olein melting curves changed to lower temperatures compared to JKO (Figure 2); while stearin melting curves changed to higher temperatures. The main differences were related to the concentration of high melting TAGs in stearins and of low melting TAGs in oleins. This effect was most relevant in solvent fractionation, where enthalpy associated with the melting of high melting TAGs represented 29.2% of total enthalpy, compared to 20.3% in dry fraction stearin and 6.8% in JKO, similar to Macaúba kernel oil fractions (Magalhães et al., 2020aMagalhães KT, Tavares TS, Nunes CA. 2020a. The chemical, thermal and textural characterization of fractions from Macauba kernel oil. Food Res. Int. 130, 108925. https://doi.org/10.1016/j.foodres.2019.108925 ). In addition, fusion enthalpy associated with low melting TAGs represented 34.2% of total enthalpy in solvent fractionation olein, compared to 10.3% and 6.8% in dry fractionation olein and JKO, respectively.

3.5. Solid fat content

The SFC curves for the JKO and their fractions (Figure 3) revealed a substantial reduction in solid fat content with increasing temperature, especially in a narrow range between 25 and 35 °C, which should be the melting temperature range of the major TAGs of these products.

The SFC curves of the fractions showed that JKO’s stearins and oleins had different melting profiles, despite similar fatty acid profiles (Table 1). It is known that TAGs can pack tightly together, which influences their melting. TAGs with larger amounts of unsaturated fatty acids cannot pack as tightly together because of kinks in the fatty acid chain, leading to a lower melting point. However, it has been reported that the SFC profile does not only depend on the level of saturation or unsaturation of fatty acids in the TAG. In a study on the thermal behavior of vegetable oils with changes in fatty acids in the TAGs, the authors observed that the solid contents tended to increase when saturated fatty acids were more symmetrically distributed between the external positions of the TAGs. They concluded that the physical properties of the oils studied depended on the fatty acid distribution in the TAGs, as well as on the composition of the fatty acids themselves (Bootello et al., 2016Bootello MA, Garcés R, Martínez-Force E, Salas JJ. 2016. Effect of the distribution of saturated fatty acids in the melting and crystallization profiles of high-oleic high-stearic oils. Grasas Aceites 67, e149. https://doi.org/10.3989/gya.0441161 ). Marikkar et al. (2013)Marikkar JMN, Saraf D, Dzulkifly MH. 2013. Effect of Fractional Crystallyzation on Composition and Thermal Behavior of Coconut Oil. Int. J. Food Prop. 16, 1284-1292. https://doi.org/10.1080/10942912.2011.585728 and Mo et al. (2016)Mo S-Y, Teng K-T, Nesaretnam K, Lai O-M. 2016. Similar physical characteristics but distinguishable sn-2 palmitic acid content and reduced solid fat content of chemically interesterified palm olein compared with native palm olein by dry fractionation: A lab-scale study. Eur. J. Lipid Sci. Technol. 118, 1389-1398. https://doi.org/10.1002/ejlt.201500305 also reported distinct thermal properties for oleins and stearins from coconut oil and palm oil, even with slight variations in the fatty acid compositions of the fractions.

The oleins were completely melted at 35 °C, while stearins SFC were about 10% at this temperature. In most of the temperature ranges, oleins had lower SFC than JKO, while stearins had higher SFC, which was to be expected considering the enthalpy percentages associated with the lowest, medium and high melting TAGs previously presented (Table 3). This can be highlighted at the temperature of 25 °C (bars, Figure 3), especially for the solvent fractionation of stearin, which presented SFC above the other products throughout the temperature range, corroborating the enthalpy percentages associated with high and low melting TAGs (Table 3). The SFC at 25 °C of the solvent stearin was 17% higher than that of MKO, whereas the SFC at 25 °C of the solvent olein decreased by 7%.

The increase and decrease in SFC is expected in the stearin and olein fractions, respectively, compared with JKO after fractional crystallization. It was reported that the SFC of stearins from a dry fractionation of palm oil was higher than that of the oil and melted completely above 45 °C; while the SFC of oleins was lower than oil and entirely melted at 20 °C (Zaliha et al., 2004Zaliha O, Chong C, Cheow C, Norizzah AR, Kellens MJ. 2004. Crystallization properties of palm oil by dry fractionation. Food Chem. 86, 245-250. https://doi.org/10.1016/j.foodchem.2003.09.032 ). This is because the olein fraction generally contains a higher content in low melting TAGs than oil and stearin.

The SFC curves of the fractions obtained in this work reveal that stearins and oleins have different profiles, making it possible to obtain fractions that cover varying ranges of solid fat as a function of temperature. With the increase in the solid fat content, the lipids tend to have a firmer texture, which is relevant to some applications in food formulation, mainly with higher contents close to 25 °C, which contributes to the fraction having greater resistance to melting in ambient conditions when compared to unfractionated oil (Sonwai et al., 2017Sonwai S, Rungprasertphol P, Nantipipat N, Tungvongcharoan S, Laiyangkoon N. 2017. Characterization of Coconut Oil Fractions Obtained from Solvent Fractionation Using Acetone. J. Oleo Sci. 66, 951-961. https://doi.org/10.5650/jos.ess16224 ).

3.6. Qualitative phase diagram

An experimental phase diagram (Figure 4) was obtained for the JKO and its fractions, where three phases were identified according to the visual appearance of the samples: solid, thick liquid and liquid. The phase diagram revealed that the stearins remained visually solid until higher temperatures (16 and 18 °C) when compared to the oleins (13 to 15 °C). This behavior again corroborates the distribution of the lowest and highest melting TAGs in the fractions, as well as their solid fat content.

The olein obtained from dry fractionation had a visual appearance similar to that of the oil, which is in agreement with the crystallization and melting thermal profiles obtained by DSC (Figures 2 and 3), indicating that the crystals may have dragged in the liquid phase. On the other hand, solvent fractionation resulted in more distinct fractions compared to JKO, with smaller liquid phase dragged in the clusters at the end of the fractionation, allowing for more effective separation and resulting in fractions with more distinct compositions. As reported by Kellens et al. (2007)Kellens M, Gibon V, Hendrix M, De Greyt W. 2007. Palm oil fractionation. Eur. J. Lipid Sci. Technol. 109, 336-349. https://doi.org/10.1002/ejlt.200600309 , the main advantage of solvent fractionation is the high separation efficiency and higher purity of the finished products. The solvent fractionation olein and stearin differed most from the oil in the phase diagram, with the olein visually solidified at 14 °C and the stearin at 18 °C.