Food Chemistry 132 (2012) Contents lists available at SciVerse ScienceDirect. Food Chemistry

Transcription

1 Food Chemistry 132 (2012) Contents lists available at SciVerse ScienceDirect Food Chemistry journal homepage: Analytical Methods Application of differential scanning calorimetry (DSC), HPLC and pnmr for interpretation primary crystallisation caused by combined low and high melting TAGs Sami Saadi a, Abdul Azis Ariffin a,, Hasanah Mohd Ghazali a, Mat Sahri Miskandar b, Huey Chern Boo a, Sabo Mohammed Abdulkarim a a Faculty of Food Science and Technology, Universiti Putra Malaysia, UPM Serdang, Selangor, Malaysia b Malaysian Palm Oil Board, 6 Persiaran Institusi, Bandar Baru Bangi, Kajang, Selangor, Malaysia article info abstract Article history: Received 23 December 2009 Received in revised form 18 October 2010 Accepted 29 October 2011 Available online 9 November 2011 Keywords: Fat agglomerations Temperature transition Enthalpy pnmr SFC Internal friction The main goal of the present work was to assess the mechanism of crystallisation, more precisely the dominant component responsible for primary crystal formations and fat agglomerations. Therefore, DSC results exhibited significant effect on temperature transition; peak sharpness and enthalpy at palm stearin (PS) levels more than 40 wt.%. HPLC data demonstrated slight reduction in the content of POO/ OPO at PS levels less than 40 wt.%, while the excessive addition of PS more than 40 wt.% increased significantly PPO/POP content. The pnmr results showed significant drop in SFC for blends containing PS less than 40 wt.%, resulting in low SFC less than 15% at body temperature (37 C). Moreover, the values of viscosity (g) and shear stress (s) at PS levels over 40 wt.% expressed excellent internal friction of the admixtures. All the data reported indicate that PPO/POP was the major component of primary nucleus developed. In part, the levels of PS should be less than 40 wt.%, if these blends are designed to be used for margarine production. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Palm oil (PO), an important edible oil source for the food industry, able to crystallize as beta prime polymorph and therefore is an attractive option for the production of margarine and shortenings (Watanabe, Tashima, Matsuzaki, Kurashige, & Sato, 1992). The major component of palm oil is 1,3-dipalmitoyl 2-oleoyl glycerol (POP) (Okawashi, Sagi, & Mori, 1985). Palm stearin (PS) is the hard fraction resulted from fractionation of palm oil after crystallisation under controlled temperature, with high melting point ranging from 44 to 56 C, giving the product low plasticity and incomplete melting at body temperature (Aini & Miskandar, 2007). Miura and Konishi (2001) studied the granular crystals composition in margarine produced by an excess of palm oil, showing that 1,3-dipalmitoyl-2-oleoyl-glycerol (POP) was one of the major TAGs content exist in the granular crystals. They found that POP and POO were immiscible with each other, resulting in POP crystals surrounded by POO. Zhang, Ueno, Sato, and Miura (2007) suggested that formations of granular crystals are referred to the agglomeration of higher-melting TAGs, reflected by more stable beta crystal polymorph. Minato, Ueno, Smith, Amemiya, and Sato (1997) studied the Corresponding author. Tel.: ; fax: address: abdulazis@putra.upm.edu.my (A.A. Ariffin). crystallisation behaviour of PPO and POP mixtures, and they reported the formation of molecular compound systems between both contents. In part, Tanaka, Miura, and Yoshioka (2007) suggested that the agglomeration of higher-melting triglycerides (TAGs), such as PPP, led to the formation of granular crystals without polymorphic emergence in the margarine product. Generally, the most common methods to study primary crystallisation of fats are differential scanning calorimetry (DSC) and pulsed nuclear magnetic resonance (pnmr) (Wiking, De Graef, Rasmussen, & Dewettinck, 2009). Differential scanning calorimetry (DSC) heating profiles give valuable information about the melting, temperature transition and solidification properties of fats. It can also be used to determine crystallisation profiles and at what temperature the crystals start to form (Lai, Ghazali, Cho, & Chong, 2000). Specific study has demonstrated that heating thermograms of crude palm oil (CPO) conducted by DSC exhibited a clear separation between a liquid part (olein) and solid part (stearin) at 31 C (Tan, Ghazali, Kuntom, Tan, & Ariffin, 2009). On the other hand, Miskandar, Che Man, Yusoff, and Abdul Rahman (2002) have reported that the crystal network was the ability to increase slightly the SFC higher than the normal solid fat content profile (SFCP). In addition, Tanaka et al. (2007) have reported also, that higher melting TAGs, such as POP, SOS or SSS become essential components of /$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi: /j.foodchem

2 604 S. Saadi et al. / Food Chemistry 132 (2012) the growing crystals and resulting agglomerates. Moreover, the viscosity of oil and fat blend can affect the rate of crystallisation, especially at high supersaturated degrees of fatty acids (Marangoni, 2005). Another study, which was reported by Mulder and Walstra (1974), showed that fat crystal morphology and polymorph of crystal developed can substantially affect rheology. As example, they reported that slow regime of crystallisation, with a SFC greater than 20%, contained large spherulites. On the other hand, several studies on crystallisation behaviour of oils and fats demonstrated an acceptable range of SFC of 10% to 15%, which can be manufactured (deman, deman, & Blackman, 1995; Chrysam, 1996). Therefore, the main goal of this work was to assess the effect of low and high melting TAGs, and their interactions on the physicochemical and rheological properties of PO/PS blend systems, more precisely the dominant component responsible for primary crystal formations using DSC, HPLC, and pnmr analytical machines. 2. Materials and methods 2.1. Materials Two types of fats were used to conduct this study. The refined bleached and deodorized (RBD) palm stearin (PS) and RBD palm oil (PO) were purchased from the Mewah Oleo Food Industries, a local refinery (Selangor, Malaysia). All reagents and solvents used were of analytical or HPLC grade and obtained from Merck (Darmstadt, Germany). Fatty acid methyl esters (FAME) and triacylglycerol standards were supplied from Sigma Chemical Co (St. Louis, MO, USA) Admixtures formulation design An analytical balance with standard error of ±0.001 g was calibrated for prior measurement. Eleven ratios of RBD palm oil and palm stearin were done in percentage of weight total (wt.%). The ratios of RBD palm oil were ranged from 100% to 0%, while for RBD palm stearin were varied from 0% to 100% in 10% increments (wt./wt.) All admixtures of RBD palm oil and palm stearin were melted at 70 C, and stirred vigorously for 5 min, then blended using low rotation speed for 2 min. These conditions were applied to give same crystallisation rate for all blends and right distribution of crystals size. All admixtures were carried out in triplicate and kept immediately at room temperature (28 C) for 24 h to ensure a good stability of molecules prior to analysis Iodine value (IV) The analysis of iodine value was determined according to AOCS methods Cd 1-25 (AOCS., 1989) Fatty acids composition (FAC) by GC The determination of fatty acid composition of palm oil, palm stearin, and their admixtures was done after the conversion of fatty acid into corresponding fatty acid methyl esters (FAME) (Ariffin et al., 2009). A sample of 50 mg was melted and weighed out separately in triplicate. An amount of 950 ll of hexane was added to a blend sample, followed by 50 ll of methoxide solution. The mixture left to stand for 5 min after shaken by vortex mixer. A diluted blend sample (1/100 v/v, in n-hexane) of 1 ll was injected manually at 20 C, and in the splitless mode. The procedure was carried out on an analytical gas chromatograph (Perkin Elmer, USA). Clarus 500 gas chromatograph fitted with an Elite 5, capillary column (30 m 0.25 mm id; 0.25 lm film thickness). Helium was the carrier gas (1 ml min 1 ). Column temperature was 130 C held for 0.5 min and increases at rate of 10 C/min to 240 C and held for 5 min. The total runtime selected was min and flame ionisation detection (FID) was performed at 240 C Triacylglycerols composition (TAG) by HPLC The identification of triacylglycerols was done according to the procedure described by Mamat, Nor Aini, Said, and Jamaludin (2005). A high performance liquid chromatograph (HPLC) model 510, supported with differential refractometer, Model 410, as the detector (Millipore Corporation, Milford, MA) was used to identify the TAGs of PO, PS, and their admixtures. The triacylglycerols were separated on a Merck Lichrosphere RP-18 Column (250 mm 4 mm, particle size 5 lm) (Darmstadt, Germany). Column temperature was maintained at 45 C during running, where the mobile phase was consisted from a mixture of acetone/acetonitrile at a ratio of 70:30 (vol/vol). The flow rate was regulated at values of 1 ml/min. The triacylglycerols were separated according to their degree of unsaturation and molecular weight (Tan & Che Man, 2002). Triacylglycerol peaks were identified referring to the retention times of TAG standards (Sigma Chemical Co., USA) DSC thermal analysis of blends formulation A Perkin Elmer Diamond differential scanning calorimetry (Shelton, CT, USA) was used for investigation the thermal characteristics of blend samples. The calibration of the instrument was done by indium and zinc. A sample of 6 8 mg was weighed into aluminium pans. An empty covered pan was used as a reference. The applied running conditions for DSC sample were as follows: for isothermal the sample was heated from 70 C to70 C, where it was held for 2 min at scan rate of 10 C/min. Concerning the cooling thermograms the blend sample was cooled from 70 C to 70 C at the same time and same scan rate used for the isothermal. All the parameter values (Onset temperature, enthalpy, and end set temperature) were calculated automatically by manufacturer s software (Diamond DSC software library), supported with program used to analyze and plot the thermal results SFC profile by pulsed nuclear magnetic resonance (pnmr) Solid fat content (SFC) was measured using a Bruker Minispec pulse Nuclear magnetic resonance (pnmr) spectrometer (Karlsruhe, Germany). The blend sample in the NMR was first melted at 70 C for 30 min, followed by chilling at 0 C for 90 min, and then held at each measuring temperature of 5, 10, 15, 20, 25, 30, 35, 37, 40, 45, and 50 C for 30 min prior to analysis (PORIM, 1995). Melting, chilling and holding of the blend samples were carried out in pre-equilibrated thermostat water baths, accurate to 0.1 C (Lai et al., 2000; Norizzah, Chong, Cheow, & Zaliha, 2004). The solid fat content values were given as means of triplicate and the final results tabulated as percentage unit (%) Determination of viscosity (g) and shear stress (s) The rheological properties of refined bleached and deodorized PO, PS and their admixtures were investigated by means of a controlled-stress rheometer (Rheostress 600, Haake, Karlsuruhe, Germany) with a sand-blasted cone sensor (C35/2 Ti; 35 mm diameter, 2 angle, mm gap) and a measuring plate cover (MPC 35) (Nor Hayati, Che Man, Tan, & Nor Aini, 2009). Each homogenized blend was subjected at different temperatures of 20, 25, 30, 40, and 50 C. Calibrations of the instrument and setting up the measurement conditions were done automatically by software (Haak software, Rheowin Job Manager Version 3.12). All

3 S. Saadi et al. / Food Chemistry 132 (2012) measurements were determined at shear rate ranging from 10s 1 to 20s 1, in triplicate Statistical analysis All statistical analyses were performed using the Minitab software (Version 14, Minitab. USA). One-way analysis of variance and Tukey s honest multiple comparison test with family error rate of 5% were performed to test the differences among the treatment groups. A multiple regression model was developed for the analytical responses (solid fat content, viscosity and shear stress) according to Hare (1974), Andreia et al. (2009), Silva et al. (2009), represented by the following equation: y ¼ b 1 x 1 þ b 2 x 2 þ b 12 x 1 x 2 Where y is the variable and b are the coefficients generated by multiple regressions; and x is the proportion of the components. 3. Results and discussion 3.1. Effect of PO, PS, and their admixtures on FAC and degree of unsaturation Table 1 shows the iodine value and fatty acids composition of RBD palm oil, palm stearin, and their admixtures. The results showed that the IV of palm oil was ± 0.19 g I 2 /100 g, while for palm stearin it was ± 0.51 g I 2 /100 g. The decreasing in the IV of PO/PS admixtures refers to a decreasing of the contents of unsaturated fatty acids or the average numbers of double bonds that existed in the admixtures. The unsaturated fatty acids are chemically more active than the saturated fatty acids, which directly the activity increases as the number of double bonds increase (O Brien, 2004). Two principles fatty acids that were clearly affected by the addition of PS. Palmitic acid progressively increased from ± 0.31% to ± 0.54% at low PS levels less than 40 wt.%, then increased significantly from ± 0.34% to ± 0.20% when the stearin ratio exceeded 40 wt.% of the blend. The palmitic acid content was related with texture degradation at high concentration among the fatty acid species (Tanaka, Isogai, Miura, & Murakami, 2010). An opposite effect was observed for oleic acid (omega-9), where low PS levels less than 40 wt.% slightly reduced omega-9 from ± 0.12% to ± 0.78%. The excessive addition of PS more than 40 wt.% significantly lowered the content of omega-9 from ± 0.78% to ± 0.12%. According to Siew and Ng (1999) the low level of linoleic acid and virtual absence of linolenic acid made the oil relatively stable to oxidative deterioration. The high saturated TAGs are gradually aggregated in the solid form (stearin) during crystallisation proceeds, leaving behind a more unsaturated TAGs liquid part (olein) (Zaliha, Chong, Cheow, Norizzah, & Kellens, 2004). It is apparent from these results that the major dominant fatty acid content is palmitic acid which can significantly affect the ratio of saturated/unsaturated fatty acids at high PS levels more than 40 wt.% Effect of PO, PS, and their admixtures on TAGs contents Comparative TAG profiles of RBDPO and RBDPS is shown in Fig. 1a and b. The TAG profile of PO and PS contains fifteen types of TAG in total for palm oil/palm stearin. The detection of each component was done according to the degree of unsaturation and molecular weight of fatty acids. Therefore, the HPLC analysis based on its chromatogram coupled with retention time showed two principal peaks. The first peak was unsaturated TAG (POO/OPO), while the second peak was saturated TAG (PPO/POP). Both peaks were detected for PO/PS, and their admixtures. The omparison between peaks height of POO/OPO and PPO/POP contents showed significant differences for PS rather than PO, where the differences between them in PO was not significant. This point may explain the natural equilibrium between liquid part (olein) and solid part (stearin). These results agreed well with those reported by Tan and Che Man (2002) that palm oil is a balanced oil in which the ratio of saturated fatty acids equal to the unsaturated fatty acids. Table 1 shows the total composition of TAG of the eleven admixtures. The level of PPO/POP increased progressively from 32.4 ± 0.4% at PS level zero to 35.0 ± 0.4% at PS levels less than 40 wt.%. The excessive addition of hard fraction (PS) more than 40 wt.% in the blend showed significant increasing of PPO/POP content to 41.2 ± 0.0% at PS level 100 wt.%. An opposite effect has been detected for POO/OPO content where the level of unsaturated TAG slightly reduced from 25.7 ± 0.3% at PS level zero (100% PO) to 23.5 ± 0.4% at PS levels 40 wt.%. The excessive addition of PS levels more than 40 wt.% significantly drops the content of POO/OPO to 16.3 ± 0.1 at PS level 100 wt.% of the blend. Triacylglycerol species with excess presence of palmitic acid such as tripalmitin (PPP) and 1,3-dipalmitoyl-2-oleoyl-glycerol (POP) and 1,2-dipalmitoyl-3- oleoyl-glycerol (PPO) were related with texture degradation (Tanaka et al., 2010). Moreover, the content of POL as well was significantly affected by addition of high levels of PS more than 40 wt.%, resulting in a decrease of POL content from 9.6 ± 0.2% to 7.0 ± 0.0%. It is apparent from these results that two major TAG contents (PPO/POP and POO/OPO) that were clearly affected by the addition of high PS levels more than 40 wt.%. The diunsaturated TAGs (POO/ OPO) content which is related with the amount of oleic acid (omega-9) and disaturated TAGs (PPO/POP) content which is also dominated by palmitic acid content (Table 1). This conclusion agreed well with that made by Tanaka, K-Tanaka, Yamato, Ueno, and Sato (2009) that from the data of fatty acid and TAG compositions the major fats and their polymorphic forms in palm oil based system are PPP, POP, and PPO, all of which are in the metastable b 0 form of the double chain length structure Effect of PO, PS, and their admixtures on crystallisation properties Differential scanning calorimetry (DSC) is an efficient thermoanalytical method applied for monitoring different temperature transition during cooling and heating thermograms. The most measured parameters during cooling stage (crystallisation) of PO/ PS admixtures are: Onset temperature, enthalpy (heat flow status), and end set temperature. The onset temperature of crystallisation profile is characterized by the beginning of fat crystal formation, which is related with the rearrangement of molecules due to the presence of high saturated TAGs, while the end set temperature may explain by the end of crystallisation, which is generally reflected by aggregation and compaction of molecules. On the other hand, and based on thermodynamic rules the enthalpy can be defined as the amount of energy which is lost from the sample during crystallisation and gained from the air surrounded during melting. The changes in the temperature transition of cooling thermograms displayed the behaviour of triacylglycerols contents as principal (PPO/POP, POO/OPO, and PLO). Tan and Che Man (2002) has reported that oils and fats may show an extremely complex thermal behaviour, depending highly on the chemical composition and the protocol for the DSC experiment. The crystallisation curve of RBD palm oil/palm stearin, and their admixtures exhibited two principal peaks namely P1, and P2 (Fig. 2a). At low PS levels less than 40 wt.% the peaks were appeared smoother than those containing PS more than 40%. In contrast, the sharpness of the peaks and the increasing of peak area at high PS levels more than 40 wt.% may refer to the presence of high content of saturated fatty acid as palmitic acid and saturated TAGs

4 Table 1 Total fatty acids and triacylglycerols composition coupled with iodine value (IV) of palm oil, palm stearin, and their admixtures. FA 1 /TAG 1 FA and TAGs in% 100%:0% 90%:10% 80%:20% 70%:30% 60%:40% 50%:50% 40%:60% 30%:70% 20%:80% 10%:90% 0%:100% FA C:14: ± 0.53 a 1.31 ± 0.19 a 1.34 ± 0.23 a 1.36 ± 0.37 a 1.50 ± 0.27 b 1.62 ± 0.16 bc 1.87 ± 0.14 c 2.26 ± 0.20 cd 2.85 ± 0.10 d 3.27 ± 0.19 de 3.51 ± 0.11 e C:16: ± 0.31 a ± 0.37 b ± 0.38 bc ± 0.60 c ± 0.54 c ± 0.34 d ± 0.48 de ± 0.27 e ± 0.40 e ± 0.96 e ± 0.20 e C:18: ± 0.28 a 2.61 ± 0.60 b 2.63 ± 0.18 b 2.66 ± 0.11 b 2.82 ± 0.11 b 2.97 ± 0.20 bc 3.16 ± 0.10 c 3.35 ± 0.13 d 3.57 ± 0.10 d 3.74 ± 0.70 d 3.86 ± 0.10 d C:18: ± 0.12 a ± 0.85 b ± 0.70 bc ± 0.28 c ± 0.78 c ± 0.30 d ± 0.53 d ± 0.39 de ± 0.40 e ± 0.73 e ± 0.12 e C:18: ± 0.56 a 9.19 ± 0.30 b 8.60 ± 0.70b c 8.31 ± 0.30b c 7.98 ± 0.74 c 7.77 ± 0.33 cd 7.20 ± 0.10 d 7.18 ± 0.10 d 7.16 ± 0.30 d 7.12 ± 0.50 d 7.05 ± 0.30 d P SFA ± 0.66 a ± 0.37 b ± 0.47 bc ± 0.84 c ± 0.53 c ± 0.70 cd ± 0.96 cd ± 0.39 d ± 0.40 de ± 0.92 de ± 0.12 e P USFA ± 0.66 a ± 0.37 b ± 0.47 bc ± 0.84 c ± 0.53 c ± 0.70 cd ± 0.96 cd ± 0.39 d ± 0.40 de ± 0.92 de ± 0.12 e IV ± 0.19 a ± 0.50 b ± 0.68 bc ± 0.52 bc ± 0.82 c ± 071 cd ± 0.63 cd ± 0.92 d ± 0.28 de ± 0.82 de ± 0.51 e TAGs MPL 1.4 ± 0.2 a 1.5 ± 0.0 a 1.5 ± 0.0 a 1.6 ± 0.0 a 1.7 ± 0.0 a 1.8 ± 0.0 ab 1.7 ± 0.4 ab 1.9 ± 0.0 b 1.7 ± 0.0 ab 1.6 ± 0.0 a 1.5 ± 0.0 a OOL 0.5 ± 0.0 a 0.4 ± 0.0 a 0.4 ± 0.0 a 0.3 ± 0.0 b 0.3 ± 0.0 a 0.3 ± 0.2 a 0.3 ± 0.0 a 0.3 ± 0.0 a 0.3 ± 0.0 a 0.3 ± 0.0 a 0.3 ± 0.0 a MMP 1.7 ± 0.0 a 1.9 ± 0.0 a 1.7 ± 0.0 a 1.9 ± 0.1 b 1.6 ± 0.0 a 1.4 ± 0.0 c 1.4 ± 0.0 c 1.3 ± 0.0 cd 1.2 ± 0.0 d 1.3 ± 0.0 de 1.1 ± 0.0 e POL 12.4 ± 0.2 a 11.8 ± 0.0 ab 11.2 ± 0.0 ab 10.8 ± 0.2 b 10.8 ± 0.1 b 9.6 ± 0.2 c 9.4 ± 0.5 c 8.5 ± 0.0 d 8.4 ± 0.5 d 7.2 ± 0.0 d 7.0 ± 0.0 d PPL 7.8 ± 0.5 a 8.1 ± 0.1 a 8.2 ± 0.0 a 8.3 ± 0.0 a 8.3 ± 0.0 a 10.7 ± 0.5 b 10.9 ± 0.2 bc 10.9 ± 0.0 bc 10.7 ± 0.3 bc 11.4 ± 0.0 c 11.3 ± 0.0 c MPP 0.4 ± 0.0 a 0.6 ± 0.0 ab 0.6 ± 0.0 ab 0.5 ± 0.3 ab 0.6 ± 0.1 ab 0.5 ± 0.4 ab 0.8 ± 0.0 ab 0.9 ± 0.0 b 1.0 ± 0.0 b 1.0 ± 0.0 b 1.2 ± 0.1b c OOO 5.2 ± 0.0 a 4.0 ± 0.0 a 4.0 ± 0.0 a 4.1 ± 0.0 a 3.6 ± 0.0 a 3.0 ± 0.2 b 3.2 ± 0.1 b 2.9 ± 0.0 b 3.0 ± 0.2 b 2.7 ± 0.1 c 2.4 ± 0.0 c POO/OPO 25.7 ± 0.3 a 24.9 ± 0.4 ab 24.4 ± 0.3 b 24.1 ± 0.0 b 23.5 ± 0.4 b 21.8 ± 0.5 c 20.8 ± 0.2 c 19.8 ± 0.1 cd 18.3 ± 0.2 d 17.7 ± 0.3 d 16.3 ± 0.1 e PPO/POP 31.4 ± 0.4 a 33.0 ± 0.4 b 34.3 ± 0.1 c 34.4 ± 0.4 c 35.0 ± 0.4 cd 35.9 ± 0.3 cd 37.4 ± 0.3 d 38.0 ± 0.0 d 40.0 ± 0.5 e 40.0 ± 0.2 ef 41.2 ± 0.0 f PPP 4.1 ± 0.0 a 4.9 ± 0.5 ab 4.5 ± 0.3 a 4.8 ± 0.1 ab 4.9 ± 0.0 ab 5.3 ± 0.5 b 4.3 ± 0.4 a 5.5 ± 0.2 ab 5.4 ± 0.0 ab 6.3 ± 0.0b c 7.0 ± 0.1 c SOO 3.1 ± 0.0 a 2.1 ± 0.0 b 2.1 ± 0.0 b 2.2 ± 0.2 b 2.1 ± 0.0 b 2.0 ± 0.4 b 1.8 ± 0.0 bc 1.7 ± 0.0 bc 1.4 ± 0.2 c 1.5 ± 0.0 c 1.4 ± 0.0 c PSO 4.9 ± 0.4 a 5.2 ± 0.1 ab 5.6 ± 0.2 ab 5.5 ± 0.3 ab 5.8 ± 0.0 b 5.9 ± 0.2 b 6.1 ± 0.3 b 6.0 ± 0.0 b 5.9 ± 0.4 b 6.8 ± 0.2 c 6.4 ± 0.0 bc PPS 0.6 ± 0.0 a 0.7 ± 0.0 a 0.6 ± 0.0 a 0.7 ± 0.0 a 0.8 ± 0.0 a 1.0 ± 0.1 b 1.1 ± 0.2 b 1.5 ± 0.0 c 1.6 ± 0.0 c 1.3 ± 0.0 bc 2.0 ± 0.0 d SSO 0.4 ± 0.0 a 0.4 ± 0.0 a 0.5 ± 0.0 a 0.5 ± 0.0 a 0.6 ± 0.0 a 0.5 ± 0.0 a 0.5 ± 0.0 a 0.5 ± 0.0 a 0.6 ± 0.0 a 0.6 ± 0.2 a 0.6 ± 0.0 a 606 S. Saadi et al. / Food Chemistry 132 (2012) Abbreviations: PO, palm oil; PS, palm stearin; IV, iodine value (g I 2 /100 g); FA, fatty acids; P SFA, sum of saturated fatty acids; P USFA, sum of unsaturated fatty acids; C:14:0, myristic acid; C:16:0, palmitic acid; C:18:0, stearic acid; C:18:1, oleic acid (omega-9); C:18:2, linoleic acid (omega-6). TAG 1, triacylglycerol; MMM, trimyristic; MPL, myristoyl-palmitoyl-linoleoyl-glycerol; OOL, dioleolyl-linoleoyl-glycerol; MMP, dimyristoyl-palmitoyl-glycerol; POL, palmitoyl-oleoyl-linoleoyl-glycerol; PPL, dipalmitoyl-linoleoylglycerol; MPP, dipalmitoyl-myristoyl-glycerol; OOO, triolein; POO/OPO, dioleoyl-palmitoyl-glycerol; PPO/POP, dipalmitoyl-oleoyl-glycerol; PPP, tripalmitic; SOO, dioleolyl-stearoyl-glycerol; PSO, palmitoyl-stearoyl-oleoyl-glycerol; PPS, dipalmitoyl-stearoyl-glycerol; SSO, distearoyl-oleoyl-glycerol. a-f Means with different letters within each row are significantly different (n =3;p < 0.05). 1 Each value in the table represents the mean standard deviation of triplicate analysis (mean ± SD).

5 S. Saadi et al. / Food Chemistry 132 (2012) Fig. 1. Triacylglycerol (TAG) profiles (chromatograms) of (a) refined-bleached-deodorized (RBD) palm oil (PO) and (b) palm stearin (PS) as a function of retention time using high performance liquid chromatography (HPLC) fitted by RP C18 (25 cm length 4 mm diameter) column and mobile phase of 70:30 (v/v) acetone and acetonitrile HPLC grade mixtures. as PPO/POP content. These observations are in accordance with the suggestion made by Tanaka et al. (2010) that the agglomeration of higher-melting TAGs such as POP and PPO become components of the growing crystals and promoted the formation of granular crystals as crystal nuclei. The cooling thermograms clearly show a significant difference in the end set temperature of peaks for blends containing PS less than 40 wt.%, and those contained high PS levels more than 40 wt.% (Fig. 2a). These results agreed well with other researchers that peaks crystallisation temperature refers to the temperature at which the highest proportion of fat species crystallizes with maximum thermal effect (Campos, Kawachi, Oliveira, & Thim, 2005). Moreover, the energy (enthalpy) liberated for crystal formation (aggregation) differ from peak to another. At the first peak enthalpy liberated for 0%PO:100%PS was four times more than that of 100%PO:0%PS, and decreased around two times for that of 70%PO:30%PS (Table 2). On the other hand, indicates the wide contribution of high saturated triacylglycerols of blends those containing PS more than 40 wt.%. These results are in accordance with those reported by Tan and Che Man (2002) that the subdivision of crystallisation curve of any oil or fat in different exothermic regions, corresponds to different types of triacylglycerols. Consequently, it is apparent from this discussion that the rate of crystallisation clearly increased by high melting TAGs as PPO/POP content, when the amount of % PS exceeded 40 wt.% and totally decreased with low melting triacylglycerols as POO/OPO content Effect of PO, PS, and their admixtures on melting properties According to Fredrick, Foubert, Van De Sype, and Dewettinck (2008) the melting profiles give an indication of the amount of crystallized fat and the occurrence of polymorphic transitions. All heating thermograms obtained by DSC for PO, PS, and their admixtures are shown in Fig. 2(b). The melting profiles exhibited two endothermic events separated by an exothermic event at the centre. The exothermic event, which located between 15 C and 25 C was found shifted gradually towards higher temperatures and almost disappeared at stearin level zero. In contrast, two

6 608 S. Saadi et al. / Food Chemistry 132 (2012) Fig. 2. (a) Cooling thermograms or crystallisation profile of refined-bleached-deodorized palm oil, palm stearin, and their admixtures. According to Tan and Che Man (2002) the DSC parameters are illustrated by constructed lines, where the onset and offset (end) temperatures corresponded closely to the intersection of the extrapolated baseline and the tangent line of the peak. (b) Heating thermograms or melting profile of refined-bleached-deodorized palm oil, palm stearin, and their admixtures. endothermic events were clearly distinguishable. The first event was detected at low temperature in the temperature interval ranging from ( 10) C to15 C. This event was possibly associated to the liquid part olein and unsaturated TAGs such as POO/OPO, PLO and OOO contents. In palm oil the increasing peak area suggests that after a polymorphic transition extra crystals are formed from the melt (Fredrick et al., 2008). The second event was detected at high temperature levels ranging from 25 C to 50 C, which may refer to the solid fraction stearin and saturated TAGs such as PPO/POP, PPP and PPS. All these results have been found in accordance with those reported by other researchers for crude palm oil (CPO), RBD palm oil and RBD palm stearin (Che Man, Haryati, Ghazali, & Asbi, 1999; Tan & Che Man, 2002; Tan et al., 2009). New event was detected at high temperature levels ranging from 50 C to 56 C, resulting in a significant retardation in the end set temperature of blends those containing palm stearin levels more than 40 wt.% (Fig. 2b). These results are in accordance with those reported by other researchers that the decrease in the fat ratio with supersaturated degree such as mono-saturated TAGs, disaturated TAGs, palmitic and stearic acids that are expected to melt in the interval temperature of the minor shoulder peak (Che Man et al., 1999). Table 2 shows the enthalpy requirement for each peak during melting stage. As a result the second peak of melting profile

7 Table 2 Crystallisation and melting parameters (onset temperature, end temperature and peaks enthalpy) measured for palm oil, palm stearin, and their admixtures using: differential scanning calorimetry (DSC). PO%:PS%(w/w) 1st peak 2nd peak 3rd peak 1 T on ( C) 1 DH P1 (J/g) 1 T end ( C) 1 T on ( C) 1 D H P2 (J/g) 1 T end ( C) 1 T on ( C) 1 DH P3 (J/g) 1 T end ( C) Crystallisation properties 100% PO:0% PS ± 0.09 a ± 0.69 a ± 0.14 a 5.94 ± 0.08 a ± 0.81 a 8.33 ± 0.57 a ND ND ND 90% PO:10% PS ± 0.46 a ± 0.51 ab ± 0.52 a 6.08 ± 0.02 b ± 0.16 b 8.36 ± 0.08 a ND ND ND 80% PO:20% PS ± 0.34 ab ± 0.24 ab ± 0.28 ab 6.24 ± 0.14 c ± 0.14 c 7.68 ± 0.10 b ND ND ND 70% PO:30% PS ± 0.23 ab ± 0.70 b ± 0.56 b 6.27 ± 0.07 d ± 0.56 c 7.23 ± 0.11 b ND ND ND 60% PO:40% PS ± 0.24 b ± 0.39 b ± 0.18 b 6.51 ± 0.03 e ± 0.64 d 7.18 ± 0.08 c ND ND ND 50% PO:50% PS ± 0.20 b ± 0.29 bc ± 0.30 c 6.77 ± 0.31 ef ± 0.45 e 6.20 ± 0.36 d ND ND ND 40% PO:60% PS ± 0.21 bc ± 0.41 c ± 0.45 d 6.87 ± 0.14 f ± 0.21 f 5.39 ± 0.40 de ND ND ND 30% PO:70% PS ± 0.44 bc ± 0.81 bc ± 0.17 de 6.96 ± 0.03 g ± 0.22 g 5.16 ± 0.15 e ND ND ND 20% PO:80% PS ± 0.33 c ± 0.73 d ± 0.41 e 7.42 ± 0.47 gh ± 0.30 h 4.59 ± 0.09 ef ND ND ND 10% PO:90% PS ± 0.27 cd ± 0.57 d ± 0.48 ef 7.92 ± 0.39 h ± 0.27 i 4.19 ± 0.36 f ND ND ND 0% PO:100% PS ± 0.14 d ± 0.83 e ± 0.87 f 8.18 ± 0.51 i ± 0.80 j 3.55 ± 0.53 f ND ND ND Melting properties 100% PO:0% PS ± 0.41 a ± 0.67 a ± 0.03 a ± 0.07 a ± 0.59 a ± 0.04 a ND ND ND 90% PO:10% PS ± 0.10 a ± 0.17 ab ± 0.32 a ± 0.27 b ± 0.69 b ± 0.49 b ND ND ND 80% PO:20% PS ± 0.13 a ± 0.17 ab ± 0.14 ab ± 0.22 c ± 0.34 c ± 0.74 c ND ND ND 70% PO:30% PS ± 0.67 a ± 0.10 b ± 0.04 ab ± 0.21 cd ± 0.48 c ± 0.70 c ND ND ND 60% PO:40% PS ± 0.14 b ± 0.45 b ± 0.50 ab ± 0.24 d ± 0.84 d ± 0.02 d ± 0.26 a 0.71 ± 0.10 a ± 0.46 a 50% PO:50% PS ± 0.26 c ± 0.85 bc ± 0.53 b ± 0.70 de ± 0.59 d ± 0.24 de ± 0.73 a 1.37 ± 0.48 a ± 0.57 ab 40% PO:60% PS 9.81 ± 0.32 ac ± 0.54 c ± 0.04 b ± 0.14 e ± 0.75 e ± 0.18 e ± 0.24 b 3.83 ± 0.78 b ± 0.52 ab 30% PO:70% PS 8.51 ± 0.37 ac ± 0.47 d ± 0.05 b ± 0.22 f ± 0.51 ef ± 0.08 ef ± 0.04 b 5.65 ± 0.56 b ± 0.28 b 20% PO:80% PS 8.05 ± 0.13 c ± 0.63 e ± 0.08 bc ± 0.12 g ± 0.59 ef ± 0.05 ef ± 0.03 b 8.74 ± 0.63 c ± 0.24 c 10% PO:90% PS 6.85 ± 0.58 c ± 0.14 f ± 0.63 c ± 0.37 h ± 0.70 f ± 0.04 f ± 0.27 c ± 0.54 c ± 0.30 cd 0% PO:100% PS 1.35 ± 0.24 d ± 0.95 g ± 0.15 c ± 0.38 i ± 0.93 g ± 0.20 f ± 0.83 c ± 0.72 d ± 0.09 d Abbreviations: Ton ( C), onset temperature; Tend ( C), end set temperature; DH P1, DH P2 and DH P3, peaks enthalpy; ND, not detected. a j Means with different letters within each column are significantly different (n =3;p < 0.05). 1 Each value in the table represents the mean standard deviation of triplicate analysis (mean ± SD). S. Saadi et al. / Food Chemistry 132 (2012)

8 610 S. Saadi et al. / Food Chemistry 132 (2012) Fig. 3. (a) Solid fat content profile (SFCP) as a function of temperature ( C) and stearin amount (wt.%) formulated PO/PS admixtures. (b) Correlation between shear stress and viscosity: g = f (s), under controlled conditions of shear rate ranged between 10 and 20s 1, times of 1 min, and rotation speed of 100 rpm. (c) and (d) Represent the evolution of viscosity and shear stress as a function of temperature ( C) and palm stearin ratio (wt.%) under same controlled conditions. indicated high level of enthalpy of ± 0.93 (J/g) at PS level of 100%, whereas less enthalpy of ± 0.59 (J/g) was at PS level of 0%. Both enthalpy values of PS and PO provide a good expression on enthalpy requirements during interactions. Consequently, the excessive addition of palm stearin up to 40 wt.% of the blend caused an apparition of new event at high temperature levels (50 56 C). This event was possibly attributed to the saturated TAGs. In contrast, the addition of % PS less than 40 wt.% kept the internal chemical composition of the admixtures between low and mid melting triacylglycerols Effect of PO, PS, and their admixtures on solid fat content (SFC) Solid fat content (SFC) is considered one of the qualitative parameter of the texture of margarine. It is responsible for many characteristics of fat blend system including general appearance, organoleptic properties, and spreadability (Chu, Ghazali, Lai, Che Man, & Yusof, 2002). The evolution of this parameter as a function of temperature is considered one of the most approach ways for better understanding the behaviour of high and low melting TAGs, which in most cases are appeared as fat agglomerations and crystal aggregations. It means SFC not only the key parameter regarding the appearance, exudation and spreadability of the product, but the deep comprehensive of the chemistry of this content is associated to the other compositions such as saturated/unsaturated fatty acids ratio, and triacylglycerols, particularly POO/OPO and PPO/ POP contents. Palm oil, a promising fat resource, involves high concentrations of POP/PPO which jointly determine the physical properties of palm oil-blended fats (Minato, Ueno, Smith, Amemiya, & Sato, 1997). The subdivision of SFC profile on two stages provides a good explanation on the evolution of SFC during tempering (Fig. 3a). The first stage located between 5 C and 25 C, which is characterized by fast dropping in SFC for all admixtures. It is clearly apparent from this period that the fast drops in SFC occurred at PS levels less than 40 wt.%, due to the increasing in the temperature levels, which led to partial disaggregation of TAG particles. POP and PPO form molecular compound crystals because of eutectic effects with the liquid oil and the dissolution temperature of every TAG must be much lower than the melting points (Tanaka et al., 2009). The second stage located in sensitive interval temperatures including moderate temperature (25 C), room temperature (28 C) and body temperature (37 C). Therefore, this stage showed significant drops in SFC for blends those containing PS levels less than 40 wt.%, resulting in low SFC less than 15% at body temperature (37 C). These results are in agreements with those reported by other researchers that a SFC of 15 25% at working temperature is desirable for better creaming performance in cake (Danthine & Deroanne, 2003). Specific study demonstrated that at 40 C, a solid fat content of % of experimental blends was noted to be an advantageous for use in cake manufacturer (Nor Aini, Abdullah, & Halim, 1992). These results were found to be in accordance to the PO/PS blends 100%:0%, 90%:10%, 80%:20% and 70%:30% (Fig 3a) Effect of PO, PS, and their admixtures on viscosity (g) and shear stress (s) Fig. 3b shows the correlation between viscosity and shear stress. The results indicate that the increasing of the viscosity (g) reflected by the increasing of shear stress (s) at PS levels more than

9 S. Saadi et al. / Food Chemistry 132 (2012) Table 3 Coefficients generated by multiple regression analysis for solid fat content (SFC), viscosity (g) and shear stress (s) of RBD palm oil/palm stearin, and their admixtures. Analytical responses (SFC, g and s) Solid fat content Viscosity (g) Shear stress (s) 40 wt.%. This point may explain by the increasing of the internal friction of blend composition due to presence of high saturated TAGs, which led to the accumulation (aggregation) of TAGs molecules as solid particles. Low PS levels less than 40 wt.% screened excellent internal friction and closed values for both parameters. Fig 3c and d exhibited linearity significant for all models concerning the evolution of viscosity (g) and shear stress (s) as a function of temperature. It is logically apparent that when the temperature increases the viscosity and shear stress decrease. On the other hand, high degree of supersaturated fatty acids increase the viscosity and shear stress, while high degree of unsaturated fatty acids lower them. The findings of this study are in agreement with those reported by Mulder and Walstra (1974) that fat crystal morphology and polymorph of crystal developed can substantially affect rheology Statistical analyses interpretation The multiple regressions model of data analysis exhibited that solid fat content (SFC), viscosity (g) and shear stress (s) were not dependent on palm stearin content (p > 0.05), but were in fact dependent on palm oil content and on the interaction of the two portions (PO wt.%:ps wt.%) (p < 0.05). The interaction coefficients b 12 have been found negative for all analytical responses (SFC, g and s), indicating a monotectic interaction between palm oil and palm stearin. Moreover, the high values of R 2 and R 2 (Adj) indicate a strong correlation and excellent regressions fit between variables (PO, PS and their admixtures) and analytical responses (SFC, viscosity and shear stress) (see Table 3). 4. Conclusion Temperature ( C) Coefficients a b 1 b b 2 c b 12 R 2 R 2 (Adj) Abbreviations: a, b and c, palm oil (PO), palm stearin (PS), and their admixtures respectively; b 1, b 2 and b 12, coefficient generated from PO, PS, and their interaction respectively; R 2 : coefficient of determination; R 2 (Adj): R square (Adjusted). This study has shown that many significant variations occurred on the physicochemical and rheological properties, due to the combination of low and high melting TAGs. The results obtained from GC indicate that the ratio of saturated/unsaturated fatty acids significantly increased at PS levels more than 40 wt.%, resulting in an undesirable level of palmitic acid ranging from ± 0.35% to ± 0.20%. HPLC data demonstrated partial stabilisation of POO/OPO content of 25.7 ± 0.3 to 24.1 ± 0.0% at PS levels less than 40 wt.%, where the excessive addition of PS more than 40 wt.% significantly increased the PPO/POP content from 35.0 ± 0.4% to 41.2 ± 0.0%. On the other hand, cooling thermograms showed two events on its crystallisation profile. The first peak was detected at high temperature levels. This peak was possibly associated to the solid fraction (stearin), while the second peak was detected at low temperature, possibly referred to the liquid fraction (olein). In addition, the peak has been found gradually shifted as the amount of unsaturated TAGs increased, whereas the levels of saturated TAGs increase the sharpness of the peak. In contrast the melting thermograms exhibited a new event at high melting temperature when the levels of PS exceeded 40 wt.% of the blend. In part, the pnmr results showed significant drop in SFC during tempering, resulting in low SFC ranging from 5 15% at body temperature (37 C) for blends containing PS levels less than 40 wt.%. Moreover, rheological properties results showed low values of viscosity (g) and shear stress (s) for all blends containing PS levels less than 40 wt.%. Consequently, the development of primary crystals within oil/fat blend system was dominated by high melting TAGs namely PPO/POP, which led to formation agglomerates at high palm stearin levels more than 40 wt.%. Therefore, the best levels of PS required from PO should be less than 40 wt.%, which made it desirable to be used for margarine production. Acknowledgements The authors would like to acknowledge the Director General of the Malaysian Palm Oil Board for his permission to the authors to conduct this study, as well as Ms. Puana Nasoi for her assistance. References AOCS (1989). Official methods and recommended practices of the American Oil Chemists Society (4th ed.). Champaign Illinois, USA. Aini, I. N., & Miskandar, M. S. (2007). Utilization of palm oil and palm products in shortenings and margarines. Lipid-Fett, 109(4), Andreia Schäfer De Martini Soares, F., Claro da Silva, R., Caroline Guimarães da Silva, K., Bertolessi Lourenço, M., Ferreira Soares, D., & Antonio Gioielli, L. (2009). Effects of chemical interesterification on physicochemical properties of blends of palm stearin and palm olein. Food Research International, 42(9), Ariffin, A. A., Bakar, J., Tan, C. P., Rahman, R. A., Karim, R., & Loi, C. C. (2009). Essential fatty acids of pitaya (dragon fruit) seed oil. Food Chemistry, 114(2), Campos, A. L., Kawachi, E. Y., Oliveira, T. C., & Thim, G. P. (2005). Mullite crystallization mechanism obtained from kinetic parameters determination for seeded and non-seeded gel. Materials Science & Engineering B, 122(3), Chrysam, M. M. (1996). Margarine and spreads. In Y. H. Yui (Ed.), Bailey s Industrial Oil and Fat Products (5th ed., pp ). Toronto: John Wiley and Sons. vol. 3. Che Man, Y. B., Haryati, T., Ghazali, H. M., & Asbi, B. A. (1999). Composition and thermal profile of crude palm oil and its products. Journal of the American Oil Chemists Society, 76(2), Chu, B. S., Ghazali, H. M., Lai, O. M., Che Man, Y. B., & Yusof, S. (2002). Physical and chemical properties of a lipase-transesterified palm stearin/palm kernel olein blend and its isopropanol-solid and high melting triacylglycerol fractions. Food Chemistry, 76(2), Danthine, S., & Deroanne, C. (2003). Physical and textural characteristics of hydrogenated low-erucic acid rapeseed oil and low-erucic acid rapeseed oil blends. Journal of the American Oil Chemists Society, 80(2), deman, L., dema, J. M., & Blackman, B. (1995). Effect of tempering on the texture and polymorphic behavior of margarine fats. Fat Science Technology, 97, Fredrick, E., Foubert, I., Van De Sype, J., & Dewettinck, K. (2008). Influence of monoglycerides on the crystallization behavior of palm oil. Crystal Growth & design, 8(6), Hare, L. B. (1974). Mixture designs applied to food formulation. Food Technology, 28, Lai, O. M., Ghazali, H. M., Cho, F., & Chong, C. L. (2000). Physical properties of lipasecatalyzed transesterified blends of palm stearin and anhydrous milk fat. Food Chemistry, 70(2),

10 612 S. Saadi et al. / Food Chemistry 132 (2012) Mamat, H., Nor Aini, I., Said, M., & Jamaludin, R. (2005). Physicochemical characteristics of palm oil and sunflower oil blends fractionated at different temperatures. Food Chemistry, 91(4), Marangoni, A. G. (2005). Crystallization kinetics. In A. G. Marangoni (Ed.), Fat crystal networks (pp ). New York: Marcel Dekker, Inc. Miskandar, M. S., Che Man, Y. B., Yusoff, M. S. A., & Abdul Rahman, R. (2002). Effect of scraped-surface tube cooler temperatures on the physical properties of palm oil margarine. Journal of the American Oil Chemists Society, 79(9), Mulder, H., & Walstra, P. (1974). The Milk Fat Globule Vol. 159, pp Wageningen, The Netherlands: Pudoc. Miura, S., & Konishi, H. (2001). Crystallization behavior of 1,3-dipalmitoyl-2-oleoylglycerol and 1-palmitoyl-2,3-dioleoylglycerol. European Journal of Lipid Science and Technology, 103, Minato, A., Ueno, S., Smith, K., Amemiya, Y., & Sato, K. (1997). Thermodynamic and kinetic study on phase behavior of binary mixtures of POP and PPO forming molecular compound systems. Journal of Physical Chemistry B, 101, Nor Aini, I., Abdullah, A., & Halim, A. H. (1992). Evaluation of palm oil quality: Correlating sensory with chemical analyses. Journal of the American Oil Chemists Society, 69, Nor Hayati, I., Che Man, Y. B., Tan, C. P., & Nor Aini, I. (2009). Droplet characterization and stability of soybean oil/palm kernel olein O/W emulsions with the presence of selected polysaccharides. Food Hydrocolloids, 23(2), Norizzah, A. R., Chong, C. L., Cheow, C. S., & Zaliha, O. (2004). Effects of chemical interesterification on physicochemical properties of palm stearin and palm kernel olein blends. Food Chemistry, 86(2), O Brien, R. D. (2004). Fats and oils: Formulating and processing for applications. CRC. Okawashi, T., Sagi, N., & Mori, H. (1985). Confectionary fats from palm oil. Journal of the American Oil Chemists Society, 62, PORIM. (1995). PORIM test methods. Malaysia: Ministry of Primary Industries, pp Siew, W. L., & Ng, W. L. (1999). Influence of diglycerides on crystallisation of palm oil. Journal of the Science of Food and Agriculture, 79(5), Silva, R. C., Cotting, L. N., Poltronieri, T. P., Balcão, V. M., de Almeida, D. B., Goncalves, L. A. G., et al. (2009). The effects of enzymatic interesterification on the physicalchemical properties of blends of lard and soybean oil. LWT-Food Science and Technology, 42(7), Tan, C. H., Ghazali, H. M., Kuntom, A., Tan, C. P., & Ariffin, A. A. (2009). Extraction and physicochemical properties of low free fatty acid crude palm oil. Food Chemistry, 113(2), Tan, C. P., & Che Man, Y. B. (2002). Differential scanning calorimetric analysis of palm oil, palm oil based products and coconut oil: Effects of scanning rate variation. Food Chemistry, 76(1), Tanaka, L., Miura, S., & Yoshioka, T. (2007). Formation of granular crystals in margarine with excess amount of palm oil. Journal of the American Oil Chemists Society, 84(5), Tanaka, L., Isogai, T., Miura, S., & Murakami, M. (2010). Effect of triacylglycerol species on the crystallizing behavior of a model water/oil emulsion. European Journal of Lipid Science and Technology, 112, Tanaka, L., K-Tanaka Yamato, S., Ueno, S., & Sato, K. (2009). Microbeam x-ray diffraction study of granular crystals formed in water-in-oil emulsion. Food Biophysics, 4, Watanabe, A., Tashima, I., Matsuzaki, N., Kurashige, J., & Sato, K. (1992). On the formation of granular crystals in fat blends containing palm oil. Journal of the American Oil Chemists Society, 69(11), Wiking, L., De Graef, V., Rasmussen, M., & Dewettinck, K. (2009). Relations between crystallisation mechanisms and microstructure of milk fat. International Dairy Journal, 19, Zaliha, O., Chong, C. L., Cheow, C. S., Norizzah, A. R., & Kellens, M. J. (2004). Crystallization properties of palm oil by dry fractionation. Food Chemistry, 86(2), Zhang, L., Ueno, S., Sato, K., & Miura, S. (2007). Binary phase behavior of 1,3- dipalmitoyl-2-oleoyl-sn-glycerol-glycerol and 1,2-dioleoyl-3-palmitoyl-racglycerol. Journal of the American Oil Chemists Society, 84,