Recent advances in plant-based fat formulation as substitute for lard

3.1. Binary mixtures of EF:CO

EF is a tropical hard fat extracted from the seeds of Shorea macrophyilla. This has been previously known for its characteristic properties being similar to cocoa butter. Canola oil, a major vegetable oil, has been previously identified as a suitable candidate for novel fat formulations (Campbell et al., 2002Campbell SD, Goff H.D, Rousseau D. 2002. Comparison of crystallization properties of a palm stearin/canola oil blend and lard in bulk and emulsified form. Food Res. Intel. 35, 935-944. https://doi.org/10.1016/S0963-9969(02)00156-4. ; Marangoni and Rousseau, 1998Marangoni AG, Rousseau D. 1998. Chemical and enzymatic modification of butterfat and butterfat-canola oil blends. Food Res. Int. 31, 595-599. https://doi.org/10.1016/S0963-9969(99)00033-2 ). Several efforts have been made in the past to expand the use of EF as it is considered an under-utilized plant lipid (Yanty et al., 2013bYanty NAM, Marikkar JMN, Shuhaimi M. 2013b. Effect of fractional crystallization on composition and thermal properties of engkabang (Shorea macrophylla) seed fat and cocoa butter. Grasas Aceites 64, 546-553. https://doi.org/10.3989/gya.023213. ; Nesaretnam and Mohd Ali, 1992Nesaretnam K, Mohd Ali AR. 1992. Engkabang (Illipe)-an excellent component for cocoa butter equivalent fat. J. Sci. Food Agric. 60, 15-20. https://doi.org/10.1002/jsfa.2740600104.). In a previous attempt, Gani et al. (2011)Gani SSA, Basri M, Rahman MBA, Kassim A, Rahman RNZRA, Sallih AB, Ismail Z. 2011. Enkabang fat esters for cosmeceutical formulation. J. Surf. Deter. 14, 227-233. https://doi.org/10.1007/s11743-010-1233-4 demonstrated the usefulness of EF as a raw material for preparing moisturizers used in cosmetic products as well as hair conditioners. In a separate study, EF was fractionated into a hard stearin and a liquid olein to explore its suitability in other applications (Yanty et al., 2013bYanty NAM, Marikkar JMN, Shuhaimi M. 2013b. Effect of fractional crystallization on composition and thermal properties of engkabang (Shorea macrophylla) seed fat and cocoa butter. Grasas Aceites 64, 546-553. https://doi.org/10.3989/gya.023213. ).

In the search for lard substitutes, engkabang fat was mixed with canola oil in proportions ranging from 25 to 40% (w/w) (Table 1). The formulated blends appeared pale yellow in color, but displayed varying degrees of unsaturation (IV from 104.7-92.5). In fact, the IV of the formulated blends of EF:CO resulted higher than that of lard (IV, 73.6), affecting the oxidative stability. Among the physical characteristics, SMP is a parameter of considerable importance as it is indicative of the point of transition from solid to liquid. According to Nur Illyin et al. (2013)Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058., the SMP of the formulated blends of EF:CO fell within the range of 24.8-31.9; while the SMP of lard was 29.5. As shown in Table 2, the SMP of the EF:CO=30:70 (w/w) blend was found to tally with that of lard, despite some deviation in the degree of unsaturation.

Blending canola oil with engkabang fat brought significant deviations in chemical compositions, as shown in Tables 2 and 4. According to several previous reports, trioleoyl glycerol (OOO), dioleoyl-3-linoleoyl glycerol (OOL), and dioleoyl-3-linolinoyl glycerol (OOLn), were the major TAG molecular species of canola oil (Marikkar and Sohel Rana, 2014Marikkar JMN, Sohel R. 2014. Use of differential scanning calorimetry to detect canola oil (Brassica napus L.) adulterated with lard stearin. J. Oleo Sci. 63, 867-873. doi: 10.5650/jos.ess14064.; Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.; Marikkar et al., 2005Marikkar JMN, Ghazali HM, Che Man YB, Peiris TSG, Lai OM. 2005. Distinguishing lard from other animal fats in admixtures of some vegetable oils using liquid chromatographic data coupled with multivariate data analysis. Food Chem. 91, 5-14. 10.1016/j.foodchem.2004.01.080 ). The addition of engkabang fat into canola oil caused gradual but significant increases in the proportions of palmitoyl-oleoyl-stearoyl glycerol (POS) and disearoyl-2-oleoyl glycerol (SOS) with concurrent decreases in the proportions of OOO, OOL and OOLn (Table 4). For instance, the proportions of triunsaturated TAG molecular species decreased remarkably as the proportion of engkabang fat in canola oil was increased from 25 to 40% (w/w) (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). When compared to canola oil, the binary blends of EF:CO were found to possess a high amount of saturated fatty acid with reduced amounts of unsaturated fatty acids. From the nutritional point of view, this would have a negative effect on the availability of essential fatty acids. This was mainly due to the increases in the proportions of palmitic and stearic acids with concurrent reductions in the amounts of oleic and linoleic acids (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). Similar situations were reported previously for blending canola oil with butter oil (Marangoni and Rousseau, 1998Marangoni AG, Rousseau D. 1998. Chemical and enzymatic modification of butterfat and butterfat-canola oil blends. Food Res. Int. 31, 595-599. https://doi.org/10.1016/S0963-9969(99)00033-2 ) and canola oil with lard stearin (Marikkar and Sohel Rana, 2014Marikkar JMN, Sohel R. 2014. Use of differential scanning calorimetry to detect canola oil (Brassica napus L.) adulterated with lard stearin. J. Oleo Sci. 63, 867-873. doi: 10.5650/jos.ess14064.). In addition to this, blending canola oil with other fats was found to steadily change the unsaturated to saturated fatty acid ratio.

The impact on DSC thermal profiles and solidification characteristics of canola oil caused by changing chemical compositions is another important aspect. As reported previously, the DSC curve of canola oil is characterized by the occurrence of two overlapping endotherms: a large-higher temperature transition at -17.86 ^oC and a small -lower temperature transition at -28.50 ^oC (Marikkar and Rana, 2014Marikkar JMN, Sohel R. 2014. Use of differential scanning calorimetry to detect canola oil (Brassica napus L.) adulterated with lard stearin. J. Oleo Sci. 63, 867-873. doi: 10.5650/jos.ess14064.). Since the end-set of the melting of canola oil was at -6.79 ^oC, thermal transition beyond this point was hardly seen. All formulated fat blends, however, displayed a high-melting thermal transition in the temperature region above 10 °C and the enthalpy and peak temperature of this high-melting peak were found to increase proportionately with the increasing proportion of EF in the fat blends. This would be most probably due to the increases noticed earlier in the proportion of saturated TAG molecules in the admixtures. Out of the four blends, the peak maxima of EF:CO (35:65) showed the closest value to that of lard, which had its maximum peak at 29.3 °C (Table 3). The SFC profiles usually describe the solidification behavior of fatty substances within a specified temperature range. For fat blends of the EF:CO series, the SFC values tended to increase throughout the temperature range with the increasing proportion of EF. Further, the increase in SFC values were found to correlate well with the increasing proportion of disaturated TAG molecules in the fat blends (Table 4). According to Rao et al. (2001)Rao R, Sankar KU, Sambaiah K, Lokesh BR. 2001. Differential scanning calorimetric studies on structured lipids from coconut oil triglycerides containing stearic acid. Eur. Food Res. Technol. 212, 334-343. https://doi.org/10.1007/s002170000254. , the SFC of a fat at 25 °C should be within the range of 15-35% to achieve the desired spreadability and texture. Further, SFC at 20 °C needs to be adjusted to more than 10% in order to avoid any oil separation in the blend. Although the SFC values of EF:CO (30:70) and lard were not similar at 0 °C, they were found to be quite similar within the range of 30-40 °C and become tallied at 25 °C. Apart from this, the results from X-ray diffractograms also showed both EF:CO (30:70) and lard possessed both β and β’ polymorphs, which were ideal for a plant-based shortening (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). This was because both lard and EF:CO (30:70) had roughly similar amounts of SPO (Table 4). According to Timms (1984)Timms RE. 1984. Phase behaviour of fats and their mixtures. Prog. Lipid Res. 23 (1), 1-38. https://doi.org/10.1016/0163-7827(84)90004-3., the presence of SPO could significantly contribute to the β’ crystal polymorph’s development in lard.

3.2 Binary mixtures of PS:GO

Guava seeds (Psidium guajava L.), discarded after juice production can be a useful material for oil extraction. According to a previous study reported from India, the Guava seed is rich in essential fatty acids such as linoleic acid (76.4%) (Prasad and Azeemoddin, 1994Prasad NBL, Azeemoddin G. 1994. Characteristics and composition of guava (Psidium guajava L.) seed and oil. J. Am. Oil Chem. Soc. 71, 457-458. https://doi.org/10.1007/BF02540531 ). Although guava oil is known for this important nutritional attribute, its commercial exploitation has not been advanced to product formulations. In search for a substitute for lard, mixing guava oil with palm stearin has been performed in varying proportions ranging from 45 to 60% (w/w) (Table 1). The SMP of all fat blends were found to be higher (SMP = 43.6 to 50.3) than that of lard (29.25), but lower than that of palm stearin (52.7) (Table 2). With regard to the degree of unsaturation, all fat blends displayed significantly lower IV values (p < 0.05) (IV= 42.3 to 50.6) than that of lard (73.76) (Table 2). Hence, none of the binary mixtures of PS:GO was found to have either SMP or IV similar to those of lard. As the added portion of palm stearin into guava oil increased from 45 to 60%, a clear change in the unsaturated to saturated fatty acid ratio was noticed. For instance, there was a gradual decrease in linoleic (from 46.16 to 33.91%) acid content with a concurrent increase in palmitic acid content (from 31.35 to 40.72%) (Table 2). This pattern of change was very similar to those previously observed for the EF:CO blend formulation (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). Simultaneously, decreases in the proportions of TAG molecules such as LLL, OLL, PLL and POL occurred with concurrent increases in the amounts of PPO, tripalmtoyl glycerol (PPP), POS and dipalmitoyl-3-stearoyl glycerol (PPS) (Table 4). The changes were happened in such a way that the proportions of triunsaturated TAG molecular species declined from 36.79 to 27.79 % (Table 4). In fact, increases in the proportions of disaturated and trisaturated TAG molecular species led to the increase in palmitic acid content as noticed before in the overall fatty acid distribution (Table 2). As noted previously in EF:CO blend formulations (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.), the proportions of POS and SOS were increased while those of OOO, OOL and OOLn were decreased. In this aspect, both of these formulation efforts yielded somewhat similar patterns of changes (Table 4).

Serval previous studies indicated that the SFC values of fats are mainly influenced by their respective fatty acid and TAG compositions. According to Graef et al. (2012)Graef VD, Vereecken J, Smith KW, Bhaggan K, Dewettinck K. 2012. Effect of TAG composition on the solid fat content profile, microstructure, and hardness of model fat blends with identical saturated fatty acid content. Eur. J. Lipid Sci. Technol. 114, 592-601. https://doi.org/10.1002/ejlt.201100215 , the proportional distribution of disaturated and trisaturated TAG molecular species would have greater influence on the SFC values of fats as more saturated TAG molecules would create a stronger crystal network. Raihana et al. (2017a)Raihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. observed that the SFC value at 5 ^oC of the formulated blends of PGO-1 (PS:GO=60:40), PGO-2 (PS:GO = 55:45), PGO-3 (PS:GO = 50:50) and PGO-4 (PS:GO = 45:55) were 47.19, 47.19, 38.16 and 34.38%, respectively. The increasing values of SFC presented herein were found to correlate well with the increasing proportion of disaturated TAG molecules in the fat blends (Table 4). This observation is in conformity with the previous findings reported on EF:CO fat blends (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.; Nur Illyin et al., 2014Nur Illiyin MR, Marikkar JMN, Lok MK, Shuhaimi M, Mahiran B, Miskandar MS. 2014. Interesterification of engkabang (Shorea macrophylla) fat - canola oil blend with lipase from candida antarctica to simulate the properties of lard. J. Oleo Sci. 63, 39-46. https://doi.org/10.5650/jos.ess13115. ). Of these blends, PGO-2 was found to display a SFC value similar to lard at temperatures 15, 20, and 25 ^oC. Particularly, at 30 ^oC, the SFC value for lard (12.43%) was between the SFC values for PGO-3 (13.70) and PGO-4 (11.81%). Based on some calculations performed with regard to least difference in SFC values, out of the four blends, PGO-1 and PGO-2 were found to display the smallest differences compared to lard in terms of SFC values.

Several previous studies pointed out that even a little change in fatty acid and TAG compositions would affect the DSC thermal profile of pure oils (Marikkar, 2015Marikkar JMN. 2015. DSC as a valuable tool for the evaluation of oils and fats adulterations. In: Differential Scanning Calorimetry: Applications in Fat and Oil Technology, Emma Chiavaro (Editor). CRC Press, Taylor & Francis Group, Florida, USA. pp. 159-178. ISBN: 9781466591523). Aziz et al. (2011)Aziz A, Mohmud Y, Roselina K, Boo HC, Nyuk LC, Che Man YB. 2011. Rheological, chemical and DSC thermal characteristics of different types of palm oil/ palm stearin-based shortenings. Int. Food Res. J. 18, 189-200. indicated that the mixing of palm stearin with palm oil would affect the DSC cooling profile of the formulated shortening because the high melting TAGs are introduced by palm stearin. Separately, Nur Illyin et al. (2013)Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058. and Nur Illyin et al. (2014)Nur Illiyin MR, Marikkar JMN, Lok MK, Shuhaimi M, Mahiran B, Miskandar MS. 2014. Interesterification of engkabang (Shorea macrophylla) fat - canola oil blend with lipase from candida antarctica to simulate the properties of lard. J. Oleo Sci. 63, 39-46. https://doi.org/10.5650/jos.ess13115. noticed that a little mixing of enkabang fat into canola oil caused fatty acid and TAG compositional changes which affected the thermal profile of canola oil. Likewise, the mixing of palm stearin into guava oil affected the DSC profile in such a way that all fat blends displayed high melting thermal transitions above 10 ^oC. The maximum peak of the high-melting thermal transitions corresponding to PGO-1, PGO-2, PGO-3 and PGO-4 shifted from 46.26 to 50.32 ^oC. According to this observation, out of the four formulated-blends, the peak maxima of PGO-4 (PS:GO=45:55) was at 44.20 ^oC and that of PGO-2 (PS:GO=55:45) was at 50.0 ^oC (Table 3). The results of the SMP and SFC, as well as fatty acid composition indicated that PGO-2 would become a better fat blend to substitute lard. Apart from this, the result of X-ray diffractograms also showed both PGO-2 and lard possessed with β and β’ polymorphs, which were ideal for a plant-based shortening. This was quite similar to what was previously observed in the case of EF:CO (30:70), which exhibited both β and β’ polymorphs (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). The occurrence of a large amount of disaturated TAG molecules such as PPO in PGO-2 (Table 2) contributed to the existence of β and β’ polymorphs in PGO-2.

3.3. Binary mixtures of MF:PS

Mee fat extracted from the seeds of Madhuca longifolia is said to be an analogue of Indian Mahua butter (Ramadhan et al., 2016Ramadan, M.F., Mohadaly, A.A.A., Assiri, A.M.A., Tadros, M., and Neimeyer, B. 2016. Functional characteristics, nutritional values and industrial applications of Madhca longifolia seeds. J. Food Sci. Technol. 532149-2157. https://doi.org/10.1007/s13197-015-2095-6.; Marikkar et al., 2010Marikkar JMN, Ghazali HM, Long K. 2010. Composition and thermal characteristics of Madhuca longifolia seed fat and its solid and liquid fractions. J. Oleo Sci. 59, 7-14. https://doi.org/10.5650/jos.59.7 ). Although mee fat is preferred as a fat ingredient in the formulations of balms and ointments used in folk medicine (Ramadhan et al., 2016Ramadan, M.F., Mohadaly, A.A.A., Assiri, A.M.A., Tadros, M., and Neimeyer, B. 2016. Functional characteristics, nutritional values and industrial applications of Madhca longifolia seeds. J. Food Sci. Technol. 532149-2157. https://doi.org/10.1007/s13197-015-2095-6.), its edible uses are yet to be realized (Marikkar et al., 2010Marikkar JMN, Ghazali HM, Long K. 2010. Composition and thermal characteristics of Madhuca longifolia seed fat and its solid and liquid fractions. J. Oleo Sci. 59, 7-14. https://doi.org/10.5650/jos.59.7 ). In an effort to find a substitute for lard, Marikkar and Yanty (2012)Marikkar JMN, Yanty NAM. 2012. Seed fat from Madhuca longifolia as raw material for halal alternative fat. Borneo Sci. 31, 84-94. compared the solid and liquid fractions of mee fat with those of lard. Although both of these exhibited compatibility in terms of SFC values at some temperatures, they displayed some disparity at certain other temperatures. Despite this, these findings provided the impetus to formulate three fat blends by mixing mee fat with palm stearin in proportions ranging from 0.5 to 2 % (w/w) (Yanty et al., 2014aYanty NAM, Marikkar, JMN, Shuhaimi, M, Miskandar MS. 2014a. Composition and thermal analysis of binary mixture of Mee fat and palm stearin. J. Oleo Sci. 63, 325-332. https://doi.org/10.5650/jos.ess13193. ). Although the IV of formulated-blends of MF:PS were significantly lower (54.27 to 57.81) (p < 0.05) than that of lard (73.76), their SMP were significantly higher (37.25 to 40.25) (p < 0.05) than that of lard (29.5 °C) (Table 2). In conformity to this, Raihana et al. (2017a)Raihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. also noticed that the fat blends of PS:GO displayed significantly lower IV values (p < 0.05) (IV= 42.3 to 50.6) than that of lard (73.76). However, Nur Illyin et al. (2013)Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058. and Nur Illyin et al. (2014)Nur Illiyin MR, Marikkar JMN, Lok MK, Shuhaimi M, Mahiran B, Miskandar MS. 2014. Interesterification of engkabang (Shorea macrophylla) fat - canola oil blend with lipase from candida antarctica to simulate the properties of lard. J. Oleo Sci. 63, 39-46. https://doi.org/10.5650/jos.ess13115. reported that all formulated blends of EF:CO were found to display (104.7-92.5) higher IV values than lard (73.76). This could be due to the high contribution of monounsaturated fatty acids such as oleic in the formulated blends. When considering the thermal properties, the SMP of all fat blends of MF:PS were found to be higher (37.25 to 40.25) than that of lard (29.25) in conformity with Raihana et al. (2017a)Raihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. , who also subsequently noticed that all fat blends of PS:GO had significantly higher SMP values (p < 0.05) (SMP= 43.60 to 50.3) than that of lard (29.5). In contrast to these two studies, some formulated blends of EF:CO were found to display roughly similar SMP (24.8 - 31.1.5) to that of lard (29.5) (Table 1). As the proportion of palm stearin mixed into mee fat increased from 0.5 to 2%, the proportions of POL, dioleoyl-3-palmitoyl glycerol (POO), dioleoyl-3-steroyl glycerol (SOO) and SOS were decreased in such a way that the proportions of di-unsaturated TAG molecular species declined from 43.13 to 34.65 % (Table 4). In the meantime, the proportions of disaturated TAG molecular species increased, leading to the increase in palmitic acid in the overall fatty acid distribution.

As fractional crystallization would yield novel solid and liquid fractions of mee fat and lard, they were compared in terms of their physicochemical parameters and thermal properties (Marikkar and Yanty, 2012Marikkar JMN, Yanty NAM. 2012. Seed fat from Madhuca longifolia as raw material for halal alternative fat. Borneo Sci. 31, 84-94.). According to this study, the mee fat and lard used had displayed compatibility in terms of SFC profiles, especially, SFC values at 5, 20 and 25 °C. However, their SFC values differed within the temperature ranges from 10 to 15 °C as well as 30 to 35 °C, which could be due to their differences in TAG distributional pattern as discussed previously (Yanty et al., 2012Yanty NAM, Marikkar JMN, Miskandar MS. 2012. Comparing the thermo-physical properties of lard and selected plant fats. Grasas Aceites. 63, 328-334. https://doi.org/10.3989/gya.023712. ). Owing to this fact, Yanty et al. (2014a)Yanty NAM, Marikkar, JMN, Shuhaimi, M, Miskandar MS. 2014a. Composition and thermal analysis of binary mixture of Mee fat and palm stearin. J. Oleo Sci. 63, 325-332. https://doi.org/10.5650/jos.ess13193. hypothesized that mixing a small amount of palm stearin into mee fat would adjust the SFC values of mee fat at those temperatures. With increasing proportion of palm stearin from 0.5 to 2%, the SFC values of the three formulated blends increased gradually. The observed increases in SFC values were well-correlated with the increasing proportions of di- and tri-saturated TAG molecules (PPO, PPSt, and StPO) in the binary blends (Table 4). Out of the three formulated-blends, MF:PS (99:1) showed better compatibility with lard at most temperatures within the specified range. The pattern of changes in the SFC profiles of these fat blends was more or less comparable to those of the binary blends of EF:CO (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.; Nur Illyin et al., 2014Nur Illiyin MR, Marikkar JMN, Lok MK, Shuhaimi M, Mahiran B, Miskandar MS. 2014. Interesterification of engkabang (Shorea macrophylla) fat - canola oil blend with lipase from candida antarctica to simulate the properties of lard. J. Oleo Sci. 63, 39-46. https://doi.org/10.5650/jos.ess13115. ) and PS:GO (Raihana et al., 2017aRaihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. ) as reported previously. However, changes in the fatty acid composition during blend formulation were only marginal (Table 2). For instance, with increasing proportions of palm stearin in the blends, only slight changes occurred in the contents of total SFA (from 44.25 to 45.77%) and total USFA (from 55.75 to 54.23%). This led to slight increases in the amount of palmitic acid with concurrent decreases in the amounts of stearic, oleic and linoleic acids.

The DSC thermal profiles of pure lipids are generally influenced by chemical compositional changes. Hence, the changes in both fatty acid and TAG compositions would affect the DSC thermal profiles of the formulated binary blends in such a way that the maximum peak of all melting transitions were shifted to higher temperature regions, resulting in higher end-sets of melting (T_endset) (Yanty et al., 2014aYanty NAM, Marikkar, JMN, Shuhaimi, M, Miskandar MS. 2014a. Composition and thermal analysis of binary mixture of Mee fat and palm stearin. J. Oleo Sci. 63, 325-332. https://doi.org/10.5650/jos.ess13193. ). The T_endset value is often regarded as an alternative for slip melting point. Nur Illyin et al. (2013)Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058. previously noticed a high-melting thermal transition in the temperature region above 10 °C in all binary blends of EF:CO. Since there was hardly any thermal transition in the original DSC curve for canola oil above 10 °C, it was considered as a remarkable feature in the formulated-blends. Owing to the increasing amounts of more saturated TAG molecules, the maximum peak of the high-melting thermal transitions present in the blends were increased gradually (Table 3). This is in conformity with the previous findings reported by Raihana et al. (2017a)Raihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. , who observed that the maximum peak of the high-melting thermal transition of PS:PGO blends also shifted from 46.26 to 50.32 ^oC. In a separate study, Norizzah et al. (2004)Norizzah AR, Chong CL, Cheow CS, Zaliha O. 2004. Effects of chemical interesterification on physicochemical properties of palm stearin and palm kernel olein blends. Food Chem. 86, 229-235. https://doi.org/10.1016/j.foodchem.2003.09.030 made a similar observation in the case of palm olein blended with palm stearin, giving rise to the appearance of a broad peak in the high-melting region of the thermal curve of palm stearin. Yanty et al. (2017b)Yanty NAM, Marikkar JMN, Miskandar, MS, Van Bockstaele F, Dewettinck K, Nosontoro BP. 2017b. Compatibility of selected plant-based shortenings as lard substitute: microstructure, polymorphic forms and textural properties. Grasas Aceites. 68 (1), e181. http://dx.doi.org/10.3989/gya.0813162. conducted a X-crystallographic study and reported that both LD and MF:PS (99:1) blends displayed both β’ and β-form polymorphs, of which the β’ form was dominant. These researchers believed that the occurrence of the disaturated TAG (USS) molecular species such as PPL, PPO, StOP, and StOSt in both of these could be the probable reason for this similarity.

3.4. Ternary mixtures of Avo:PS:CB

In recent years, the oil of avocado pulp has received much attention due to its uses in food, cosmetics and health care products (Qin and Zhong, 2016Qin X, Zhong J. 2016. A review of extraction techniques for avocado oil. J Oleo Sci. 65, 881-888. https://doi.org/10.5650/jos.ess16063 ). Although avocado oils from some varieties exist in liquid form, avocado lipids of some Malaysian cultivars were found to exist in a semi-solid state at room temperature. This difference in their physical state could be due to the compositional differences based on TAG and fatty acid profiles (Yanty et al., 2011bYanty NAM, Marikkar JMN, Long K. 2011b. Effect of varietal differences on composition and thermal behavior of avocado oil. J. Am. Oil Chem. Soc. 88, 1997-2003. https://doi.org/10.1007/s11746-011-1877-x.). Qin and Zhong (2016)Qin X, Zhong J. 2016. A review of extraction techniques for avocado oil. J Oleo Sci. 65, 881-888. https://doi.org/10.5650/jos.ess16063 made some further comments regarding this matter in a survey presented on the efficiency of different extraction methods of avocado fat. A plant-based substitute for lard was proposed by Yanty et al. (2017a)Yanty NAM, Marikkar JMN, Shuhaimi M, Miskandar MS. 2017a. Composition and thermal analysis of ternary mixtures of avocado fat:palm sterin:cocoa butter (Avo:PS:CB). Int. J. Food Prop. 20, 465-474. https://doi.org/10.1080/10942912.2016.1166130 by making use of avocado fat as the major component and palm stearin / cocoa butter as minor components. According to a previous study by Yanty et al. (2013b)Yanty NAM, Marikkar JMN, Che Man YB. 2013a. Effect of fractional crystallization on composition and thermal characteristics of avocado (Persea americana) butter. J. Therm. Anal. Calorim. 111, 2203. https://doi.org/10.1007/s10973-011-2055-y , Malaysian cocoa butter was found to possess POP, POS, and SOS as its major TAG molecular species. Likewise, another previously mentioned that when compared to lard, Avo of different Malaysian cultivars exhibited lower SFC values throughout the temperature range (Yanty et al., 2014aYanty NAM, Marikkar, JMN, Shuhaimi, M, Miskandar MS. 2014a. Composition and thermal analysis of binary mixture of Mee fat and palm stearin. J. Oleo Sci. 63, 325-332. https://doi.org/10.5650/jos.ess13193. ). As such, the addition of hard fats such as cocoa butter and palm stearin is inevitable to complement or adjust the SFC profiles.

When studying the properties of novel fat blend formulations, the degree of unsaturation is an index of considerable importance as it is an indicator of the oxidative stability of food lipids. With regard to degree of unsaturation, Avo:PS:CB blends displayed slightly lower IV (65.47 to 70.27) (p < 0.05) than the lard (73.76) (Table 2). This characteristic feature was somewhat similar to those reported earlier by Yanty et al. (2014a)Yanty NAM, Marikkar, JMN, Shuhaimi, M, Miskandar MS. 2014a. Composition and thermal analysis of binary mixture of Mee fat and palm stearin. J. Oleo Sci. 63, 325-332. https://doi.org/10.5650/jos.ess13193. and Raihana et al. (2017a)Raihana ARN, Marikkar JMN, Jaswir I, Nurrulhidaya AF, Miskandar MS. 2017a. Effect of pink guava oil-palm stearin blends and lard on dough properties and cookies quality. Int. Food Res. J. 24, S355-S362. in binary blends of MF:PS and PS:GO, respectively. The SMP of the ternary blends studied by researchers were also found to be higher (38.25 to 40.5 °C) than that of lard (29.25) (Table 2). Hence, the SMP and IV of these ternary blends and lard were not found to be similar. Additions of palm stearin and cocoa butter into avocado fat caused slight increases in the amounts of palmitic (from 30.37 to 33.08%), stearic (from 1.30 to 4.57%) and oleic (from 43.64 to 44.04%) acids with concurrent decreases in the amounts of linoleic acids (from 17.45 to 14.28%). The increases in palmitic and stearic acids of these ternary blends was probably due to the presence of higher amounts of palmitic and stearic acids in palm stearin and cocoa butter, respectively.

DSC thermograms are generally helpful in monitoring the thermal properties of natural fats and their blends (Marikkar, 2015Marikkar JMN. 2015. DSC as a valuable tool for the evaluation of oils and fats adulterations. In: Differential Scanning Calorimetry: Applications in Fat and Oil Technology, Emma Chiavaro (Editor). CRC Press, Taylor & Francis Group, Florida, USA. pp. 159-178. ISBN: 9781466591523). The addition of palm stearin and cocoa butter into avocado fat would affect the melting peaks of avocado fat throughout the temperature region. Compositional changes such as increases in saturated fatty acids (Table 2) and disaturated/trisaturated TAG contents in the formulated blends are attributed to this (Table 4). On a comparative basis, the melting transitions of lard and the formulated ternary blends were found to display more differences than similarities. For instance, the end-sets of melting (T_endset) transition of all three Avo:PS:CB blends were higher (at around 47 °C) than that of lard (37.5 °C). However, there was a similarity noticed between Avo:PS:CB (84:7:9) blends and lard with regard to the maximum peak of thermal transitions at around -3.59 °C. Although it was a positive attribute, this alone cannot be sufficient to make a decision regarding which blend would be best in compatibility. A more relevant parameter to measure the compatibility of the ternary blends with lard would be the SFC profile as monitored by NMR. The sample containing the NMR tube was melted at 70 °C for 15 min, followed by chilling at 0 °C for 60 min, and then held at each measuring temperature for 30 min prior to measurement.

Past studies found that the SFC values of avocado fat of Malaysian cultivars were always lower than those of lard throughout the specified temperature range (Yanty et al., 2012Yanty NAM, Marikkar JMN, Miskandar MS. 2012. Comparing the thermo-physical properties of lard and selected plant fats. Grasas Aceites. 63, 328-334. https://doi.org/10.3989/gya.023712. ). The incorporation of a hard fat like palm stearin or cocoa butter could help enhance the SFC values of avocado fat to a level compatible with those of lard. By the addition of palm stearin (7%) and cocoa butter (5 to 9%) in small quantities into avocado fat, some of the TAG molecular groups were increased (e.g. PPO, PPP, SOO, SPO and PPS) while others tended to decrease (e.g. LLn, LLL, OLL, PLL, OOL, POL, PPL, OOO and POO) (Table 4). As PPP, PPO and SPO were originally present in high amounts in PS, their amounts would certainly go up. The excessive presence of these TAG molecules in PS had been previously confirmed by other researchers (Podchong et al., 2018Podchong P, Tan CP, Sonwai S and Rousseau D. 2018. Composition and crystallization behavior of solvent-fractionated palm stearin. Int. J. Food Prop. 21, 496-509. https://doi.org/10.1080/10942912.2018.1425701 ; Miskandar and Nor, 2010Miskandar MS, Nor A. 2010. Palm stearin as low trans hard stock for margarine. Sains Malay. 39, 82-827. http://journalarticle.ukm.my/7412/1/01_Md_Yeaminhossain.pdf ). In addition, Segall et al. (2005)Segall SD, Artz WE, Raslan DS, Ferraz VP, Takahashi JA. 2005. Analysis of triacylglycerol isomers in Malaysian cocoa butter using HPLC-mass spectroscopy. Food Res. Int. 38, 167-174. https://doi.org/10.1016/j.foodres.2004.09.008. stated that the excessive amounts of SPO, PPO and SOS TAG molecules present in cocoa butter would create a strong fat crystal network. With the gradual increases in the proportion of cocoa butter and palm stearin, the SFC values of the ternary fat mixtures tended to increase throughout the temperature region (Yanty et al., 2017aYanty NAM, Marikkar JMN, Shuhaimi M, Miskandar MS. 2017a. Composition and thermal analysis of ternary mixtures of avocado fat:palm sterin:cocoa butter (Avo:PS:CB). Int. J. Food Prop. 20, 465-474. https://doi.org/10.1080/10942912.2016.1166130 ). In fact, there was a good correlation (r=+0.97; p < 0.05) between the increasing SFC values of the fat blends and the increasing proportion of disaturated/trisaturated TAG molecules (Yanty et al., 2017aYanty NAM, Marikkar JMN, Shuhaimi M, Miskandar MS. 2017a. Composition and thermal analysis of ternary mixtures of avocado fat:palm sterin:cocoa butter (Avo:PS:CB). Int. J. Food Prop. 20, 465-474. https://doi.org/10.1080/10942912.2016.1166130 ). Of the three fat blends formulated, Avo:PS:CB (84:7:9) showed better compatibility to lard at most of the temperatures in the range. In addition to this, both lard and Avo:PS:CB (84:7:9) were found to possess almost similar proportions of PPO and SPO (Table 4). The occurrence of these two TAG molecular species in significant amounts caused both LD and Avo:PS:CB (84:7:9 blends to display both β’ and β-form polymorphs, of which the β’ form was dominant.

3.5. Quaternary mixtures of PO:PS:SBO:CB

Palm oil is a semi-solid that has been used worldwide for its multifaceted food and nonfood applications. The ever-increasing popularity of palm oil across the globe has resulted in its inclusion in various products. As palm oil can undergo fractionation, it would be an ideal substance for novel product formulation for specific needs (Nor Aini and Miskandar, 2007Nor Aini I, Miskandar MS. 2007. Utilization of palm oil and palm products in shortenings and margarines. Eur. J. Lipid Sci. Technol. 109, 422-432. https://doi.org/10.1002/EJLT.200600232 ). For instance, palm oil has been used for different types of shortenings, margarines, vanaspati, deep-frying fat and other specialty fats. Nevertheless, it cannot become a direct substitute for lard due to the occurrence of PPO and POO in high amounts, which can cause an unusually high solidification profile (Siew, 2002Siew WL. 2002. Palm oil. In Vegetable Oils in Food Technology: Composition, Properties and Uses, ed. F.D. Gunstone, pp. 59-97. Florida: CRC Press. ). The SFC values for palm oil were found to be higher than that of lard within the temperature ranges of 0 to 20 °C (Yanty et al., 2012Yanty NAM, Marikkar JMN, Miskandar MS. 2012. Comparing the thermo-physical properties of lard and selected plant fats. Grasas Aceites. 63, 328-334. https://doi.org/10.3989/gya.023712. ). In fact, a tremendous gap existed between these two at the beginning of solidification and melting end-point. For instance, at 0 ºC the SFC of lard and palm oil were 30.8 and 68.63%, respectively, but their SFC values tended to become 0% at 40 and 55 ºC, respectively. It is generally assumed that by blending an appropriate amount of liquid oil such as soybean oil with palm oil, the SFC value can be adjusted to similar to that of lard at all temperatures (Siew, 2002Siew WL. 2002. Palm oil. In Vegetable Oils in Food Technology: Composition, Properties and Uses, ed. F.D. Gunstone, pp. 59-97. Florida: CRC Press. ).

Soybean oil can be an oil of choice to be used for blending with palm oil because the SFC value of soybean oil at 0 ºC was 0.31%, and would become 0% at 5 ºC. Based on this hypothesis, an attempt was made to develop a quaternary fat blend comprising of palm oil and soybean oil as its major components (Marikkar et al., 2018Marikkar JMN, Yanty NAM, Peciulli M, Miskandar MS, Chiavaro E. 2018. Composition and thermal properties of quaternary mixtures of palm oil:palm stearin:soybean oil:cocoa butter. Ital. J. Food Sci. 30, 740-751. https://doi.org/10.14674/IJFS-1242 ). The question related to how much soybean oil would be just enough to make a proper blending with palm oil indeed needs to be worked out experimentally. For this purpose, the SFC profiling of fat blends at various temperatures would be inevitable. SFC is generally expressed as the percentage that corresponds to the total meltdown of a fat into liquid (Madison and Hill, 1978Madison BL, Hill RC. 1978. Determination of solid fat content of commercial fats by pulsed nuclear magnetic resonance. J. Am. Oil Chem. Soc. 55, 328-331. https://doi.org/10.1007/BF02669922.). The SFC profiles of three binary blends of PO:SBO made at the beginning are shown in Figure 1, which might give some indication for the selection of the best blending proportions. Out of these three binary blends prepared, the addition of 50% of soybean oil into palm oil resulted in SFC values similar to lard from the range 30 to 45 °C, but lower than that of lard within the temperature range 0 to 30 °C (Figure 1B). This deviation, however, could be overcome by adding a hard fatty substance like palm stearin or cocoa butter in smaller amounts due to their role in the adjustment of various physical and functional properties. Although the additions of 5% of PS into PO:SBO blends tended to increase the SFC values at 0 to 27 °C, the SFC values were found to increase higher than those of lard in the range of 27 to 35 °C, as shown in Fig. 1C. However, the addition of 5% CB into PO:SBO blends was found to help adjust the SFC values of the formulated fat blends and lard to become similar within the range of 27 to 35 °C, as shown in Figure 1A.

With regard to degree of unsaturation, IV of all quaternary fat blends were found to be significantly (p < 0.05) higher (92.26 to 96.95) than that of lard (73.76) (Table 2). This characteristic feature was somewhat similar to the observation made previously in the case of EF:CO binary blends (Nur Illyin et al., 2013Nur Illiyin MR, Marikkar JMN, Shuhaimi M, Mahiran B, Miskandar MSA. 2013. Comparison of the thermo physical behavior of Engkabang (Shorea macrophylla) seed fat - canola oil blends and lard. J. Am. Oil Chem. Soc. 90, 1485-1493. https://doi.org/10.1016/j.jfoodeng.2006.01.058.). This was primarily due to the presence of a high amount of soybean oil in the blends. On the other hand, SMP values for all quaternary blends (18.0 to 21.25 °C) were lower than that of lard (29.25) (Table 2). Hence, the SMP and IV of none of the quaternary blends were found to become similar to that of lard (Marikkar et al., 2018Marikkar JMN, Yanty NAM, Peciulli M, Miskandar MS, Chiavaro E. 2018. Composition and thermal properties of quaternary mixtures of palm oil:palm stearin:soybean oil:cocoa butter. Ital. J. Food Sci. 30, 740-751. https://doi.org/10.14674/IJFS-1242 ). The additions of soybean oil into palm oil would cause significant increases in the amounts of linoleic acid (from 10.25 to 34.44%) with concurrent decreases in the proportions of palmitic (from 43.99 to 25.72%) and oleic acids (from 39.24 to 27.56%). The palmitic and linoleic acid contents in the quaternary blends were significantly higher (p < 0.05) than in lard. These changes in the degree of unsaturation and fatty acid composition led to changes in the IV of these mixtures as noted in Table 2.

The preparation of quaternary fat blends by adding palm stearin, soybean oil and cocoa butter into palm oil caused some increases in certain TAG molecules (Marikkar et al., 2018Marikkar JMN, Yanty NAM, Peciulli M, Miskandar MS, Chiavaro E. 2018. Composition and thermal properties of quaternary mixtures of palm oil:palm stearin:soybean oil:cocoa butter. Ital. J. Food Sci. 30, 740-751. https://doi.org/10.14674/IJFS-1242 ). Increases in the proportions of PLL, OOL and POL could be due to the influence of soybean oil in the blends as soybean oil possesses TAG molecules namely, LLL (23.56%), OLL (17.77%), PLL (15.82%) and POL (13.69%) in high amounts (Table 4). Significant increases in the proportions of PPP and SOS were also noticed in the fat blends due to the fact that these two were major TAG species of palm stearin and cocoa butter (Nor Aini and Miskandar, 2007Nor Aini I, Miskandar MS. 2007. Utilization of palm oil and palm products in shortenings and margarines. Eur. J. Lipid Sci. Technol. 109, 422-432. https://doi.org/10.1002/EJLT.200600232 ; Segall et al., 2005Segall SD, Artz WE, Raslan DS, Ferraz VP, Takahashi JA. 2005. Analysis of triacylglycerol isomers in Malaysian cocoa butter using HPLC-mass spectroscopy. Food Res. Int. 38, 167-174. https://doi.org/10.1016/j.foodres.2004.09.008. ). Among the TAG molecular species, PPO (ranging from 13.48 to 12.09%) and POO (from 10.87 to 9.70%) showed reductions in proportions (Table 4). These changes in composition would certainly have an impact on the melting peaks of the DSC thermograms of the fat blends.

The melting profiles of the formulated fats blends were considerably different from that of the original palm oil sample. For instance, they have one additional minor peak at around -20 °C with a shoulder peak (Table 3). This emerging feature in the melting curve of quaternary blends could be due to compositional changes caused by soybean oil. According to previous studies, soybean oil was reported to have all of its thermal peaks in the low-melting region of the DSC curve. Moreover, the T_endset was found to shift to the higher temperature region after the addition of palm stearin and cocoa butter into palm oil. When compared to the melting profile of lard (T_endset =37.5 °C), all three quaternary blends had higher end-set of melting (T_endset) (at around 44 °C) as well as lower onset of melting T_onset (at around -45 °C). These DSC features could be attributed to the influence of the low melting TAG molecules of soybean oil, which tended to crystallize at lower temperatures. Although there were great differences in melting transitions between lard and the three quaternary blends, a closer similarity was seen between them at the maximum peak of some peaks at around -3.59 °C (Marikkar et al., 2018Marikkar JMN, Yanty NAM, Peciulli M, Miskandar MS, Chiavaro E. 2018. Composition and thermal properties of quaternary mixtures of palm oil:palm stearin:soybean oil:cocoa butter. Ital. J. Food Sci. 30, 740-751. https://doi.org/10.14674/IJFS-1242 ).