Journal of Oleo Science Copyright ©2012 by Japan Oil Chemists’ Society J. Oleo Sci. 61, (12) 671-679 (2012)

Characterization of Physicochemical and Thermal Properties and Crystallization Behavior of Krabok ( Malayana) and Rambutan Sopark Sonwai＊ and Punnee Ponprachanuvut Department of Food Technology, Faculty of Engineering and Industrial Technology, Silpakorn University, Nakhonpathom, 73000, Thailand

Abstract: composition, physicochemical and thermal properties and crystallization behavior of fats extracted from the of krabok () and rambutan (Nephelium lappaceum L.) grown in Thailand were studied and compared with cocoa butter (CB). The krabok seed , KSF, consisted of 46.9% lauric and 40.3% myristic acids. It exhibited the highest saponification value and slip melting point but the lowest iodine values. The three fats displayed different crystallization behavior at 25℃. KSF crystallized into a mixture of β’ and pseudo-β’ structures with a one-step crystallization curve and high solid fat content (SFC). The fat showed simple DSC crystallization and melting thermograms with one distinct peak. The rambutan seed fat, RSF, consisted of 42.5% arachidic and 33.1% oleic acids. Its crystallization behavior was more similar to CB than KSF, displaying a two-step crystallization curve with SFC lower than that of KSF. RSF solidified into a mixture of β’ and pseudo-β’ before transforming to β after 24 h. The large spherulitic microstructures were observed in both KSF and RSF. According to these results, the Thai KSF and RSF exhibited physicochemical, thermal characteristics and crystallization behavior that could be suitable for specific applications in several areas of the food, cosmetic and pharmaceutical industries.

Key words: Irvingia Malayana, krabok, rambutan, cocoa butter, fat, solid fat content, crystallization

1 INTRODUCTION kijang’, which means‘ of the stags’5）. The most in- Vegetable fats and oils are extensively used as raw mate- teresting part of the krabok seems to be its fat, which rial and inputs for industrial food, medicine, and cosmetic is the most abundant component of kernels. Only a few production1）. The possibility of an acute shortage in edible studies have been done on the compositions and properties and industrial vegetable fats and oils together with a clear of krabok seed fat（KSF）and it was reported that KSF is tendency favoring the use of natural vegetable fats for the rich in lauric and myristic acids6）, indicating its potential replacement of those from hydrogenated vegetable oil has uses as confectionery and cosmetic fat. The nutritional encouraged the investigation of edible as possible value of KSF was evaluated by Laohawinit5）. Bandelier et sources of usable fats and oils1, 2）. Krabok（Irvingia Malay- al.7） studied the biochemical and oil composition of the ana）is a large-grown and woody wild widely KSF obtained from Cambodia and found that it contained distributed in tropical and subtropical areas. It is from the more than 83％ of lauric and myristic acids. However, due Irvingia family, which consists of two species. The other to the lack of technical information regarding to its proper- species is Ivingia gabonensis or African wild mango, ties, especially thermal properties and crystallization be- which has been studied extensively. It is a commercial and havior, and potential use, KSF has so far been under uti- indigenous fruit tree of West and Central , which has lized for oil production. been identified as the most important tree for domestica- The rambutan（Nephelium lappaceum L.）is a medium- tion3）. The fruits are used as condiment and are highly sized evergreen tropical tree growing to a height of 12-20 valued for their food thickening properties. In Thailand, m. It is believed to be native to the Malay Archipelago, the krabok tree is commonly used for and charcoal from where it spread to Thailand, Burma, Sri Lanka, India, production, whereas the seeds, after peeling, are consumed Vietnam, Philippines and Indonesia. The fruit produced by by people4）. In Cambodia, the tree is known as the‘ pauh the tree, also called rambutan, is generally consumed fresh.

＊Correspondence to: Sopark Sonwai, Department of Food Technology Faculty of Engineering and Industrial Technology Silpakorn University 6 Rajmakkanai Road Nakhonpathom, 73000 Thailand E-mail: [email protected], [email protected] Accepted June 27, 2012 (recieved for review May 24, 2012) Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 online http://www.jstage.jst.go.jp/browse/jos/ http://mc.manusriptcentral.com/jjocs

671 S. Sonwai and P. Ponprachanuvut

However, in the main producing countries like Malaysia mercially and were of the highest purity available. and Thailand, which has become the leading producer（0.6- 0.7 million tons a year）and exporter（～10-15 million 2.2 Crude fat extraction and puri cation dollars US）of the fruit8）, rambutan fruit is industrially pro- Krabok seed fat（KSF）and rambutan seed fat（RSF）were cessed to obtain juices, jams, jellies, marmalades, etc9）. extracted from the dried and finely ground krabok and There are well over 200 varities of rambutan available rumbutan seed kernels, respectively, using the Soxhlet ex- throughout tropical Asia. In Thailand, Rongrien and traction method at 60℃ for 6 h with n-Hexane as a solvent. Chompu are the most popular varieties, both of which have The crude fats were purified employing the method de- crisp arils and are well suited for fresh consumption and scribed by Solís-Fuentes and Durán-de-Bazúa12）. The puri- canning. After the process of canning, the residues consist fied fats were kept away from light and air at 4℃ until mainly of seeds and peels. Although, the seeds are consid- further analysis. ered as agro-industrial waste in many countries, but in others, like some Asian countries, they are edible after 2.3 Characterization of fatty acid composition roasting1）. Some studies on the seed have showed that KSF, RSF and CB were converted into fatty acid methyl rambutan possesses a relatively high amount of fat esters using AOAC official method 969.3313）. The fatty acid （between 17％ and 39％）that can be used for manufactur- methyl esters analysis was performed in a Shimadzu GC ing candles, soaps, and fuels9, 10）. In one of the most recent with flame ionization detector（GC-FID）. The system had studies, Solís-Fuentes et al.1）revealed that rambutan seed an VertiBondTM wax capillary column（50 m long, 0.25 mm fat obtained from RI-104 variety grown in Mexico com- internal diameter and 0.20 mm film thickness）. Compound prised mainly of oleic and arachidic acids and exhibited po- identification was carried out using external standards of tential use for different branches of industries from confec- fatty acids methyl esters. Helium was used as a carrier gas tionery to cosmetics. Sirisompong et al.11）reported that, with a flow rate of 1 mL/min and with a controlled initial

with high level of arachidic acid and low iodine value, the pressure of 93.2 kPa at 120℃. N2 and air were makeup chemical characteristics as well as the physical properties gases. The injection temperature was 210℃, and the oven of a fat obtained from a Thai variety of rambutan compared temperature program was holding at 120℃ for 3 min well with those of conventional fats. This would permit the before increasing at a rate of 10℃/min to 220℃, holding at use of the fat, especially where oxidation may be a concern, this temperature for 30 min, increasing at a rate of 5℃/min without its being subjected to hydrogenation. to 240℃, followed by holding at 240℃ for 30 min. The split In this study, fatty acid composition, physicochemical ratio was 100:1, the injection volume was 1 μl, and the de- properties and crystallization behavior of the fats extracted tector temperature was 280℃. After the fats were ana- from the krabok and rambutan seeds from trees grown in lyzed, their chromatograms were acquired and the fatty Thailand were investigated for the purpose of evaluating acid contents were calculated based on percentage of peak the potentiality of these fruit by-products as sources of area. natural edible fats with possible industrial use. 2.4 Characterization of physicochemical properties Iodine value（Iv）was analyzed using automatic titrator （Mettler Toledo DL58 titrator）. Saponification value and 2 EXPERIMENTAL PROCEDURES slip melting point were analyzed following Palm Oil Re- 2.1 Materials search Institute of Malaysia Test Method no. p3.1（1995） Mature krabok seeds were collected from the northern and no. p4.2（1995）, respectively14）. Acid value was ana- region of Thailand, and rambutan seeds from Rongrien lyzed following American Oil Chemists’ Society Official variety were kindly supplied by Malee Sampran Public Method Ca 5a-4015）. Company Limited（Nakhonpathom, Thailand. The seeds were cut open and the seed kernels were removed. The al- 2.5 Characterization of solid fat content mond-like decorticated seed kernels were crushed by a Changes in the solid fat content as a function of temper- hand mill and dried in a vacuum oven for 4 h at 60℃. Then, ature between 15℃ and 40℃ and the melting behavior of the dried kernels were finely ground and stored in polyeth- the fats were determined by pulse-nuclear magnetic reso- ylene bags at 4℃ until extraction. Cocoa butter（CB）was nance（p-NMR）spectrometer（Minispec-mq20, BRUKER, purchased from Sino-Pacific Trading（Thailand）Co., Ltd. Karlsruhe, Germany）following a method for measuring The standard fatty acid methyl esters for fatty acid analysis solid fat content of CB and similar fats developed by the using gas chromatography（GC）were purchased from Ac- “Joint Committee for the Analysis of Fats, Oils, Fatty Prod- cuStandard, Inc.（USA）. Acetone and acetonitrile were of ucts, Related Products and Raw Materials（GA FETT）” as GC grade from Burdick and Jackson（Muskegon, MI, USA）. described by Fiebig and Luttke16）. All other chemicals and solvents used were obtained com-

672 J. Oleo Sci. 61, (12) 671-679 (2012) Characterization of Krabok and Rambutan Seed Fats

2.6 Characterization of crystallization and melting pro les 0.02 °2θ, respectively. The crystallization and melting profiles of the fat samples were determined with a Perkin-Elmer differential scanning 2.8 Crystal morphology calorimeter（DSC）（ model DSC 8000, PerkinElmer Co., The crystal morphology of the KSF, RSF and CB crystal- Norwalk, CT）following a procedure employed by Solís- lized under static conditions at 25℃ was observed by po- Fuentes et al.1）. The heat flow of the instrument was cali- larized light microscopy（PLM）（ Olympus BX51, Olympus brated with indium（mp 156.6℃）as a reference standard. A Optical Co., Ltd., Tokyo, Japan）equipped with a digital fat sample of 3-5 mg was placed in an aluminum pan（20 μL camera a digital camera（Olympus C-7070, Olympus Optical capacity）and hermetically sealed. An empty pan served as Co., Ltd., Tokyo, Japan）. The fat samples were melted at reference. Prior to the analysis of thermal behavior, the 80℃ for 10 min. A small droplet（about 20 μL）of the melted samples were heated from 20℃ to 80℃ at 30℃/min and fats was placed on a glass slide preheated to 80℃. A pre- held for 10 min in order to ensure homogeneity and to heated glass cover slip was carefully placed over the destroy any crystal memory. Then, the samples were sample to produce a film of uniform thickness. The slides cooled to －40℃ at the rate of 5℃/min. The crystallization were incubated for 24 h at 25℃ in a temperature-con- profile, enthalpy of crystallization, and the phase changes trolled cabinet. A 10× lens was employed to image the between onset and offset temperatures were recorded. gray scale photographs of the fat crystals. This was followed by heating the sample at the rate of 5℃/ min from －40℃ to 80℃ during which the melting profile, 2.9 Statistical analysis the enthalpy of melting, and the temperatures of the phase The obtained data was analyzed by Analysis of Variance changes were recorded. The crystallization onset（TCO）and with Least Significant Difference（ANOVA/LSD）at 95％ melting completion temperatures（TMC）were obtained from confidence interval. the peaks located at the highest temperature of each fat sample. TCO and TMC were considered to be the tempera- tures at which the crystallization began and the melting ended, respectively17）. 3 RESULTS AND DISCUSSION 3.1 Fatty acid composition analysis 2.7 Crystallization behavior under static isothermal con- Table 1 exhibits the fatty acid contents of KSF, RSF dition compared with CB, which was high in stearic（35.9％）, The crystallization behavior of KSF and RSF under static oleic（32.4％）and palmitic（25.3％）acids. KSF consisted conditions at 25℃ was investigated using p-NMR and x-ray more than 90％ of saturated fatty acids. It contained diffraction（XRD）techniques. For the p-NMR study, the fat mainly （46.7％）and （40.3％）. The samples, which were contained in p-NMR tubes, were amount of lauric acid in the Thai KSF reported here heated to 80℃ for 10 min in a water bath and then were matched the lauric acid content in and palm transferred to a cooling bath set below 25℃. Once the kernel oils18, 19）, which are two most important lauric fats temperature of the samples inside the tubes decreased to for the confectionery industry. Also present are palmitic, 25.5℃, the tubes were removed from the cooling bath, stearic, oleic, and behenic acids. This gives KSF a hard wiped dry and rapidly put into the p-NMR sample port with consistency with high solid fat content and high melting the temperature set at 25℃. The timing then started and point as shown in the following sections. The combined the SFC was recorded once per min for 60 min, allowing amount of lauric and myristic acids found in the Thai KSF in-situ observation of the crystallization of the fat samples. （87％）was higher than that of the Cambodian KSF（83％） Since 25℃ was too high for CB to solidify enough to reported by Bandelier et al.7）but lower than that of the achieve good SFC reading during the p-NMR measurement, African wild mango（92％）published by Njoku and Ug- the study of the crystallization behavior of CB using the wuanyi2）. p-NMR was omitted here. Two main fatty acids in RSF were oleic（33.1％）and ara- For the XRD study, each fat sample was put inside an chidic（42.5％）acids. Lauric, myristic, palmitic, stearic and XRD capillary sample holder using a pipette. The sample behenic acids were also present. The contents of oleic and was then heated to 80℃ for 10 min by dipping in a water arachidic acids in the Thai RSF from Rongrien variety re- bath, after which it was transferred to a temperature-con- ported here were significantly lower and higher, respec- trolled cabinet set at 25℃ and the timing started immedi- tively, than the contents of the two fatty acids in the ately. The XRD characterization was performed at 0, 1.5, Mexican RSF from RI-104 variety as reported by Solís- 24 h and 7 days using an x-ray diffractometer（Rigaku Fuentes et al.1）. The high content of arachidic acid, a fatty TTRAX III, Rigaku Corporation, Tokyo, Japan）. Scans were acid with a long chain and a relatively high melting point, performed in wide angle x-ray scattering from 15 °2θ to 25 of the Thai RSF would have allowed it to exhibit a harder °2θ with a scan speed and a step width of 2.7 °2θ/min and consistency with higher SFC than the Mexican variety as

673 J. Oleo Sci. 61, (12) 671-679 (2012) S. Sonwai and P. Ponprachanuvut

Table 1 Fatty acid compositions of krabok seed fat (KSF), rambutan seed fat (RSF) and cocoa butter (CB). All data are mean values±standard deviations of triplicate measurements. a-cValues with the same letter in each row are not significantly different (p＞0.05). Fatty acid content（area％） Fatty acid type CB KSF RSF a b Lauric acid（C12） － 46.907±0.831 2.476±0.019 c a b Myristic acid（C14） 0.083±0.004 40.281±0.790 2.152±0.042 a b b （C16） 25.269±0.785 3.182±0.091 3.389±0.109 a c b （C18） 35.872±0.463 1.621±0.028 8.701±0.112 b c a （C18:1） 32.439±0.161 2.260±0.238 33.131±0.575 a c b Linoleic acid（C18:2） 2.406±0.003 0.016±0.012 0.025±0.006 a b Linolenic acid（C18:3） 0.123±0.001 0.036±0.005 － b c a Arachidic acid（C20） 0.890±0.051 0.012±0.001 42.506±0.604 c b a Behenic acid（C22） 0.135±0.001 1.675±0.069 2.628±0.078 a c b Lignoceric acid（C24） 0.368±0.007 0.012±0.002 0.036±0.007 Others 2.414±0.104c 3.996±0.084b 4.955±0.188a

Table 2 Physicochemical properties of krabok seed fat (KSF), rambutan seed fat (RSF) and cocoa butter (CB). a-cValues with the same letter in each row are not significantly dif- ferent (p＞0.05). Physicochemical properties CB KSF RSF Acid value（％ as oleic acid） 2.78±0.028a 0.28±0.002c 0.77±0.066b Iodine value（g I2/g fat） 34.03±0.467a 9.26±0.215b 32.31±0.771a Saponification value（mg KOH/g fat） 190.70±0.534b 223.86±0.253a 199.38±0.972b Slip melting point（℃） 31.03±0.058c 38.70±0.000a 38.47±0.058b

can be seen later. At 32.31 g I2/g fat, the Iv of the Thai RSF reported here was

significantly lower than that of the Mexican variety（47 g I2/ 3.2 Physicochemical properties g fat）1）. The difference in the rambutan cultivars and the The Soxhlet extraction of the crushed and dehydrated geographical locations where the rambutan trees were krabok and rambutan gave white fats that exhibited planted could have been responsible for the difference of a soft to the touch consistency. Table 2 shows the acid the values. value, the iodine value, the saponification value and the Saponification value（Spv）is a measurement of the slip melting point of KSF, RSF and CB. CB exhibited the average molecular weight（or chain length）of all the fatty highest acid value（2.78％ as oleic acid）followed by RSF acids present in a fat. Spv increases when molecular weight （0.77％ as oleic acid）and KSF（0.28％ as oleic acid）. The decreases. KSF exhibited the highest Spv（223.86 mg KOH/ iodine value（Iv）is useful in determining the degree of un- g fat）and this was most likely due to a high content of me- saturation and hardness of fats. The higher the Iv, the more dium-chain fatty acids（lauric and myristic acids）in the fat unsaturated fatty acid bonds present in a fat. From the as discussed earlier. The value was close to the Spv range

table, KSF exhibited the lowest Iv（9.26 g I2/g fat）whereas of African wild mango（212.7 -219.2 mg KOH/g fat）reported 20） the Iv of CB and RSF（34.03 and 32.31 g I2/g fat, respective- by Joseph . The Spv of RSF from the Thai rambutans ly）were much higher than that of KSF and not significantly （199.38 mg KOH/g fat）and of CB（190.70 mg KOH/g fat） different from each other（p＞0.05）. This indicates that were much lower than that of KSF due to the their high KSF contained much lower amount of unsaturation than contents of long-chain fatty acids such as oleic and ara- CB and RSF. The Iv of KSF given here was well within the chidic acids for RSF and stearic and oleic acids for CB. The

Iv range（3.5-19 g I2/g fat）of African wild mango grown in Spv of RSF reported in this work was relatively close to the as reported by Joseph20）and Njoku and Ugwuanyi2）. Spv of the Mexican RSF（186 mg KOH/g fat）given in the lit-

674 J. Oleo Sci. 61, (12) 671-679 (2012) Characterization of Krabok and Rambutan Seed Fats

erature1）. ture, indicating that it can resist high temperature, and a Slip melting point（SMP）of a fat is the temperature at sharp melting characteristic between 30 and 40℃. This which a column of fat in an open capillary tube softens or was in agreement with the way the SFC of the Cambodian becomes sufficiently fluid to slip or run up the tube when it KSF decreased within the same temperature range as re- is subjected to a controlled heating. SMP of KSF（38.7℃） ported by Bandelier et al.7）. However, the Cambodian KSF and RSF（38.5℃）were higher than CB（31.0℃）. The less melted completely at around 35℃ whereas the complete content of unsaturated fatty acids in KSF and higher melting of Thai KSF occurred just above 40℃. Though not content of saturated, long-chain fatty acid（arachidic acid） as high as KSF’s, SFC of RSF was much higher than CB at in RS when compared with CB（see Table 1）must have room temperature（～30℃）. At a body temperature, the contributed to the higher SMP of the two former fats. SFS of both KSF and RSF were still high above 20％. This would make them unattractive for consumption if they 3.3 Characterization of solid fat content were to be used as a sole fat ingredient in a confectionery Solid fat content（SFC）measured at a given temperature product due to the waxy mouthfeel they could generate. represents the percentage of solid fat crystallized at that Hence, the fats’ triacylglycerol composition/structure particular temperature. A plot of SFC versus temperature should be modified by one or more of various modification also provides information on the melting behavior of fats. methods such as blending, fractionation and chemical/en- Figure 1 shows SFC of KSF, RSF and CB measured at dif- zymatic interesterification, etc. before use. ferent temperatures between 15-40℃. It can be seen from the figure that SFC of KSF was the highest at all tempera- 3.4 Analysis of crystallization and melting pro les tures. The low content of unsaturated fatty acids contained The DSC thermograms for crystallization of KSF, RSF in the fat（Table 1）was the key contribution for its high and CB are given in Fig. 2（a）and the temperature of phase SFC. SFC of CB was the lowest at temperatures above 15℃ transition points for the fat samples were summarized in due to its high content of unsaturated fatty acids（～35％） Table 3. The crystallization of KSF showed a simple solidi-

whereas SFC of RSF, which was high in both unsaturated fication profile with only one sharp peak at 14.23℃（ T1） fatty acids（～33％）and long-chain saturated fatty acids whilst the crystallization thermogram of RSF and CB ex- （especially arachidic acid）was in between those of the hibited 3-4 peaks that appeared non-distinct, indicating former two fats. that heterogeneous types of triglycerides were crystallized At temperature range 15-30℃, SFC of KSF remained at different temperatures. The crystallization onset tem-

relatively unchanged as the temperature increased while perature, TCO, was highest for RSF（28.13℃）, possibly a SFC of CB dropped dramatically by 98.7％ and SFC of RSF result of its high arachidic acid content, followed by KSF decreased slowly by 34.8％. The fast drop in SFC of CB （24.61℃）and CB（19.10℃）. The crystallization enthalpies with temperature and its low SFC as the temperature ap- in a decreasing order were 111.1, 83.27 and 72.81 J/g for proaches a body temperature are well-known characteris- KSF, RSF and CB, respectively. RSF had a major thermal

tics that allow chocolate, which contains more than 30％ of transition at 26.83℃（ T1）and minor transitions at 16.6℃

CB, to melt quickly with a pronounced cooling effect when （T2）, 6.47℃（ T3）and 0.95℃（ T4）. CB exhibited a broad

consumed. KSF demonstrated high SFC at room tempera- main peak at 14.58℃（ T2）, which was located at a lower temperature than that of RSF, with a shoulder on the high-

temperature side at 17.76℃（ T1）. TCO for RSF reported here was lower than that of the Mexican RSF published by Solís-Fuentes et al.1）. In addition, a sharp peak represent- ing a thermal transition at a very low temperature of －18.1℃, indicating a group of triacylglycerols with an abundance of unsaturated fatty acids, recorded by Solís- Fuentes et al.1）was not observed here. This might have been the result of the difference in the cultivars of rambu- tan seeds used in the two studies and the geographical areas where the rambutan trees were planted. The DSC thermograms for melting of the fat samples are given in Fig. 2（b）. The melting completion temperatures,

TMC, for KSF（42.22℃）and RSF（41.55℃）were almost the

same and were much higher than TMC of CB（29.64℃）. KSF

Fig. 1 Solid fat content measured at 15-40℃ of krabok had only one sharp melting peak at 39.88℃（ T1）, suggest- seed fat (KSF), rambutan seed fat (RSF) and co- ing that the fat contained a group of triacyglycerols with coa butter (CB). similar phase behavior. In addition, the sharp DSC melting

675 J. Oleo Sci. 61, (12) 671-679 (2012) S. Sonwai and P. Ponprachanuvut

profile where the melting completed at 42.2℃ reflected what was observed earlier with the SFC study when the SFC of KSF decreased rapidly at the temperatures above 30℃ and came down to almost zero once the temperature reached just above 40℃（ Fig. 1）. Both RSF and CB showed three or four overlapped peaks with shoulders on both sides. The peaks corresponded to the melting of different groups of triacylglycerols that had difference fatty acid compositions. The melting enthalpies in a decreasing order were 131.66, 97.98 and 87.71 J/g for KSF, RSF and CB, re- spectively.

3.5 Static crystallization behavior Figure 3 displays crystallization curves of KSF and RSF which followed changes in SFC of the fat samples during

Fig. 2 DSC crystallization thermograms (a) and melt- ing thermograms (b) of krabok seed fat (KSF), rambutan seed fat (RSF) and cocoa butter (CB).

TCO and TMC indicate crystallization onset tem- Fig. 3 Crystallization curves for krabok seed fat (KSF) perature and melting completion temperature, and rambutan seed fat (RSF) obtained during

respectively. T1, T2, T3 and T4 indicate transition static isothermal crystallization at 25℃ for 60 temperature points as shown in Table 3. min.

Table 3 Temperature of phase transition points (T) and crystallization and melting enthalpies (ΔH) for krabok seed fat (KSF), rambutan seed fat

(RSF) and cocoa butter (CB). TCO and TMC represent crystallization onset and melting completion temperatures, respectively. 1, 2, 3 and 4 are points of phase transition, based on Figure 2 for DSC thermo- grams of crystallization and melting. Transition point temperature（℃） Sample ΔH（J/g） TCO T1 T2 T3 T4 TMC Crystallization KSF 111.61 24.61 14.23 RSF 83.27 28.13 26.83 16.60 6.47 0.95 CB 72.81 19.10 17.76 14.58 2.73 Melting KSF 131.66 39.88 42.22 RSF 97.98 9.42 29.40 37.16 39.42 41.55 CB 87.71 17.49 21.09 22.35 29.64

676 J. Oleo Sci. 61, (12) 671-679 (2012) Characterization of Krabok and Rambutan Seed Fats

static isothermal crystallization at 25℃ for 60 min. Both samples started to solidify before the crystallization tem- perature was reached with RSF demonstrating a higher SFC at the starting point of the crystallization time. The crystallization of KSF then developed rapidly with its SFC surpassing that of RSF within 1 min and the crystallization of the fat continued until the SFC attained a plateau at ～ 45 min, suggesting that the equilibrium had been reached. It can be seen from the figure that the crystallization of KSF was a one-step process. This was possibly due to the fact that KSF was comprised of a group of triglycerides with an abundance of saturated fatty acids that began to crystallize at approximately the same time. On the con- trary, the crystallization of RSF, which contained more complex mixtures of triglycerides, was clearly divided into two parts with different slopes, suggesting that the crystal- lization of the fat was a two-step process. The two-step crystallization curves have been observed before with veg- etable fats21, 22）. The change in the slope of the crystalliza- tion curve could be related to either the formation of dif- ferent polymorphic form of the fats, or the crystallization of different fractions22）. The SFC of RSF reached a plateau at around 40 min with an SFC value that was ～35％ lower than that of KSF. Fig. 4 Changes in polymorphic structure of krabok Changes in the polymorphic structure of KSF, RSF and seed fat (a), rambutan seed fat (b) and cocoa CB during static isothermal crystallization at 25℃ for 7 butter (c) obtained during static isothermal crys- days was demonstrated in Fig. 4. KSF began to crystallize tallization at 25℃ for 7 days. before the temperature decreased to 25℃（ Fig. 4a）. At t＝ 0, the fat crystallized into a mixture of polymorphs with dominating diffraction peak at 4.60 Å, a typical diffraction four diffraction peaks located at 3.84（1）, 4.04（2）, 4.24（3） pattern of β structure27）. This suggested that polymorphic and 4.41（4）Å. The peaks at 3.84 and 4.24 Å corresponded transition had taken place between 1 and 7 days of the to β’ structure23）. The 4.04 Å peak was close to peak range crystallization time. The fact that the polymorphic struc- of β’（ 3.70-4.00 Å）reported by Ghotra et al.24）. The peak at ture of RSF remained unchanged during the first 1.5 h 4.41 Å matched one of the diffraction peaks of pseudo-β’ implies that the two-step crystallization curve observed structure reported by D’Souza et al.25）and O’Brien26）. earlier with RSF during the SFC study（Fig. 3）was more During 7 days of crystallization, the diffraction profiles of likely related to the crystallization of different fractions KSF did not display any significant change with the four than the formation of different polymorphic forms of the diffraction peaks remained, indicating that the fat’s struc- fat. ture was a mixture of β’ and pseudo-β’ throughout the CB did not crystallized immediately as the temperature crystallization time. descended to 25℃（ Fig. 4c）. At 1.5 h of crystallization, the RSF also began solidification before the crystallization fat had solidified into form IV in β’ structure with two dif- temperature was reached（Fig. 4b）. The fat crystallized fraction peaks at 4.15（5）and 4.35 Å（6）29）. At 24 h, the fat into a mixture of polymorphs with diffraction peaks at 3.85 structure was still in β’ structure. After 7 days, CB had （1）, 4.15（2）, 4.24（3）and 4.55 Å（4）. The diffraction transformed into β structure with four typical diffraction pattern remained the same for at least 24 h. The peaks at peaks at 3.67（1）, 3.75（2）, 3.87（3）, 3.98（4）and 4.58 Å（7）, 3.85 and 4.24 Å corresponded to β’ structure23）while the corresponding to a stable form V of the fat29）. 4.15 Å peak could be related to either α structure27）or It can be seen here that the crystallization behavior of pseudo-β’ structure25）. However, considering that fats the three fats were quite different from one another. usually crystallize into α structure at a low temperature However, the crystallization behavior of RSF was more before transforming to a more stable polymorph, it was similar to CB than KSF. Both RSF and CB exhibited poly- more likely that the diffraction peak at 4.15 Å of RSF crys- morphic transition during crystallization and after 7 days tallized at 25℃ for 24 h corresponded to pseudo-β’ struc- had transformed into β structure. The phase transforma- ture. The diffraction peak at 4.55 Å corresponded to β tion did not occur with KSF during the experiment and structure28）. At 7 days, the diffraction profile showed one after 7 days the fat was mostly still in β’ structure, indicat-

677 J. Oleo Sci. 61, (12) 671-679 (2012) S. Sonwai and P. Ponprachanuvut

ing that β’ was probably the most stable structure of the crostructure of CB crystals was loosely-packed spherulites fat. consisting of needle-like crystals radiating and branching outward from central nuclei（Fig. 5a）. The crystal size of 3.6 Morphology study the fat ranged from 15 to 60 μm. Some degree of crystal The microstructure of fat crystal networks has a great aggregation, presumably by van der Waals forces31）, is effect on macroscopic properties30）. The PLM micrographs evident in the figure. The microstructure of KSF crystals of the fat samples obtained after static crystallization at was densely-packed spherulites with near-perfect spheru- 25℃ for 48 h are given in Fig. 5. The solid phase appears litic shape and smooth appearance that strongly aggregated white or gray while the liquid phase appears black. The mi- to one another（Fig. 5b）. The crystal size of KSF was ap- pearently much larger than that of CB. The largest crystal in the image was as large as ～375 μm. Finally, the micro- structure of RSF crystals was very similar to KSF but with a slightly smaller average crystal size（Fig. 5c）. A small extent of crystal aggregation had also been observed.

4 CONCLUSION From the results obtained in this study, KSF exhibited the highest saponification value and slip melting point but the lowest iodine values. KSF, RSF and CB displayed differ- ent crystallization behavior at 25℃. KSF displayed a one- step crystallization curve and simple DSC crystallization and melting thermograms. It started melting rapidly around 30℃ and completed just after 40℃. The crystallization be- havior of RSF was more similar to CB than KSF with a two- step crystallization curve. The amount of lauric acid in KSF situated well within the range of lauric acid content in coconut and palm kernel oils. This would give it potential to be used as a starting material for manufacturing lauric- fat cocoa butter replacer. According to these results, the Thai KSF and RSF have physicochemical and thermal characteristics and crystallization behavior that may be useful for specific applications in several segments of in- dustry, from confectionery to cosmetic and pharmaceuti- cal.

REFERENCES 1） Solís-Fuentes, J. A.; Camey-Ortíz, G.; Hernández-Me- del, M. R.; Pérez-Mendoza, F.; Durán-de-Bazúa, M. C. Composition, phase behavior and thermal stability of natural edible fat from rambutan（Nephelium lap- paceum L.）seed. Bioresource Technol. 101, 799-803 （2010）. 2） Njoku , O. U.; Ugwuanyi, J. O. Nutritional and toxico- logical properties of Dika fat（）. J. Herbs Spices Med. Plants. 4, 53-58（2010）. 3） Nangue, T. J.; Hilaire, M. W.; Felicite T. M.; Jacques F.; Fig. 5 Polarized light microscope images showing Linder M. Irvingia gabonensis fat: nutritional proper- crystal morphology of cocoa butter (a), krabok ties and effect of increasing amounts on the growth seed fat (b) and rambutan seed fat (c) obtained and lipid metabolism of young rats wistar sp. Lipids after static crystallization at 25℃ for 48 h. Health Dis. 10, 43-52（2011）.

678 J. Oleo Sci. 61, (12) 671-679 (2012) Characterization of Krabok and Rambutan Seed Fats

4） Wongsuthavas, S.; Yuangklang, C.; Wittayakun, S.; Va- oils-determination by pulsed nuclear magnetic reso- supen, K.; Mitchaothai, J.; Srenanual, P.; Beynen A. C. nance spectroscopy［C-IV 3g（2003）］. Eur. J. Lipid Dietary oil, but not krabok oil, diminishes ab- Sci. Tech. 105, 377-380（2003）. dominal fat deposition in broiler chickens. Int. J. 17） Kim, B. H.; Lumor, S. E.; Akoh C. C. Trans-free marga- Poultry Sci. 6, 792-795（2007）. rines prepared with canola oil/palm stearin/palm ker- 5） Laohawinit, S. Nutritional evaluation of Irvingia nel oil-based structured lipids. J. Agri. Food Chem. malayana seed. Ph. D. Thesis, Chulalongkorn Univer- 56, 8195-8205（2008）. sty, Bangkok, Thailand, p. 478-479（1989）. 18） Codex Standard 126, Edible Palm Kernel Oil（1981）. 6） Peangpra, P. Extraction and Analysis of fat from 19） Codex Standard 124, Edible Coconut Oil（1984）. Kabok. M. Sc. Thesis（Chemistry Education）Faculty of 20） Joseph, J. K. Physico-chemical attributes of wild Science, Chaingmai University, Chaingmai, Thailand, mango（Irvingia gabonensis）seeds. Bioresource p. 792-795（1977）. Technol. 53, 179-181（1995）. 7） Bandelier, J.; Chunhieng, T.; Olle, M.; Montet, D. Origi- 21） Marangoni A. G.; Mcgauley S. E. Relationship between nal Study of the Biochemical and Oil Composition of crystallization behavior and structure in cocoa butter. the Cambodia Nut Irvingia malayana. J. Agri. Food Cryst. Growth Des. 3, 95-108（2003）. Chem. 50, 1478-1482（2002）. 22） Foubert, I.; Dewettinck, K.; Janssen, G.; Vanrolleghem, 8） Salakpetch, S. Rambutan production in Thailand. In P. A. Modelling two-step isothermal fat crystallization. Proceedings of the inter- national tropical fruit J. Food Eng. 75, 551-559（2006）. conference. Florida: Hawaii Tropical Fruit Growers, 23） Czerniak, A. S.; Karlovits, G.; Lach, M.; Szlyk, E. X-ray pp. 13-17（2000）. diffraction and differential scanning calorimetry stud- 9） Morton, J. F. Rambutan. In: Morton, J. F.（Ed.）, Fruits ies of β’→ β transitions in fat mixtures. Food Chem. of Warm Climates. Julia F. Morton, Miami, FL., pp. 92, 133-141（2005）. 262-265（1987）. 24） Ghotra, B. S.; Dyal, S. D.; Narine, S. S. Lipid shorten- 10） Zee, F. Rambutan and pili nuts: potential crops for ings: a review. Food Res. Int. 35, 1015-1048（2002）. Hawai. In: Janick, J., Simon, J. E.（Eds.）, New Crops. 25） D’souza, V.; Deman, J. M.; Deman, L. Short Spacings J. Wiley and Sons Inc., New York, pp. 461-465（1993）. and Polymorphic Forms of Natural and Commercial 11） Sirisompong, W.; Jirapakkul, W.; Klinkesorn, U. Re- Solid Fats-a Review. J. Am. Oil Chem. Soc. 67, 835- sponse surface optimization and characteristics of 843（1990）. rambutan（Nephelium lappaceum L.）kernel fat by 26） O’Brien, R. D. Fats and Oils. Technomic Pub. Co., hexane extraction. LWT Food Sci. Technol. 44, 1946- Lancaster, PA（USA）, pp. 266-269（1998）. 1951（2011）. 27） Larsson, K. Classification of Glyceride Crystal Forms. 12） Solís-Fuentes, J. A.; Durán-de-Bazúa, M. C. Mango ACTA Chem. Scand. 20, 2255-2260（1966）. seed uses: thermal behaviour of mango seed almond 28） Ahmadi, L.; Marangoni, A. G. Functionality and physi- fat and its mixtures with cocoa butter. Bioresource cal properties of interesterified high oleic shoetening Technol. 92, 71-78（2004）. structured with stearic acid. Food Chem. 117, 668- 13） AOAC. Official Methods of Analysis of AOAC Inter- 673（2009）. national. AOAC International, Arlington, VA（1995）. 29） Wille, R. L.; Lutton, E. S. Polymorphism of Cocoa But- 14） PORIM. Palm Oil Research Institute of Malaysia ter. J. Am. Oil Chem. Soc. 43, 491-496（1966）. （PORIM）test methods. Kuala Lumpur. Malaysia 30） De Man, J. M.; Beers, A. M. Fat crystal networks: （1995）. Structure and rheological properties. J. Texture Stud. 15） AOCS. Official Methods and Recommended Practic- 18, 303-318（1987）. es of the American Oil Chemists’ Society. 6th ed. 31） Berger, K. G.; Jewell, G. G.; Pollitt, R. J. M. Oils and American Oil Chemists’ Society. Champaign, USA Fats. In Vaughan JG（ed.）Food Microscopy. Academic （1997）. Press, London, pp. 445-497（1979）. 16） Fiebig, H. J.; Luttke J. Solid fat content in fats and

679 J. Oleo Sci. 61, (12) 671-679 (2012)