Comparative Analysis of the Physico-Chemical, Thermal, and Oxidative Properties of Winged Bean and Soybean Oils

Abstract

To explore possible food applications, the oxidative stability, antioxidants contents (tocopherols and tocotrienols), thermal properties, and solid fat content of winged bean oil were investigated along with soybean oil for comparison. Results showed that winged bean oil was significantly (p < 0.05) resistant to oxidation (27 h) compared to soybean oil (9 h) heated at 110^oC for 32 h, due presumably to the presence of alpha tocotrienol and the high behemic acid content. The high content of tocopherols, 230 mg/100 g in soybean oil did not contribute much to its oxidation stability. At around 25^oC, winged bean oil contained about 15% solid fat content with <1% in soybean oil. Soybean oil, however, had better levels of the omega-3 and omega-6 fatty acids with a ratio of 7.6:1, falling within the range of 5:1 and 10:1 recommended by the Food and Agriculture Organization/World Health Organization, when compared to winged bean oil having a ratio of 34.4:1. The positional fatty acids esterified as stearoyl, palmitoyl, behemoyl or lignoceroyl triacylglycerides species, PSP, PPP, PLB, POB and SSlg+PBP overwhelmed the unsaturated FAs of winged bean oil for free radical reaction and also conferred it high thermal conductivity. The high oxidative stability, solid fat content, and thermal conductivity of winged bean oil coupled with its crystalization characteristics confirmed it to be good material for frying and for making zero-trans-fat margarines and spreads.

Keywords:

• Winged bean oil
• Solid fat content
• DSC
• Physicochemical properties
• Antioxidants

INTRODUCTION

Legumes are dicotyledonous seeds of plants consumed whole or dehulled in many parts of the world as cheap sources of edible lipids and proteins in comparison to other conventional crops.^[1] Because of their adaptability to different climatic conditions, they are found wild in some parts of the world as underutilized crops except for some such as soybean, cowpea, pea, and lentils. As a result, extensive researches into their compositions as sources of micro- and macronutrients, as novel food sources for mitigating malnutrition and as neutraceuticals have been conducted, with reports on their phytochemical, nutraceutical, and therapeutical properties,^[2,3] significant nutritional value,^[4,5] as a source of essential oils,^[6] and as sources of edible fats and oils^[7,8] are well documented.

Recently, fats and oils from non-conventional sources are gaining recognition for their role in the functional food and healthcare industries due to their therapeutic properties,^[9] frying properties, and for their unique phytochemical composition and antioxidant properties.^[10,11] As a result, lipids obtained from non-traditional sources such as Morienga oliefera,^[12,13] culinary melon seed (Cucumis melo var. acidulous),^[14] safflower,^[15] conophor,^[16] and pumpkin seeds^[17] have been reported for their important food use properties. Accordingly, when Morienga oliefera was compared with soybean oil (SBO) and palm olein for oxidative stability, Morienga oliefera oil showed lower p-AV and TOTOX values than either SBO or palm olein heated at 185^oC for 30 h,^[12] whereas culinary melon seed oil, in another study, showed a large amount of linoleic acid which is essential in synthesis of prostaglandin in addition to boasting the immune system.^[14] Others such as Sclerocarya birrea kernel oil and Sorghum bug (Agonoscelis pubescens) oil showed superior oxidative stabilities and frying qualities when investigated along with SBO at 17^oC for 24 h.^[18] Winged bean [Psophocarpus tetragonolobus (L.) DC,], also a non-conventional leguminous crop, has numerous nutritional and medicinal benefits^[7,19] and offers a good substitute to soybean because in addition to its nutritive similarity with the soybean, the crop is highly versatile as it grows where soybean fails to grow. The crop is, however, cultivated primarily for its immature edible pods that are cooked and eaten (sometimes raw) as a vegetable, whereas the young leaves and flowers are used in soups.^[19] The tubers (or roots) contain proteins more any other known root or tuber food plant.^[8] The fat content of winged bean seed ranged between 18 to 20%, resembling soybean,^[8,20] and the fatty acid (FA) composition of the its oil as previously reported, suggested it a potential oil for both the foods and the oleo-chemical industries.^[21,22] Recent studies have also placed winged bean oil (WBO) as a healthy fat-replacer for use in reduced-fat foods because of its content of behenic acid, which has been shown to have low absorption and consequently low adipose-fat deposit especially when the acid is esterified at the Sn-1, 3, positions.^[23,24] It is, therefore, essential that other important properties of the WBO that could enable its use in food be investigated. Physicochemical, thermal, and oxidative stability of an oil and its solidity relative to its constituent FAs are critical in fat or oil intended for frying application and for making fat-based products, such as margarines and spreads. Thus, an insight into the physicochemical and thermos-oxidative stability of WBO could provide a new platform for effective utilization of the oil. In this study, the physicochemical and thermal properties, oxidative stability, and solid fat content (SFC) relative to the tocols contents of winged bean (Psophocarpus tetragonolobus L. DC) oil were evaluated with reference to SBO for possible application in food processing.

MATERIALS AND METHODS

Chemicals

The α-, β-, δ-, and γ-tocopherols and tocotrienols standards were purchased from Sigma Aldrich (Supelco Park, Bellefonte, PA, USA), and Sigma Aldrich (Sigma-Aldrich Pte. Ltd., Singapore), respectively. Fatty acids methyl esters (FAMEs) standard (Supelco-37) was purchased from Supelco (Supelco Park, Bellefonte, PA, USA), whereas methanol, acetonitrile, 2-propanol, n-hexane, acetone, and ethanol were from Fisher (Pittsburgh, PA, USA). All other chemicals used were either of high-performance liquid chromatography (HPLC) or analytical grade obtained from Merck (Merck, Darmstadt, Germany).

Sample Preparation

Winged bean and soybean seeds were purchased from the Malaysian Agricultural Research and Development Institute (MARDI) and Giant Hypermarket, respectively, both in Serdang, Malaysia. The seeds were cleaned and dehulled manually, oven-dried (Memmert, UFB-500, Schwabach, Gemany) at 35^oC for 22 h, and ground using variable speed rotor mill (Pulveristte-14, Fritsch, GmbH, Germany). Milled flour was passed through a 500 µm test sieve (ASTM Grade) mounted onto an electromagnetic sieve shaker (Analysette 3, Fritsch, Germany) to yield 0.5 mm particle size flour.

Oil Extraction Process

Fat was extracted from 100 g lot of winged bean or soybean seed flour using 300 mL n-hexane in a 500 mL Schott bottle placed in a shaking water bath (SW 22 Julabo, Labortechnik, Germany) at 30^oC for 3 h, and the resulting slurry vacuum filtered using an aspirated Buchner funnel fitted with Whatman No. 2 filter paper. The n-hexane filtrate was evaporated under vacuum in a rotary evaporator (Buchi Rotavapor, R-210, Switzerland) at 50% of total speed and 45^oC to recover the oil. The extracted oils were further passed over anhydrous sodium sulphate and dried at 45^oC before storage in amber glass bottles at below –4^oC until analyzed.

Physical Properties

Determination of the refractive index (RI) was done based on Association of Official Analytical Chemists (AOAC) method Cc 7–25^[25] using Abbe’s refractometer (ATAGO, DR-A1, Tokyo, Japan). The color characteristics of the oil samples were measured using a HunterLab colorimeter (UltraScan Pro 1092, USA) equipped with a tri-stimulus coordinates (L* a* b*) software (EasyMatch QC, HunterLab Inc. USA), where L* (0 to 100) scale denotes color from dark (0–50) to bright (50–100); a* value denotes redness (+) to greenness (–); and b* value denotes yellowness (+) to blueness (–). Determinations were carried out at least three times and results reported as means and standard deviations.

Chemical properties

The saponification value (SV; mg/g; Cd 3-25, AOCS), iodine value (IV; g/100 g; Cd 1-25, AOCS), peroxide value (meq/kg; Cd 8-53, AOCS), and free fatty acids (% FFA; Ca 5a-40, AOCS) were all determined following AOCS (1998) method.^[26] All analyses were conducted in triplicates and results reported as means and standard deviations.

Analysis of α-, β-, δ-, γ-Tocopherols and Tocotrienols Contents

Sample preparation for HPLC

Vitamin E isomers were determined as described earlier^[27,28] with slight modification of the saponification procedure. Accordingly, 100 mg oil sample was each placed in a 20-mL screw-capped aluminium foil-wrapped test tube, followed by the addition of 0.1 g ascorbic acid, 5 mL ethanol, and 0.15 mL of 80% KOH under nitrogen and infrared light. The mixture was incubated at 80^oC for 15 min in a shaking water bath oscillating at 100 rpm. After saponification, the mixture was removed and immediately placed in an ice bath to cool, and thereafter, 5 mL each of water and n-hexane were added. The mixture was allowed to stand for 5 min and then centrifuged in a 20 mL centrifuge tube at 1500 × g for 1 min. The upper layer was transferred to another 20 mL screw-capped test-tube using a Pasteur pipette and extracted with 5 mL n-hexane three times each. The top layers after three extractions were pooled and evaporated using a rotary evaporator (Buchi Rotavapor, R-210, Switzerland) attached to a heating bath (B-491, Buchi, Switzerland). The residue after solvent evaporation was dissolved in 1.5 mL n-hexane, passed through a 25 µm cellulose filter and injected (20 µL) in a HPLC system.

Preparation of tocopherols and tocotrienols standards

Standards stock solutions were prepared separately using ethanol as solvent as described earlier.^[28] About 1.5 mg of each standard solution was diluted in 10 mL ethanol to make 150 μg/mL stock solution of the eight vitamin E isomers: α-, β-, δ-, γ-tocopherols and α-, β-, δ-, γ-tocotrienols, respectively, and stored at –20^oC. The concentration (µg/mL) of each tocotrienol standard stock solution was confirmed using ultraviolet-visible (UV/Vis) spectrophotometer (Perkin Lambda-25, Shelton, USA) from their known absorption coefficient [ρ = A × 10⁴] as follows: α-T(75.8) at 292 nm, β-T(89.4) at 296 nm, γ-T (91.4) at 298 nm, δ-T (87.3) at 298 nm, α-T3 (86.0) at 292 nm, β-T3 (86.2) at 296 nm, γ-T3 (91.0) at 297 nm, and δ-T3 (91.0) at 297 nm.^[27] Working standards were prepared from these stock solutions and diluted with absolute ethanol to obtain isomer concentrations of 4, 10, 20, 30, and 40 µg/mL. For the quantification, four standard curves were made using the reference tocopherols and tocotrienols isomers (α-, β-, δ-, γ). All the standard calibration curves showed good linearity (R²) in the range of concentrations studied, and the absolute contents of tocopherols and tocotrienols were quantitated by referring to the calibrated standard curves.

HPLC conditions

Chromatographic separation of the tocopherols and tocotrienols were carried out on an Agilent Infinity series 1200 high-performance liquid chromatograph (Agilent Technologies Inc., USA), equipped with a quaternary pump, an auto-sampler (SIL-20A), a degasser, and a diode-array spectrophotometric detector (DAD) in series with a fluorescence detector (RF-10AXL) at an excitation and an emission wavelength of 295 and 330 nm, respectively. Separation of the compounds was achieved at 25°C on a reverse-phase Cosmosil pi-NAP column (internal diameter 4.6 mm, length of 25 cm, particle size of 5.0 μm) at a maximum pressure of 400 bar, pH 9 and sample injection volume of 20 μL. The column was conditioned using the mobile phase for 30 min to achieve a linear baseline and run with the mobile phase [water-methanol-acetonitrile, 13:80:7 (v/v)] in an isocratic mode at 1 mL/min flow rate.Tocopherols and tocotrienols were eluted in about 30 min running time.

Determination of FAs as Methyl Esters

FAMEs were prepared by trans-esterification of the oil using sodium-methoxide complex as a catalyst, according to the AOCS method Ce 1-62.^[26] Using a Pasteur pipette, 100 mg (±0.5 mg) of oil was transferred into a screw-cap vial to which was added 5 mL of hexane and vortexed briefly. Sodium methoxide (250 µL) was then added and the mixture vortexed for 1 min followed by the addition of 5 mL saturated sodium chloride solution. The mixture was capped and shaken vigorously for 15 s and allowed to stand for 10 min. Using a Pasteur pipette, the top hexane layer was passed over anhydrous sodium sulphate (Na₂SO₄) granules for 20 min before injecting into a gas chromatograph (GC) system.

The GC system was an Agilent (Agilent 6890N, Network GC-System, Wilmington, USA) fitted with an auto sampler (Agilent 7683-Series), an injector (Agilent 7683-B Series), and flame ionization detector. Separation was carried out on a capillary column DB-Wax (length 30 m, internal diameter of 0.30 mm) and the carrier gas was helium at a flow rate of 1 mL/min. Sample was injected at 1 µL in a split ratio of 1:20 and flow rate of 2 the mL/min (at room temperature). The injector and detector temperatures were 250°C while the oven temperature initially at 100^oC for 2 min and then increased to 230^oC at a rate of 5^oC/min. The oven temperature was subsequently held at 230^oC for 10 min. Identification of FAMEs was achieved by comparing their retention times with those of known standards (37-Component FAME Mix, SUPELCO) and quantified as percent-weight of total FAs.

Triacylglyceride (TAG) Composition

The TAGs composition of the oils was determined as described previously^[29] using a reversed phase Jasco HPLC system (JASCO 1500 series, Japan) coupled with pump (PU-1558), detector (Jasco RI-1530), a degasser (Jasco DG-2080-54), and an oven (Jasco-2065). The TAGs were separated on a Purospher Star RP-18e column (250 mm in length × 4.6 mm internal diameter, 5 µm particle size, Merck-Millipore) and the column temperature was maintained at 45^oC. The mobile phase used was a mixture of acetone-acetonitrile [65:35 (v/v)] at a flow rate of 1 mL/min. Aliquots of 20 µL sample in HPLC grade acetone (0.1 g/1 mL) were injected into the HPLC system and TAG were eluted according to their chain length (ECN) and degree of unsaturation of the FAs in the glycerol moiety^[30] while peaks were identified according to their retention times relative to that of TAG standards (Sigma Chemical Co., USA).

Determination of Viscosity (ƞ) and Shear stress (τ)

The rheological properties of the oils were determined using a controlled-stress rheometer (Rheostress RS600, Haake, Germany), with a sand-blasted cone sensor (pp 35/2, Ti-35 mm diameter, 0.50 mm gap) and a measuring plate cover (MPC 35).^[29] Measuring conditions and instrument calibration were made automatically by the Rheowin software (Rheowin Job Manager Version 3.12) based on the measurement variables. Determination of the flow properties was conducted at shear rate in the range of 10–1000 s^–1 while the effect of temperature on the viscosity was investigated by subjecting the oils over temperatures of 20, 25, 30, 40, 50, 60, and 70^oC at constant shear rate of 100 s^–1. Temperature conditions were controlled by a temperature programmer (HAAKE RheoStress 6000, Thermo Scientific, Karlsruhe, Germany). Samples were equilibrated to 25^oC prior to measurement and the experiments run by gradually heating the samples to 70^oC and cooling back to 25^oC. All determinations were conducted in triplicates and results reported as means and standard deviations.

Differential Scanning Calorimetry (DSC) of Oil Samples

The thermal properties of the winged bean and SBOs were determined using a Mettler Toledo differential scanning calorimeter (Mettler Toledo Model 822e, Columbus, USA) equipped with thermal analysis data station software (Star Software Version 3.0 Mettler Toledo). The equipment was calibrated with indium and zinc prior to analysis. Nitrogen (99% purity) was used as the carrier gas at a flow rate of 100 mL/min. Sample weights between 15–20 mg were accurately weighed in 40 µL aluminium pans (ME-26763) and placed in the heating chamber of the DSC unit along with an empty pan as a reference both previously equilibrated to 20^oC. Analysis was run as follows: heating from –70 to 70^oC, scanned at 5^oC/min and held for 8 min; cooling: cooled from 70^oC to –70^oC, scanned at 5^oC/min and held for 8 min. The onset, denaturation, crystallization, and end set temperatures were computed automatically by the data analysis software (equipped with the equipment) after imputing the necessary variables. All determinations were conducted in triplicates and results reported as means and standard deviations.

Oxidative Stability Index (OSI)

The OSI was determined using a Metrohm Rancimat (Model 743, Herisau, Switzerland) following AOCS method Cd 12b-92^[26] To avoid any form of contamination, all glassware, electrodes and connecting tubes were rigorously cleaned before and between runs using detergent solution then rinsed with tap water, acetone, and distilled water, respectively, before oven-dried at 80°C prior to use. A stream of purified air (20 L/h) was passed through oil sample (5 g) heated at 110^oC and the effluent air containing volatile organic acids was bubbled in deionized water whose electrical conductivity increased as oxidation of the oil takes place with formation of peroxidases and hydroperoxidases. The stability test was run for 32 h and water conductivity measured continuously until the maximum point was reached. The oxidative stability of the oil was recorded as the point of inflexion of the tangents of the time and point of surge in oxidation (h).

Determination of SFC

The SFC of the oil samples was measured following the Malaysian Palm Oil Board (MPOB) Test Method^[31] using a Bruker NMS 120 Mini Spec pulsed nuclear magnetic resonance (pNMR) spectrometer (Karlsruhe, Germany) based on direct measurement of the ratio of solid and liquid component of the sample observed in the NMR free induction decay (FID) at different temperatures. Samples were prepared to produce 10-points melting curves and the SFC of the oil samples measured at each melting point. Oil samples were first placed in NMR tubes and melted at 80°C for 30 min, held at 70^oC for 30 min, and then chilled at 0°C for 90 min. Melting, holding, and chilling of the samples were carried out in a pre-equilibrated thermostatic water bath held at the measurement temperatures of 5, 10, 15, 20, 25, 30, 35, 37, 40, and 45^oC for 30 min.

Statistical Analysis

All determinations were carried out at least in triplicates and analysis of data was done using Minitab statistical software (Minitab Inc. Version 16). Means of triplicate determinations between samples were compared using two sample t-test at 5% level of significance. Results are presented as means ± standard deviations of multiple determinations.

RESULTS AND DISCUSSION

Physico-Chemical Properties

The physico-chemical properties of the WBO and SBO are shown in Table 1. WBO bean had significantly (p < 0.05) higher SV of 277.12 mg/g than SBO (225.56 mg/g) indicating that when employed as a soap base^[32] WBO could yield higher soap stock. SV is a measure of the amount of saponifiable material in the fat that will form soap on reaction with NaOH. The IV of WBO and SBO were 85.35 and 127.55, respectively, showing that more iodine were absorbed by the adjoining unsaturated bonds of SBO glyceride chains than was absorbed by WBO, indicating more unsaturation in the SBO. Though the IV of WBO seemed quite low given that it contained 70.4% unsaturated fatty acid (USFA), this could be due to the fact that the amount of monounsaturated fatty acids (38.4%) was higher than polyunsaturated fatty acids (PUFA; 32%), when compared to SBO in which the amount of PUFA (59.8%) was much higher than the amount of monounsaturated fatty acid (23.6%). Thus, the contribution of the PUFAs in WBO in absorbing iodine will likely be less, giving that the level of IV in WBO was proportionate to the total contributions of the USFAs and the PUFAs, as shown in Table 2. The peroxide values (PV) of both WBO and SBO showed that both oils contain low levels of peroxides and hydroperoxides as primary decomposition products of fat oxidation, an indication of absence of aldehydes, ketones, and secondary alkyls.^[33] These latter compounds, determined and expressed as p-ansidine value (p-AV), are secondary fat decomposition products that usually become noticeable only at advanced stage of fat deterioration, much later after the induction period, formed especially at frying conditions than when exposed to air alone.^[34] The higher IV of SBO increases its susceptibility to oxidation than do WBO, and that might explain why WBO had lower levels of both PV and FFA. FFA content of SBO was also higher indicating that it can easily be oxidized. The RI, also a measure of the unsaturation of oil, did not differ greatly between the WBO and the SBO, the latter showing higher value, again, indicating more unsaturation in SBO than in WBO. Color data for winged bean and SBOs as presented in Table 1 showed significant (p < 0.05) differences between color variables, L*, a*, and b* of both oils. SBO was significantly (p < 0.05) lighter (higher L*) in color and had more yellowness (higher b*) than WBO. In addition, WBO had significantly (p < 0.05) lower a* value (23.47) than SBO (33.30), and thus had less “redness” than SBO. The corresponding higher values of 83.31, 33.30, and 70.32 for L* (brightness), a* (redness), and b* (yellowness) for SBO, than for WBO 50.73, 23.47, and 40.79, respectively, gave SBO a bright, golden-yellow color and WBO a dull-yellow color.

TABLE 1 Physicochemical properties of winged bean and soybean oils

Property**		Winged bean oil	Soybean oil
Sap val (mg/g oil)		277.12 ± 1.72^a	225.56 ± 0.42^b
Iodine value (gI/100 g oil)		85.35 ± 0.70^a	127.55 ± 0.74^b
Peroxide value (g/kg)		1.81 ± 0.03^a	2.1 ± 0.11^b
Free fatty acid (g/100 g)		1.16 ± 0.00^a	1.82 ± 0.03^a
Refractive index (30^oC)		1.463 ± 0.00^a	1.459 ± 0.00^a
Color*	L	50.73 ± 0.71^a	83.31 ± 0.97^b
	a	23.47 ± 0.16^a	33.303 ± 0.30^b
	b	40.79 ± 1.14^a	70.32 ± 0.62^b
Oxidative stability index (32 h, 110^oC)		27.80 ± 0.35^a	9.17 ± 0.08^b
Alpha (α) tocopherol		nd	14.6 ± 1.23
Gamma (γ) tocopherol		158.7 ± 0.23^a	142.8 ± 0.21^a
Delta (δ) tocopherol		14.0 ± 0.56^a	72.6 ± 1.00^b
Alpha (α) tocotrienol		19.6 ± 0.65	nd*
Total tocopherols		172.70 ± 0.01^a	230.00 ± 0.00^b
Total tocotrienols		19.6	nd*

Values are means ± standard deviations of triplicate determinations;

Values in the same row with different superscript are significantly different at p < 0.05;

*L: dark (0–50) –light (51–100); a: red (+) – green (–); b: yellow (+) – blue (–);

**α-, γ-, and δ-tocopherols and tocotrienols measured in mg 100 g^–1 oil.

TABLE 2 Fatty acid compositions of winged bean and soybean oils

	Fatty acids composition (wt. %)
	Winged bean oil	Soybean oil
C14:0 (myristic acid)	0.00 ± 0.00	0.2 ± 0.02
C16:0 (palmitic acid)	7.3 ± 0.21^a	10.04 ± 0.23^b
C18:0 (stearic acid)	4.7 ± 1.32^a	4.7 ± 0.92^a
C18:1 (oleic acid)	34.0 ± 0.26^a	23.4 ± 0.51^b
C18:2 (linoleic acid)	31.0 ± 0.03^a	52.9 ± 0.20^b
C18:3 (linolenic acid)	0.9 ± 0.07^a	6.9 ± 1.21^b
C20:0 (arachidic acid)	1.3 ± 0.12^a	0.4 ± 0.23^a
C20:1 (eicosenoic acid)	3.5 ± 0.05^a	0.2 ± 0.01^b
C22:0 (behemic acid)	15.2 ± 0.21^a	0.3 ± 0.30^b
C22:1 (erusic acid)	0.9 ± 0.53	0.0 ± 0.00
C24:0 (lignoceric acid)	0.9 ± 0.20^a	0.2 ± 0.45^b
MONOUNSAT	38.4 ± 1.32^a	23.6 ± 0.70^b
POLYUNSAT	32.0 ± 0.20^a	59.8 ± 0.10^b
SFA	29.4 ± 0.50^a	16.0 ± 2.01^b
USFA	70.4 ± 0.41^a	83.4 ± 0.23^b
USFAs/SFAs	2.39 ± 0.24^a	5.21 ± 1.34^b

Values are means ± standard deviation of triplicate determinations;

Values in the same row with different superscript letters are significantly different at p < 0.05.

Tocopherol and Tocotrienol Contents

Three but differing vitamin E isomers were detected in each of the winged bean and SBOs as shown in Table 1. Alpha (α), Gamma (γ), and delta (δ) tocopherols were the main antioxidants detected in SBO at levels of 14.6 and 142.8 and 72.6 mg/100 g oil, respectively, making a total of 230 mg/100 g tocopherol. In WBO, α-tocotrienol along with both γ- and δ-tocopherols were detected at levels of 19.6, 158.7, and 14 mg/100 g oil, respectively. The high level of γ-tocopherol (158.7 mg/100 g) in WBO could also add to its oxidative stability since the isomer have high oxidation-reduction potential of all the tocopherols isomers, and that it can dimerized to compounds that still possess antioxidant activity,^[35] in addition to the α-tocotrienol content (19.6 mg/100 g) of WBO. The large amount of total tocopherols in SBO, however, did not contribute meaningfully in reducing the oxidation along its double bonds as seen in the low OSI in Table 1. This is because the different vitamin E isomers, as potent antioxidants in both in vivo and in vitro, have different biological activity and γ-tocopherol, which formed the bulk (142.8 mg/100 g) of the total tocopherol in SBO, has only about one-tenth the biological activity of α-tocopherol^[35] which constitute only 14.6 mg/100 g, representing 6.35% of the tocopherols in SBO. It will, therefore, be logical to say that the relative oxidative stability of SBO could be contribution from the antioxidant activity of α-tocopherol. Generally, however, tocotrienols have better antioxidant properties than tocopherols as shown from studies in model membrane,^[36,37] of which none (tocotrienol) was detected in the SBO (Fig. 1). This might also explains the high OSI of WBO and consequent long induction time, three-fold the induction period of SBO (Fig. 2). The values of the tocopherol homologues in SBO agree well with reported values of 12.66 to 72.37, 38 to 108, and 8.15 mg/100 g of α-tocopherol, δ-tocopherol, and β-tocopherol contents, respectively, in soybean seed and germ oils.^[38]

FIGURE 1 HPLC chromatogram of α-, β-, δ-, γ-tocopherols (T) and tocotrienols (T^–3) in (a) soybean oil; (b) winged bean oil; and (c) vitamin E isomers standards of tocopherols and tocotrienols.

FIGURE 2 Curves of shear stress vs shear rate of (a) winged bean oil and (b) soybean oil at 25^oC.

FA Composition

The FA compositions of WBO and SBO oils are presented in Table 2. The results showed that WBO had behemic acid (15.2%) as its major saturated FA but a negligible amount (<1%) in SBO. WBO contained approximately 70% unsaturated and 30%, saturated FA, respectively, with oleic (34.0%) and linoleic (31.0%) as the major unsaturated FAs, followed by small amounts of eicosenoic (3.5%), erusic (0.9%), and linolenic (0.9%) acids. The major saturated FA species were lignoceric (C24:0), behemic (C22:0), arachidic (C20:0), stearic (C18:0), and palmitic (C16:0) acids. The ratio of unsaturated to saturated FAs was about 2.4. In SBO, the major unsaturated FAs were linoleic (52.9%), oleic (23.4%), and linolenic (6.9%), whereas palmitic (10.4%) and stearic (4.7%) acids formed the major saturated FAs, Table 2. The ratio of unsaturated to saturated FAs in SBO was 5.2. The FA profile of SBO has been extensively studied, and one of such studies reported that the major FAs were linoleic (49.8–59.0%), oleic acid (17.7–28.0%), and palmitic (8.0–13.5%).^[39] Linoleic and linolenic acids together making up 32 and 60% of WBO and SBO, respectively, are essential components of omega-3 and omega-6 FAs which are indispensable for healthy tissue and cell functions. The low content of PUFA especially linolenic acid (<1%) and high amount of saturated FA (29.4%) might be responsible for the oxidative stability of the WBO compared to their respective contents, 7.0 and 16.0%, respectively, in SBO. SBO had higher amounts of both linoleic and linolenic acids^[40] than WBO, and also a good balance of the ratio of linoleic to α-linolenic acids of 7.6:1, falling within the range of 5:1 and 10:1 recommended by FAO/WHO^[41] compared to WBO with a ratio of 34.4:1. In a related study by Ayyildiz et al.,^[40] the FAs composition, tocols, oxidative stability, FFA, PV and IV of refined sunflower, soybean, corn, hazelnut, peanut, and canola oils, results showed the presence of tocopherols in all the oils, with SBO containing the highest total tocopherol levels, whereas α- and β-tocotrienols were detected in corn oil only, and that gave it a relatively higher oxidative stability (5 h/120^oC) with respect to the other oils whose oxidative stabilities ranged between 3.05 to 4.17 h at 120^oC.^[40]

TAG Composition

The TAG compositions of the WBO and SBO oils as shown in Table 3 indicated that both oils are composed of different TAG species and positional distribution. About 24 different TAGs were identified in WBO, some eluted as critical pair TAGs. The main TAGs in WBO include OLB (17.01%), LLB+OOE (16.13%), and OOL+LLE (9.60%), singlets like OOB (8.03%), OOO (6.76%), LLO (5.74%), and other minor components (≤5%). The mean total saturated TAGs (calculated as the sum of LLB+OOE, OLB, SSL+PBB, POB, OOB, PLB, OLB, OOS+LLB, PPP, SOL, OOP, PSL, POL, and LLP percentages) was 71.35%, whereas LLL+OLLn, LLO, OOL+LLE, OLE, and OOO formed 28.64% mean total unsaturated TAGs. This is based on the fact that the effect of the long-chain saturated FAs may overwhelm the contributions of the unsaturated FAs, especially where both sn-1 and sn-3 positions were esterified as stearoyl, palmitoyl, behemoyl, or lignoceroyl or any combination of two. This can be seen in TAG molecular species such as PSL, PPP, PLB, POB, and SSLg+PBB where, though, the relative percentages are small, their presence in a TAG may decrease the proportion of available double-bond for free radical reaction to occur, because the main reactants involved are the unsaturated FAs whether present as FFAs, triacylglycerols, diacyglycerols or monoacylglycerols or phospholipids.^[33] Comparatively, in SBO, LLL (17.98%), LLO (16.77%), LLP (13.33%), and one critical pair, LLS+POL (11.82%), were the major TAG species, with mean total unsaturated FA (sum of LnLnL, LLLn, LLL, OLLn, LLO, OOL, and OOO) of 58.91%. Thus, the high SFC of WBO could be due primarily to the saturated TAG species behemoyl-, lignoceroyl-, stearoyl-, and palmitoyl-glycerols, while the low SFC of SBO could be attributed chiefly to the unsaturated TAGS such as linoleoyl- and oleoyl-glycerol (Table 3) and to a lesser extent linolenoyl-glycerol. Solid fat content is one of the key properties of fats meant for making margarines, spreads, and other plastic fats used in bread and confectionery manufacture, usually obtained through partial hydrogenation of oils, which generates trans-FAs that are implicated in coronary heart diseases.^[42,43] The SFC of WBO placed it above other plants oils such as soybean, peanuts, safflower, and sunflower in terms of application in making fat-based products without the need for hydrogenation to achieve the right solid content.

TABLE 3 Triacylglyceride (TAG) compositions of winged bean and soybean seed oils

Triacylglyceride (TAG) composition (%)
Winged bean oil		Soybean oil
TAG	(%)	TAG	(%)
LLL+OLLn	2.69 ± 0.50	LnLnLn	0.80 ± 0.003
LLO	5.74 ± 0.37	LnLnL	1.62 ± 0.06
LLP	2.65 ± 0.14	LLLn	6.55 ± 0.04
OOL+LLE	9.60 ± 0.06	LLL	17.98 ± 0.41
POL	5.98 ± 0.04	OLLn	3.30 ± 0.24
OLE	3.95 ± 0.06	PLLn	2.74 ± 0.03
OOO	6.76 ± 0.06	LLO	16.77 ± 0.07
PSL	2.89 ± 0.05	LLP	13.33 ± 0.13
OOP	2.35 ± 0.05	PPLn	0.26 ± 0.01
SOL	0.84 ± 0.05	OOL	9.20 ± 0.05
PPP	0.56 ± 0.12	LLS+POL	11.82 ± 0.17
LLB+OOE	16.13 ± 0.06	PPL	1.98 ± 0.03
OOS+LLB	1.40 ± 0.06	OOO	3.49 ± 0.07
OLB	17.01 ± 0.02	SOL	4.04 ± 0.06
PLB	4.37 ± 0.66	POO	1.95 ± 0.05
OOLg	4.13 ± 0.36	PSL	1.44 ± 0.05
OOB	8.03 ± 0.37	PPO	0.51 ± 0.03
POB	2.46 ± 0.22	OOS	1.67 ± 0.03
SSLg+PBB	2.47 ± 0.14	POS	0.56 ± 0.02

Values are means ± standard deviations of triplicate determinations;

TAG composition was calculated from peak area of TAG divided by the total area of all TAGs present in the chromatogram;

TAG species: L: linoleoyl; Ln: linolenoyl; O: oleoyl; P: palmitoyl; E: elaidoyl; S: stearoyl; B: behemoyl; Lg: lignoceroyl.

Viscosity and Shear Stress

The rheological properties of the legume oils over a shear rate of 1-1000 s^–1 showed that both WBO and SBO exhibited time-independent Newtonian flow behavior within the temperature range of investigation (25–70^oC). The viscosity data for the oils as a function of shear stress versus shear rate is shown in Fig. 3. Vegetable oils are Newtonian fluids, having a linear relationship between shear rate and shear stress. This relation is in agreement with Newton’s law of viscosity, as given by the following equation:

(1)

FIGURE 3 Flow behaviors of (a) winged bean oil and (b) soybean oil at constant shear rate (100 s^–1) as a function of temperature.

where, σ is the shear stress (Pa), ý is shear rate (s⁻¹), and µ is the viscosity (Pa.s). The viscosity at each temperature point can be obtained from Eq. (1) using the slope from fits of the experimental shear stress–shear rate data. It can be seen that at a given temperature, WBO has higher viscosity compared to SBO. Expectedly, the viscosities of both oils exhibited exponential decreases with increases in temperature as shown in Fig 4.

FIGURE 4 Thermograms of (a) melting and (b) cooling of winged bean and soybean oils.

DSC of Oil Samples

Melting properties

The DSC melting thermograms Fig. 5a of both WBO and SBO were obtained after heating them from –70 to 70^oC at 5^oC/min. In WBO, two major curves and a smaller one were identified with variable temperature peaks. The first melting peak in the TAGs melting range started at –28.50 ^oC, peaked at –23.11^oC and offset at –14.83^oC. This peak could be due to the low melting point (LMP) such as linoleic acid (–12^oC) constituting up to 31% of the FAs (Table 2) esterified as tri-linoleoyl-glycerol (LLL), di-linoleoyl-oleoyl-glycerol, di-linoleoyl-palmitoyl-glycerol and other LMPs critical pairs such as LLL+OLLn and OOL+LLE (Table 3). The nearly sharp melting peak indicate the presence of TAGs including linoleoyl-glycerol esters, the major FA in that region, with enthalpy of fusion of 376.30 MJ. The integral of the curve shows the eutectic nature of the TAGs that together require a higher energy to become liquid. The second melting curve started at 5.36^oC and peaked at 11.28^oC, represents complex mixtures of medium melting point (MMP) triacylglycerol species such as trioleoyl-glycerol, dioleoyl-lignoceroyl-glycerol (the sharp point), and other oleic acid (m. p. 12.82^oC) esterified glycerols in the concave shoulder.^[44] The high enthalpy of fusion of 452.04 MJ indicates the presence of mixtures of medium to high melting points (HMPs) TAG.^[29] The HMP TAGs with adjoining glycerol esters such as stearic, palmitoyl, behemoyl, and lignoceroyl could be responsible for the broad but concave nature of the third peak with low enthalpy of –47.97 MJ (Table 4), probably due to the total contributions of behemic (15.2%), lignoceric (0.9%), and stearic (4.7%) acids (Table 2). Increased chain length, unsaturation, and branching of fat or oil molecules have been linearly correlated with increased enthalpy.^[44,45] The presence of some eutectic impurities probably gave rise to the shouldering on the curve,^[46] which seemed to have crystallized out during the later cooling stage leaving only the main components in the observed peak, Fig 4a. In both cases, glass transitions were not noticeable to be quantified and accounted for. The heat of fusion and melting point obtained from the curve depicted the nature of the mixed TAGs contained in the oil and indicate high thermal conductivity of the WBO. Probably because the liquid phase of the WBO consisted of a number of individual droplets, the degree of supercooling of each droplet was different so that several concave-shaped peaks are observed with melting points at maxima. Oils and fats exhibits complex thermal crystal behavior dictated by their chemical compositions and the protocol for the DSC experiment.^[47] The complexity of the TAGs in WBO is indicative of the melting curve exhibited by the thermograms (Fig. 4a), and consequently the HMP attributed partly to the oleic (34.0 %) and linoleic (31.0 %) acids content having melting points of 16 and –5^oC, respectively. The thermal behavior of SBO, however, showed quite a distinct melting characteristics from WBO. Four melting peaks were observed during the melting regime between –45.10 and –1.20^oC (Fig. 4a). The first melting peak started (T_onset) at –45.10^oC, peaked (T_peak) at –36.12^oC, and offset (T_end-set) at –30.5^oC, probably due to very LMP TAG species esterified mainly by linolenic (6.9 %) acid, and characterized by the small but near-concave shaped curve. The second melting curve had melting T_onset, T_peak, and T_end-set at –30.60, –27.36, and –20.93^oC, respectively, with resultant enthalpy of 158.18 MJ and constituted the majority of the TAGs in the oil with close melting temperatures. The FA with melting points between –11.50 and –12.20^oC including linoleic acid (53%) esterified as trilinoleoyl-glycerol, dilinoleoyl-linolenoyl-glycerol, and dilinolenoyl-palmitoyl-glycerol were the major TAGs contributing to the broad but near-sharp nature of the third curve (Fig. 4a). This peak had T_onset, (T_peak), and T_end-set temperatures of –20.45, –18.57, and –6.92^oC, respectively. The last peak in the melting profile characterized the HMPs TAGs in the SBO ranging between –11.22^oC (T_onset) and –3.61^oC (T_end-set), with (T_peak) at –6.55^oC. This melting curve had a sharp peak signifying the higher amount of high melting FAs such as oleic (23.4%) and some stearic (4.7%) acids, with adjoining shoulder that characterized palmitic acid, Tables 2 and 3. The higher melting points and enthalpy of fusion of WBO TAGs could be due to the crystalline packing of the TAGs compared to that of SBO TAGs, as they increased with increases in chain length, saturation, and symmetry, and are higher for trans than for cis isomers^[48,49] This finding corroborated earlier study on SBO^[47] in which three melting peaks, –38, –25 and –8^oC and a crystallization peak at –65^oC were observed. The relatively high enthalpy of fusion of WBO, Table 4, compared to SBO could impart it higher thermal conductivity when employed in frying operations as foods could cook faster when the same amout of heat is applied as exemplified in the DSC, thus reducing cost of energy.

TABLE 4 Crystallization and melting properties of winged bean and soybean oils

Property	Winged bean oil	Soybean oil
Heating**
Melting range (^oC)	–28.50 to 22.1	–45.10 to –1.20
Melting time (min)	11.3	9.80
Melting rate (^oC/min)	4.51	4.44
Enthalpy of fusion (MJ)	–915.65	–266.49
First peak (^oC)	–23.11 (–376.33)	–36.12 (–43.99)
Second peak (^oC)	11.28 (–452.07)	–27.36 (–158.18)
Third peak	19.46 (–47.97)	–18.57 (–39.42)
Fourth peak	—	–6.55 (–24.90)
Cooling**
Crystallization range (^oC)	–54.34 to –67.21	–9.91 to –55.20
Crystallization time (min)	15.10	9.21
Crystallization rate ^oC/min)	6.30	5.00
Enthalpy (MJ)	216.02	160.29
Onset (^oC)	—	–9.91
First peak (^oC)	–62.54 (216.02)	–12.70 (71.03)
Second peak (^oC)	—	–38.55 (89.26)

*Values are means ± standard deviation of triplicate determinations;

**Values in parenthesis are enthalpies obtained by integration of respective curves.

FIGURE 5 Oxidative stability indices of winged bean (WBO) and soybean (SBO) oils determined by intersections of the tangents of stability and induction curves.

Crystallization properties

On cooling from 70^oC at 5^oC/min to –70^oC, the crystallization curve of the WBO showed slight undulations (Fig. 4b). Crystallization of WBO from the molten state took place between –54.34 and –67.21^oC within 15 min of cooling, yielding a total enthalpy of 216.02 MJ with the only peak at –62.54^oC. On further cooling, mesophase transitions occurred first and no two peaks were observed as it was during the heating regime.^[50] This could be caused by diglycerides, FFAs, phospholipids, sterols, and other minor constituents in addition to the TAGs constituents with HMPs characteristic of eutectic mixture. This could be the reason why no crystallization peak occurred during the later cooling stage of the WBO. WBO crystallized faster and with narrow crystallization range (–54.34 to –67.21^oC) within the DSC measurement temperatures than SBO (–9.91 to –55.20^oC), as shown in Fig. 4b. Further crystallization through –54.34 to –67.21°C, the fat molecules exhibited supercooling with the formation of crystals in the last phase of the cooling at maximum peak of –62.54^oC. The glass transition (T_g) stage was similarly not observed during the cooling stages of the oils. The high enthalpy of WBO could have resulted from the high content of saturated FAs (52.5%) attributed mainly to behemoyl-glycerol with larger peak area in the DSC curve. The cooling profile of SBO showed two crystallization curves exhibiting different behaviors with no observable phase transition. SBO crystallized between –9.91 and –55.20^oC, a larger temperature range than that of WBO but lower enthalpy of 160.29 MJ, Table 4. The first cooling curve started at –9.91^oC (T_onset), reached peak at –12.70^oC to yield enthalpy of crystallization of 71.03 MJ. The second cooling curve had the highest peak at T_peak –38.55^oC with T_onset and T_endset of –29.55 and –48.43^oC, respectively. The broad peak exhibited by the second curve at peak maxima of –38.55^oC could be caused by the different TAGs plus di- and monoacylglycerides and other minor components.^[50] In the liquid state, TAGs are orderly oriented due to self-organization of the liquid molecules into structural entities that gave them their lamellar structures.^[48,49] In SBO cooling, the very low crystallization temperature could be due to linoleic (53%) and linolenic (7%) acids, with melting points of –5 and –11^oC, respectively. Knothe and Dunn^[44] earlier showed that crystal’s characteristics of a fat determines the primary rheological behavior of the fat, its crystal network and solidification properties in food applications. The crystallization characteristics exhibited by WBO are desirable in making margarines and spreads where fast crystallization facilitates formation of minute crystal polymorphs with even distribution that could yield smooth and consistent texture of resulting fat product.

OSI (Rancimat Measurement)

The relative oxidative stabilities of the oils (Fig. 5) showed that WBO had higher stability to oxidation than SBO determined by extrapolation of the baseline and the tangent of the peak of induction. On subjecting the oils to accelerated oxidation test at 110^oC for 32 h, thermal oxidation of the SBO begun within the first 10 h, whereas WBO remained stable until after about 27 h before oxidation set in with formation of decomposition products such as peroxidses, organic acids, and hydroperoxidases and some minor components decomposed from the primary compounds. This was indicated by the increased conductivity of the deionized water into which the oxidized volatile organic compounds were bubbled. Under the conditions of the test, WBO did not produced any lipid oxidized product until after 27 h at which time oxidation started to be noticeable. The low oxidative stability of SBO could be due primarily to the high levels of USFA especially linolenic linoleic and acid, and probably its low level of tocotrienols, compared to their low and high levels, respectively, in the WBO, as can be seen in Tables 1 and 2. Studies showed that under frying conditions, the stability of oil is not merely a function of the FA composition, but that other unsaponifiable matter such as tocopherol esters, sterols, and phytosterols do contribute to the oxidative stability of oil.^[51] On the other hand, there are also other lipid components that have pro-oxidants property and whose presence could have opposite effects. In practical applications, however, the choice of oil for food application is usually based on the FA compositions and oxidative and thermal stability, the first initial consideration,^[40] since a lot of oil (up to 40%) is taken up by food being fried, resulting in relatively high amounts of oil in fried foods.^[52] Where oil prone to oxidation is employed in food frying, oxidation and polymerization could occur too rapidly yielding both undesirable volatile products, non-volatile oxidized derivatives and dimeric, polymeric and other compounds with immediate or potential health hazards.^[53–55] WBO oil with its extended stability against oxidation relative to SBO could be a better frying fat and fried products could remain edible much longer than products fried in SBO where about 30% polar compounds accumulate after 18 h frying at 180^oC.^[52] Although the specific effects of α-tocotrienols in both WBO and SBO has not been established in this study, the absence of any of the tocotrienol in the SBO and its low oleic and high linoleic acids content could increase its susceptibility to oxidation compared to WBO because the tocotrienols have better antioxidants activity and thus, imparts more oxidative stability than the tocopherols.^[35,37] Further, Ayyildiz et al.^[40] observed that the oxidative stability of an oil depends not solely on the antioxidants content of the oil but on cummulative effect of the tocols profile, FAC, FFA, PV IV, and some minor components with antioxidative effects.

SFC

The SFC denotes the percentage of the components of lipids that are solid at certain temperatures. Determined using pNMR spectroscopy, the SFC of the WBO was significantly higher than that of SBO. As shown in Fig. 6, the SFC of WBO was about 35% at 5^oC, 30% at 10^oC and at room temperature of 25^oC, the SFC was 19%. For SBO, however, the SFC of the remained below 1% throughout the measuring conditions, an indication that SBO cannot be employed without major modification such as hydrogantion with consequent risk of trans-fats accumulation. The high SFC of WBO are contributions due mainly to behemic acid since it constitute about 15% of the total weight of the WBO, whereas for SBO, with less than 1% behemic acid, the two monosaturated acids, palmitic (10%) and stearic (4.7%), did not contribute much in the solidity of the oil. Molecular composition such as longer chain FAs will impart higher melting and crystallization temperatures than shorter chain FAs^[56] as exhibited by the WBO. SFC is a property critical in plastic fats, and in the manufacture of margarines, spreads, chocolates, and similar products. Fats intended for chocolate making must be substantially solid such that it will not melt in the hand temperature at 30°C, but melt all the more quickly in the mouth at around body temperature of 36°C. SFC, TAG composition and the polymorphic behavior exhibited by fat crystals greatly affect the texture of the fats and its products.^[45,57] With its high SFC of 14 at 25^oC, WBO can be modified at less cost to yield the right solid fat level for fat-based products such as margarine, spreads, and similar products without the risk of trans-fat generated during hydrogenation of oil for making those products. The WBO can also be fractionated into liquid and solid fractions for use as frying medium and for making margarines, respectively.

FIGURE 6 Solid fat content of winged bean (WBO) and soybean (SBO) oils determined by pulsed nuclear magnetic resonance (NMR) spectroscopy.

CONCLUSION

In the present study, WBO proved valuable as frying medium because of its high thermal conductivity; a desirable property for cooking food, in addition to its high oxidative stability, higher than SBO. In food frying operations, temperatures between 150 and 190^oC usually employed repeatedly accelerate lipid breakdown into compounds that may shorten product shelf life and/or present possible toxicity to man. Employing WBO could minimize the formation of polar compounds that are consequent of frying and, in addition, could cook food faster than other oils such as SBO. The high SFC of the WBO present another opportunity for use in the manufacture of trans-fat free fat-based products such as margarine, shortenings, spreads, and similar products. Trans FAs, implicated in coronary heart diseases, are generated during hydrogenation of oil to obtain solid fat. Use of WBO could rule out incidence of trans fats in foods. However, further investigating into the practical food frying application of the WBO and its use in producing fat-based products is suggested.

NOMENCLATURE

P	=	Palmitic
S	=	Stearic
Lg	=	Lignoceric
O	=	Oleic acid
B	=	Behemic acid
FAO	=	Food and agriculture organization
WHO	=	World health organization

ACKNOWLEDGMENTS

The author Makeri thanked the Malaysian government for the award of Malaysian International Scholarship (MIS). The assistance of Dr Kharima Ahmad and staff of the Oil and Fats Technology Laboratory, MPOB, Malaysia, in conducting the oxidative stability test is gratefully acknowledged.

SUPPLEMENTAL MATERIAL

Supplemental data for this article can be accessed on the publisher’s website.