- Email: [email protected]
Differential scanning calorimetry study—Assessing the influence of composition of vegetable oils on oxidation
Food Chemistry 194 (2016) 601–607
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Differential scanning calorimetry study—Assessing the influence of composition of vegetable oils on oxidation Baokun Qi 1, Qiaozhi Zhang 1, Xiaonan Sui, Zhongjiang Wang, Yang Li ⇑, Lianzhou Jiang ⇑ College of Food Science, Northeast Agricultural University, Harbin 150030, China National Research Center of Soybean Engineering and Technology, Harbin 150030, China
a r t i c l e
i n f o
Article history: Received 9 February 2015 Received in revised form 30 July 2015 Accepted 31 July 2015 Available online 6 August 2015 Keywords: Differential scanning calorimetry Vegetable oil Fatty acid Triacylglycerol Lipid oxidation Kinetic analysis
a b s t r a c t The thermal oxidation of eight different vegetable oils was studied using differential scanning calorimetry (DSC) under non-isothermal conditions at five different heating rates (5, 7.5, 10, 12.5, and 15 °C/min), in a temperature range of 100–400 °C. For all oils, the activation energy (Ea ) values at T p were smaller than that at T s and T on . Among all the oils, refined palm oil (RPO) exhibited the highest Ea values, 126.06 kJ/mol at T s , 134.7 kJ/mol at T on , and 91.88 kJ/mol at T p . The Ea and reaction rate constant (k) values at T s , T on , and T p were further correlated with oil compositions (fatty acids and triacylglycerols) using Pearson correlation analysis. The rate constant (k) and Ea of all oils exhibited varying correlations with FAs and TAGs, indicating that the thermal oxidation behaviors were affected by oil compositions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Vegetable oils play an important role in the human diet due to their nutritional and sensory properties. They are also one of the main ingredients in a wide range of food products, such as ice cream, sausages, and margarine (Shahidi & Zhong, 2010). However, oils are susceptible to oxidation reactions during food processing and subsequent storage of food products. The oxidation of oils leads to the development of off-flavor compounds and undesirable chemical changes of foods, resulting in decreased nutritional value of foods and potential food safety problems (Choe & Min, 2006; Dijkstra & Segrers, 2007). Differential scanning calorimetry (DSC), which belongs to the family of thermal analysis, has been applied for the last 50 years in the field of fats and oils for various purposes (Tan & Nehdi, 2014). In particular, DSC evaluates the thermal oxidation behaviors of oils through precise recording of heat flow into and out of an oil sample. The heat flow is reported as a function of either time or temperature and plotted on the DSC thermograms, in which each peak is associated with a specific physical or chemical process (Vecchio Ciprioti & Chiavaro, 2014). As a very well-established ⇑ Corresponding authors at: College of Food Science, Northeast Agricultural University, Harbin 150030, China. E-mail addresses: [email protected] (Y. Li), [email protected] (L. Jiang). 1 The first and second author contributed equally to this work. http://dx.doi.org/10.1016/j.foodchem.2015.07.148 0308-8146/Ó 2015 Elsevier Ltd. All rights reserved.
method, DSC has been widely applied in lipid studies. Cerretani et al. (2012) investigated the effect of fatty acid composition and phenol contents on the stability of extra virgin olive oil under accelerated oxidative test using DSC. Caponio et al. (2013) applied DSC in the evaluation of cooling and heating curves as well as thermal attributes of olive oil during refining process. Cerretani, Maggio, Barnaba, Toschi, and Chiavaro (2011) quantified the fatty acid in olive oils using a novel DSC-PLS (partial least square) method. Maggio, Barnaba, Cerretani, Paciulli, and Chiavaro (2014) developed a faster analytical method based on DSC, named highpressure liquid chromatography–differential scanning calorimetry–partial least square (HPLC–DSC–PLS), to investigate the DSC cooling curves of extra virgin olive oil affected by triacylglycerol composition. Oil oxidation is a complex reaction involving several simultaneous reactions. The different oxidation behaviors among oils may be affected by the various oil compositions. To the best of our knowledge, only one study conducted by Vecchio, Cerretani, Bendini, and Chiavaro (2009) investigated the relationship between oil composition and thermal oxidation properties. They reported that the onset temperatures of the thermal decomposition transition, the maximum heat flow temperatures, and as well as the sum of enthalpy were found to significantly correlate with the components of twelve different extra virgin olive oils. However, no study was found to address the relationship between oil composition and
602
B. Qi et al. / Food Chemistry 194 (2016) 601–607
thermal oxidation properties for other vegetable oils, such as soybean oil, corn oil, and sunflower oil. Therefore, the aim of this study was to evaluate the thermal oxidation properties of vegetable oils using DSC, and further investigate the relationships between oil compositions and thermal oxidation properties. 2. Materials and methods 2.1. Materials Eight different vegetable oils, namely refined palm oil (RPO), olive oil (OeO), grapeseed oil (GsO), sunflower oil (SuO), corn oil (CnO), soybean oil (SoO), safflower oil (SaO), and sesame oil (SeO), were purchased from a local market. Fatty acid methyl esters (FAME) standards were purchased from Sigma–Aldrich (St. Louis, MO, USA). All other chemicals and reagents used were of analytical grade. 2.2. Analysis of fatty acid (FA) by GC–MS The gas chromatography–mass spectrometry (GC–MS) analysis of fatty acid composition was performed using an Agilent GC–MS instrument (7890A-MSD5975C, Agilent Technologies, Palo Alto, CA, USA) equipped with a quadrupole selective mass detector. A CP-Sil 88 capillary column (100 m 0.25 mm i.d., 0.2 lm film thickness, Chrompack, Bridgewater, NJ, USA) was used in the analysis. The fatty-acid methyl esters (FAMEs) were prepared according to (ISO, 1990) method 5508. The injection volume of FAMEs was 1 ml. Helium was used as the carrier gas at a flow rate of 0.8 ml/min. Split ratio was 1:30. The injection temperature was 260 °C, and detector temperature was 280 °C. The GC oven program was as follows: first held at 70 °C for 1 min, and then increased to 100 °C at 5 °C/min and held for 2 min; then increased to 175 °C at 10 °C/min and held for 40 min; and finally increased to 220 °C at 15 °C/min and kept for 30 min. The individual fatty acids
were identified and quantified by comparing their retention times with external standards. 2.3. Identification and quantification of triacylglycerol The triacylglycerol (TAG) analysis was conducted on an Eksigent high performance liquid chromatography (HPLC) system (AB SCIEX, CA, USA) equipped with a tandem quadrupole mass spectrometer (TripleTOF 4600, AB SCIEX, CA, USA) and a 250 4.6 mm C18 column (Waters, Milford, MA, USA). Gradient elution was performed using n-hexane and dichloromethane (1:1, v/v) as mobile phase A and acetone and acetonitrile (5:95, v/v) as mobile phase B at a flow rate of 0.6 ml/min. The elution program used was as follows: from 0 to 20 min, held at 20% A; from 20 to 35 min, increased to 80% A; from 35 to 45 min, decreased back to 20% A. The column temperature was set at 40 °C. TAGs were identified according to their mass spectra data in literatures (Cunha & Oliveira, 2006; Holcapek, Lisa, Jandera, & Kabatova, 2005; Lerma-Garcia et al., 2011; Zeb, 2012). The amount of each TAG was normalized and expressed as a percentage of the total peak area (%). According to the types of fatty acids bonded to glycerol skeleton, the TAG were further grouped as trisaturated (TSTAG), disaturated (DSTAG), monosaturated (MSTAG), and triunsaturated triacylglycerol (TUTAG). 2.4. Differential scanning calorimetry analysis The DSC analysis was conducted on a STA449 F3 simultaneous thermoanalyzer (Netzsch, Bavaria, Germany) using oxygen (99.99%, 80 ml/min) as the purge gas and nitrogen (99.99%, 20 ml/min) as the protective gas. An oil sample of 2.7–3.6 mg was weighed into an aluminum pan and hermetically sealed with a pinhole lip, which can allow for constant interaction between the sample and the oxygen streamoxygen stream. The system was equilibrated at 30 °C for 5 min and then heated linearly to 400 °C at 5, 7.5, 10, 12.5, and 15 °C/min to generate the oxidative profile
Fig. 1. DSC oxidation curve of the eight different oils at the heating rate of 10 °C/min. (a) Sunflower oil; (b) soybean oil; (c) refined palm oil; (d) olive oil; (e) corn oil; (f) safflower oil; (g) sesame oil; (h) grapeseed oil. Top large plots: DSC oxidation curve. Bottom small charts: first derivative of DSC oxidation curve.
B. Qi et al. / Food Chemistry 194 (2016) 601–607
(heat flow against temperature). To precisely determine T s , T on , and T p , first derivative curves were plotted for each oil sample at each heating rate based on their corresponding DSC data (representative first derivative curves are shown as small charts in Fig. 1). The T s , T on , and T p , were located following the above descriptions with the help of the software (NETZSCH-Propeus-6.1) provided by the manufacturer. In our study, the non-isothermal oxidation of oils was conducted in the presence of excess oxygen, such that oil oxidation was independent of the oxygen concentration. This results in the oxidation of oils following a first order rate reaction (Adhvaryu, Erhan, Liu, & Perez, 2000; Martinez-Monteagudo, Saldan, & Kennelly, 2012; Ostrowska-Ligeza et al., 2010; Tan, Man, Selamat, & Yusoff, 2001). The Ozawa–Flynn–Wall method is one of the most commonly used iso-conversional methods to calculate kinetic parameters (Flynn & Wall, 1966; Ostrowska-Ligeza et al., 2010; Ozawa, 1965). Using this method and a set of data (T on and T p ) obtained for different heating rates (b), the activation energy (Ea , kJ/mol) and pre-exponential factor (A) were calculated using Eq. (1).
1 log b ¼ a þ b T
ð1Þ
where b is the heating rate (K/min), and T is the temperature T on or T p (K). The Ea and A were computed directly from the slope and intercept by plotting log b against 1=T using Eqs. (2) and (3).
a ¼ 0:4567
Ea R
Ea b ¼ 2:315 þ log A R
ð2Þ ð3Þ
where a and b are the slope and intercept from Eq. (1), respectively, and R is the universal gas constant (8.314 J/mol K). The activation energy (Ea ), reaction rate constant (k, min1), and pre-exponential factor (A) were then calculated using Eqs. 4–6.
Ea ¼ 2:19R
d log b dT 1
Ea
k ¼ AeRT A¼
bEa eEa =RT RT 2
ð4Þ ð5Þ ð6Þ
2.5. Statistical analysis All data were conducted at least in triplicate and reported as mean values ± standard deviations. One-way analysis of variance (ANOVA) and Duncan’s multiple-range test at a 95% confidence level (p < 0.05) were performed to identify significant differences among values using SPSS (Version 22.0, SPSS Inc., Chicago, IL, USA). Pearson correlation coefficients were calculated between the chemical compositions of oils and thermal kinetic parameters using SPSS (Version 22.0, SPSS Inc., Chicago, IL, USA). The calculated Pearson correlation coefficients (R) were used to generate a heat map, in which color scale indicated the degree of correlation using Tableau (Version 8.3, Tableau Software, Seattle, USA). 3. Results and discussion 3.1. Fatty acid composition analysis The fatty acid composition of the eight oils are shown in Table 1. Overall, all the tested fatty acid compositions were within the range of earlier reports (Tan & Man, 2000; Tan et al., 2001). Among
603
all oil samples, refined palm oil (RPO) had the most abundant palmitic acid (C16:0, 31.26%) and oleic acid (C18:1, 47.08%) content. Consequently, the saturated fatty acid (SFA) content (38.34%) of RPO was the highest, and the mono-unsaturated fatty acid (MUFA) content (47.08%) of RPO ranked the second highest among all the oils. The poly-unsaturated fatty acid (PUFA) content (14.02%) of RPO was significantly lower than other oils except olive oil (OeO). Olive oil contained the highest amount of oleic acid (C18:1, 74.84%). Due to the higher proportion of linoleic acid (C18:2, 65.47%), the PUFA content (65.47%) of grapeseed oil (GsO) ranked the second highest among all the oils. The fatty acid compositions of sunflower oil (SuO) was close to those of GsO, although SuO contained a low concentration of behenic acid (C22:0, 0.82%). Apart from the similar fatty acid compositions of GsO and SuO, corn oil (CnO) contained a small amount of linolenic acid (C18:3, 0.81%) and arachidic acid (C20:0, 0.46%). Soybean oil (SoO) was characterized by a significantly higher value of linolenic acid (C18:3, 9.53%), as observed by Li, Zhang, Wang, Jiang, and Sui (2013). The highest proportion of linoleic acid (C18:2, 76.79%) was observed in safflower oil (SaO), which in turn resulted in SaO having the highest PUFA content (76.80%). SaO was also characterized by having the lowest SFA content (11.21%) and MUFA content (11.60%). Sesame oil (SeO) had a similar amount of oleic acid (C18:1, 39.46%) and linoleic acid (C18:2, 43.69%). 3.2. Triacylglycerol composition analysis The triacylglycerol (TAG) composition of oil samples are shown in Table 1. TAG results showed significant compositional differences among oils, in accordance with those observed for FAs. RPO contained the highest amount of POO (31.81%) and PPO (28.32%). RPO was characterized by a high proportion of TSTAG (5.2%), MSTAG (42.8%) and DSTAG (50.39%), as well as a low content of TUTAG (1.93%). For OeO, due to the presence of a high concentration of oleic acid, the levels of OOO and OOL were remarkably high (18.37% and 38.18%, respectively), compared to the other oils. GsO exhibited a higher content of TUTAG (75.86%) and DSTAG (23.19%), and a lower content of MSTAG (0.87%). The TAG composition of SuO was similar to that of GsO. The proportions of MSTAG (2.79%), DSTAG (29.14%), and TUTAG (67.93%) in CnO were similar to OeO, GsO, and SuO, although the composition of each type of TAG was varied. SoO had a significantly higher amount of LLP (16.69%) among all the oils. SaO showed the highest LLL + OLLn percentage of 53.16% which subsequently contributed to the highest percentage of TUTAG (78.06%). SeO had a significantly higher content of OOO (14.90%). The MSTAG (7.84%) content of SeO was higher than other oils except RPO, and the DSTAG and TUTAG content of SeO were 32.95% and 59.21%, respectively. 3.3. DSC analysis The representative thermograms of the eight vegetable oils at the heating rate of 10 °C/min in the range of 100–400 °C are shown in Fig. 1 and of the corn oil at five heating rates (5, 7.5, 10, 12.5, and 15 °C/min) are shown in Fig. 2. The eight different oils exhibited different DSC thermograms at the heating rate of 10 °C/min, indicating different oxidation profiles. T s is identified at the point when the first derivative curve (red lines in Fig. 1) shows an inflexion point between a maximum and a minimum point of the curve. T on , which is the temperature at which a rapid increase occurs in heat flow, was calculated by extrapolating the tangent drawn on the steepest slope of the DSC thermogram. T p is considered when the heat flow has reached a maximum point on the DSC spectra. T on was reported to be the most suitable parameter for lipid oxidation under non-isothermal conditions as it is closely associated
604
B. Qi et al. / Food Chemistry 194 (2016) 601–607
Table 1 Fatty acid and triacylglycerol composition of the eight different vegetable oils.* RPO
OeO
GsO
SuO
CnO
SoO
SaO
SeO
Fatty acid (%) C10:0 C14:0 C16:0 C16:1 C18:0 C18:1 C18:2 C18:3 C20:0 C22:0
0.40 ± 0.02a 1.15 ± 0.05a 31.26 ± 0.10a nd 5.08 ± 0.13b 47.08 ± 0.11b 14.02 ± 0.04f nd 0.45 ± 0.08b nd
nd nd 11.52 ± 0.15c 0.79 ± 0.12a 3.81 ± 0.04d 74.84 ± 0.18a 6.21 ± 0.05g 0.72 ± 0.07c 0.43 ± 0.01c nd
nd nd 7.91 ± 0.06e nd 4.66 ± 0.09c 21.97 ± 0.14e 65.47 ± 0.21b nd nd nd
nd nd 7.30 ± 0.02e nd 3.87 ± 0.05d 29.24 ± 0.10d 57.87 ± 0.15c nd nd 0.82 ± 0.03a
nd nd 12.64 ± 0.12b nd 2.08 ± 0.07f 29.67 ± 0.22d 53.65 ± 0.10c 0.81 ± 0.05b 0.46 ± 0.02b nd
nd nd 12.14 ± 0.04c nd 5.03 ± 0.03b 22.41 ± 0.19e 49.58 ± 0.14d 9.53 ± 0.08a nd 0.44 ± 0.03b
nd nd 8.15 ± 0.03e nd 3.06 ± 0.04e 11.60 ± 0.13f 76.80 ± 0.08a nd nd nd
nd nd 9.76 ± 0.09d nd 6.05 ± 0.04a 39.46 ± 0.11c 43.69 ± 0.24e 0.40 ± 0.01d 0.63 ± 0.03a nd
SFA MUFA PUFA
38.34 ± 0.08a 47.08 ± 0.11b 14.02 ± 0.04f
15.76 ± 0.07d 75.63 ± 0.16a 6.93 ± 0.05g
12.57 ± 0.08e 21.97 ± 0.14e 65.47 ± 0.21b
11.99 ± 0.04e 29.24 ± 0.10d 57.87 ± 0.15c
15.18 ± 0.06d 29.67 ± 0.22d 54.46 ± 0.08d
17.61 ± 0.03b 22.41 ± 0.19e 59.11 ± 0.10c
11.21 ± 0.03f 11.60 ± 0.13f 76.80 ± 0.08a
16.44 ± 0.05c 39.46 ± 0.11c 44.09 ± 0.17e
Triacylglycerol (%) LLnLn nd LLLn nd MPP 0.53 ± 0.08a LLL + OLLn nd OLL + OLPo nd LLP nd MPL 2.45 ± 0.04a OOL 0.96 ± 0.01h PLO + SLL 9.01 ± 0.02d PPL 8.45 ± 0.05a OOO 0.97 ± 0.03f SLO 5.51 ± 0.08a POO 31.81 ± 0.15a PPO 28.32 ± 0.08a SOO 4.06 ± 0.04a PSO 2.36 ± 0.05a MMM 0.45 ± 0.02a PPP 4.22 ± 0.03a AOO nd SSO 0.90 ± 0.05a OOB nd
nd nd nd 1.72 ± 0.15f 6.68 ± 0.07f 1.51 ± 0.03e nd 18.37 ± 0.12a 6.94 ± 0.08e 1.48 ± 0.06d 38.18 ± 0.15a 2.21 ± 0.09e 15.63 ± 0.03b 1.90 ± 0.05b 2.64 ± 0.08b 1.45 ± 0.04c nd nd 0.35 ± 0.03b 0.78 ± 0.07b nd
nd 1.92 ± 0.10b nd 40.35 ± 0.08b 22.56 ± 0.05b 9.62 ± 0.06c nd 10.15 ± 0.15c 10.79 ± 0.06a 0.47 ± 0.04f 0.88 ± 0.06g 1.17 ± 0.07f 1.61 ± 0.03f 0.40 ± 0.02g nd nd nd nd nd nd nd
nd nd nd 39.58 ± 0.07b 23.37 ± 0.14b 9.81 ± 0.09c nd 5.63 ± 0.11f 9.90 ± 0.06b 0.92 ± 0.04e 3.41 ± 0.07d 3.66 ± 0.06b 1.36 ± 0.03f 1.15 ± 0.08d 0.53 ± 0.10d nd nd nd nd nd 0.68 ± 0.06a
nd 0.98 ± 0.05b nd 28.49 ± 0.13c 25.26 ± 0.06a 14.06 ± 0.04b nd 8.82 ± 0.15d 9.14 ± 0.03c 1.17 ± 0.09d 4.38 ± 0.05c 2.53 ± 0.08e 2.49 ± 0.04d 1.62 ± 0.10c 0.40 ± 0.02d nd nd nd 0.52 ± 0.05a nd nd
1.87 ± 0.03a 9.16 ± 0.06a nd 24.52 ± 0.09d 18.63 ± 0.16c 16.69 ± 0.07a nd 6.96 ± 0.05e 10.61 ± 0.09a 1.84 ± 0.03c 2.32 ± 0.06e 2.91 ± 0.12d 1.73 ± 0.01e 0.93 ± 0.03e 0.68 ± 0.05d 0.65 ± 0.02d nd nd nd nd 0.45 ± 0.03b
nd nd nd 53.16 ± 0.11a 17.91 ± 0.07e 10.14 ± 0.04c nd 3.76 ± 0.05g 7.80 ± 0.08e 0.95 ± 0.03e 3.23 ± 0.10d 1.12 ± 0.05f 1.24 ± 0.06f 0.68 ± 0.03f nd nd nd nd nd nd nd
nd 0.23 ± 0.01c nd 10.17 ± 0.05e 18.19 ± 0.14d 8.77 ± 0.10d nd 15.72 ± 0.05b 8.14 ± 0.05b 5.53 ± 0.02b 14.90 ± 0.07b 3.12 ± 0.15c 11.08 ± 0.09c 0.46 ± 0.02g 1.33 ± 0.05c 1.85 ± 0.03b nd nd 0.51 ± 0.08a nd nd
5.20 ± 0.09a 42.48 ± 0.05a 50.39 ± 0.07a 1.93 ± 0.04g
0.00 5.61 ± 0.09c 29.28 ± 0.10c 64.95 ± 0.04e
0.00 0.87 ± 0.03g 23.19 ± 0.15d 75.86 ± 0.08b
0.00 2.07 ± 0.05e 25.94 ± 0.11d 71.99 ± 0.12c
0.00 2.79 ± 0.03e 29.14 ± 0.07c 67.93 ± 0.05d
0.00 3.42 ± 0.03d 33.07 ± 0.10b 63.46 ± 0.08e
0.00 1.63 ± 0.08f 20.30 ± 0.03e 78.06 ± 0.09a
0.00 7.84 ± 0.05b 32.95 ± 0.09b 59.21 ± 0.09f
TSTAG MSTAG DSTAG TUTAG
Abbreviations: M, myristic acid(C14:0); P, palmitic acid(C16:0); Po, palmitoleic acid(C16:1); S, stearic acid(C18:0); O, oleic acid(C18:1); L, linoleic acid(C18:2); Ln, linolenic acid(C18:3); A, arachidic acid(C20:0); B, behenic acid(C22:0). * Mean values in the same row followed by the same superscript letters are not significantly different (p > 0.05). nd: not detected.
with the formation of peroxides, while T p was indicative of the termination stage (Kowalski, Ratusz, Kowalska, & Bekas, 2004). The DSC thermograms of SuO, SoO, OeO, CnO, SaO, SeO, and GsO exhibited similar flat and short profiles in the initial heating stage (100 to <200 °C), before a sudden increase to maximum heat flow temperatures. For RPO, its DSC spectra began with similar flat, but longer lasting curves (100 to >200 °C), compared to other oils, before gradually increasing to a maximum heat flow temperature. In greater details, the changes in T s , T on and T p for the eight oil samples under each heating rates are shown in Table 2. The T s , T on , and T p shifted towards to higher values as the heating rate increases, which was in agreement with previous studies (Martinez-Monteagudo et al., 2012; Ostrowska-Ligeza et al., 2010). Vecchio et al. (2009) also observed similar increases in peak temperature (T p ) as heating rates were increased to 10 K min1. Although the changes in T s and T on were not reported, DSC profiles were found not to vary at the other heating rates (2.5, 5, and 7.5 K min1). The increase in heating rate from 5 to 15 °C/min led to an increase in T s and T on of around 20 °C for all oil samples, with an even larger magnitude of increment for T p (up to 70 °C for CnO), which indicated that T p was more sensitive than T s and T on . Martinez-Monteagudo et al. also arrived at the same conclusion in their evaluation of the non-isothermal kinetics
of anhydrous milk fat (AMF) rich in conjugated linoleic acid (CLA). They found that T p varied (increased by 35.15 and 36.62 °C for medium and high CLA AMF, respectively) more notably than T s (increased by 22.13 and 22.19 °C for medium and high CLA AMF, respectively) and T on (increased by 25.19 and 28.68 °C for medium and high CLA AMF, respectively), while for low CLA AMF the variance for T s , T on , T p were close, 15.34 for T s , 22.36 for T on , and 21.24 for T p .
3.4. Thermal kinetic analysis The kinetic parameters, namely activation energy (Ea ), preexponential factor (A), and rate constant (k) at 200 °C, which were computed from T s , T on , and T p for each oil, are shown in Table 3. The high correlation coefficients (R2 ) indicated that experimental data was well fitted to the models (Eqs. 4–6). For all oil samples, the Ea values at T p were smaller than that at T s and T on . Similarly, the Ea values at T p were also found to be smaller than that at T s and T on in the work conducted by MartinezMonteagudo et al. (2012). RPO exhibited the highest Ea values, 126.06 kJ/mol at T s , 134.7 kJ/mol at T on , and 91.88 kJ/mol at T p , among all the oils. As expected, the highest Ea values of RPO gave
605
B. Qi et al. / Food Chemistry 194 (2016) 601–607
3.5. Pearson correlation analysis The Pearson correlation coefficients between FA and TAG, and thermal kinetic parameters (Ea and k) at T s , T on , and T p were analyzed. For ease of analysis, the calculated Pearson correlation coefficients (R) were further subjected to create a heat map indicating the degree of correlation (Fig. 3). In general, the k at the three temperatures exhibited correlations with FA and TAG at varying degrees. Positive correlation coefficients were observed between the k at T s and linoleic acid (0.85), PUFA (0.87), and LLL + OLLn (0.93) at the significance level of 0.01, and OLL + OLPo (0.77) and TUTAG (0.81) at the significance level of 0.05. In addition, the k at T s was negatively correlated with oleic acid (0.73), SFA (0.78), MUFA (0.72), SSO (0.76), and MSTAG (0.78) at the significance level of 0.05, and arachidic acid (0.86), PPL (0.85), POO (0.91), SOO (0.90), PSO (0.94), and DSTAG (0.84) at the significance level of 0.01. Both MUFA and PUFA are unsaturated fatty acids despite their different degrees of saturation. However, it is interesting to note that the k value at T s appeared to be positively correlated with PUFA and negatively correlated with MUFA. Furthermore, no significant positive and negative correlations were found between k at T on and FA & TAG. The k at T p showed similar correlation coefficients to k at T s . However, k at T p exhibited a significant positive correlation to LLP, and no negative correlation to arachidic acid (C20:0) and SFA. As for correlations between k and TAGs, the TSTAG did not show any positive correlations with k, which may be due to the low concentrations of TSTAG in oils, except RPO. Our results showed that only fully unsaturated TAG (TUTAG) positively affected the k values at T s and T p , which, for the first time, indicated the interesting relationship between unsaturated TAG and k. The Ea at T s , T on , and T p gave very similar correlations to FA and TAG. In particular, significant positive correlations were found between all the Ea values at T s , T on , and T p , and capric acid (C10:0), myristic acid (C14:0), palmitic acid (C16:0), SFA, MPP, MPL, PPL, SLO, POO, PPO, SOO, PSO, MMM, PPP, TSTAG, MSTAG, and DSTAG. The Ea at T p additionally positively correlated with SSO. Linoleic acid (C18:2), PUFA, and LLL + OLLn were negatively correlated with the Ea at T s and T p , whereas OLL + OLPo were negatively correlated with the Ea at T on and T p , and TUTAG was negatively correlated with the Ea at the three temperatures. Vecchio et al. (2009) evaluated the thermal decomposition of extra virgin
Fig. 2. DSC oxidative profiles of corn oil at five heating rates.
the lowest k values of 0.339 min1 at T s , 0.157 min1 at T on , and 0.001 min1 at T p . The oxidation kinetic parameters of RPO indicated that it was the most stable oil among the tested oil samples. This is not surprising when considering that RPO had the highest percentage of SFA (38.49%), which was more than twice as high compared to the other oils, resulting in RPO being highly stable (Nagaraj, 2009). Due to the abundant amount of PUFA (76.8%, Table 1), which is oxidized very readily (Wijesundera, 2008), SaO showed the lowest Ea values at T s (72.76 kJ/mol) and T p (31.85 kJ/ mol). Besides, the lowest Ea value at T on (69.58 kJ/mol) was found for grapeseed oil. According to the Ea values, the stability scale for the tested oils could be proposed as RPO > SeO > OeO > CnO > SuO > SoO > GsO > SaO at T s , RPO > SoO > SeO > OeO > SaO > SuO > CnO > GsO at T on , and RPO > SeO > SoO > OeO > GsO > SuO > CnO > SaO at T p . The differences observed in the stability of oils may be attributed to their various compositions. To explore the relationship between oil compositions and its thermal attributes, a Pearson correlation analysis was conducted in Section 3.5.
Table 2 The start (T s ) onset (T on ) and maximum heat flow (T p ) temperature of oils under the five heating rates.*
*
Heating rate (°C/min)
RPO
OeO
GsO
SuO
CnO
SoO
SaO
SeO
Start temperature (Ts) 5.0 7.5 10.0 12.5 15.0
203.4 ± 0.3a 206.7 ± 0.8a 211.0 ± 0.1a 214.3 ± 0.2a 219.3 ± 0.8a
178.8 ± 0.6c 187.1 ± 1.1c 189.8 ± 0.1c 195.5 ± 0.7b 199.8 ± 0.4b
161.3 ± 0.1f 168.4 ± 0.4e 172.6 ± 0.3f 178.9 ± 1.0e 182.9 ± 0.6e
167.6 ± 0.8e 173.5 ± 1.3d 179.5 ± 0.6e 182.6 ± 0.4d 174.9 ± 0.6g
172.1 ± 0.3d 176.7 ± 0.1d 183.2 ± 0.4d 188.2 ± 0.9c 190.9 ± 0.2c
167.3 ± 0.7e 171.8 ± 0.8d 177.1 ± 0.5e 183.5 ± 0.2d 186.8 ± 1.1d
159.7 ± 0.5f 162.1 ± 0.1f 171.0 ± 0.6f 177.8 ± 0.7e 179.4 ± 0.4f
184.5 ± 0.2b 190.0 ± 0.4b 197.8 ± 0.9b 194.1 ± 1.4b 202.3 ± 0.5b
Onset temperature (Ton) 5.0 7.5 10.0 12.5 15.0
212.3 ± 0.1a 216.4 ± 0.2a 221.5 ± 0.1a 224.1 ± 0.4a 227.8 ± 0.6a
187.3 ± 0.5b 194.8 ± 0.3b 203.2 ± 0.2b 202.8 ± 0.8b 208.6 ± 1.0b
165.2 ± 0.6d 172.5 ± 0.3d 177.1 ± 1.6e 185.6 ± 0.2e 189.8 ± 0.8d
171.4 ± 1.0c 176.6 ± 1.2c 182.5 ± 0.1d 191.1 ± 0.9d 189.0 ± 0.8d
172.5 ± 0.1c 177.2 ± 0.4c 187.6 ± 0.2c 192.6 ± 0.1d 193.5 ± 0.8c
169.2 ± 0.5c 175.3 ± 0.1c 178.5 ± 0.8e 184.4 ± 0.1e 187.1 ± 0.2e
164.5 ± 0.1d 165.1 ± 0.1e 175.0 ± 0.3f 178.9 ± 0.1f 180.0 ± 0.4f
188.6 ± 0.4b 192.3 ± 0.5b 201.6 ± 0.2b 198.1 ± 0.5c 207.6 ± 0.1b
maximum heat flow temperature (Tp) 5.0 337.1 ± 0.7a 7.5 348.8 ± 0.1a 10.0 354.1 ± 0.4a 12.5 368.3 ± 0.3a 15.0 372.7 ± 0.4a
253.1 ± 0.2b 284.8 ± 0.1b 290.2 ± 0.7b 293.0 ± 0.6b 299.7 ± 0.2b
220.8 ± 0.3e 254.2 ± 0.8c 255.0 ± 0.2d 270.4 ± 1.0d 277.9 ± 0.3d
244.7 ± 0.1c 252.2 ± 0.1d 263.4 ± 0.6c 292.2 ± 0.3e 263.1 ± 0.2e
233.5 ± 0.4d 250.2 ± 0.6d 261.0 ± 0.5c 265.8 ± 0.8e 303.4 ± 0.2b
206.9 ± 0.5f 217.0 ± 0.2f 220.5 ± 0.1e 235.3 ± 0.8f 246.3 ± 0.3f
221.3 ± 0.4e 249.8 ± 0.1d 251.4 ± 0.6d 288.4 ± 0.3c 290.3 ± 0.5c
234.6 ± 0.7d 238.7 ± 0.2e 263.4 ± 0.8c 264.5 ± 0.4e 265.6 ± 0.7e
Mean values in the same row followed by the same superscript letters are not significantly different (p > 0.05).
606
B. Qi et al. / Food Chemistry 194 (2016) 601–607
Table 3 The kinetic parameters (k at 200 °C) calculated from start (T s ) onset (T on ) and peak (T p ) temperature under each heating rate. Ts Ea . RPO OeO GsO SuO CnO SoO SaO SeO
126.06 90.16 79.26 89.16 89.79 85.18 72.76 100.43
Ton A
k 13
2.85 10 8.82 109 1.19 109 1.49 1010 1.23 1010 4.66 109 2.33 108 1.06 1011
0.339 0.970 2.092 2.118 1.487 1.821 2.147 0.857
2
R
Ea .
0.9789 0.9922 0.9949 0.7468 0.9919 0.9865 0.9671 0.9264
134.7 89.23 69.58 82.62 74.23 96.93 86.05 93.21
Tp A
k 14
1.18 10 4.31 109 6.92 107 1.83 109 1.77 105 1.01 1011 7.57 109 1.31 1010
0.157 0.601 1.427 1.37 1.117 1.985 2.366 10.664
2
R
0.9941 0.9800 0.9890 0.9648 0.9777 0.9929 0.9445 0.9199
Ea . 91.88 52.70 44.37 40.81 36.21 52.84 31.85 56.74
A 7
1.85 10 4.82 104 6.85 103 6.01 103 2.32 103 2.05 105 9.77 102 2.38 105
k
R2
0.001 0.073 0.216 0.187 0.232 0.300 0.297 0.129
0.9851 0.9330 0.9635 0.7312 0.9300 0.9616 0.9626 0.9272
Fig. 3. Heat map for the Pearson correlation analysis.
olive oils and they also found positive correlations between Ea calculated from II deconvoluted peak, palmitic acid (C16:0, 0.74), SFA (0.78), and MSTAG (0.65) at the significance level of 0.01. The increased percentage of saturated CH2 carbons in oil molecules would improve the resistance of oil against the initial thermal breakdown, leading to a higher Ea value (Adhvaryu et al., 2000). However, our results showed that the Ea values had no statistical correlation with stearic acid (C18:0), arachidic acid (C20:0), and behenic acid (C22:0), although their contents in oils were close to capric acid (C10:0) and myristic acid (C14:0). Additionally, it is worthy to point out that although k values at T s and T p were negatively correlated with MUFA, the Ea values showed no significant correlation with MUFA. The TSTAG showed a significantly positive correlation with Ea values. However, MSTAG and DSTAG, which contained two and one unsaturated fatty acids in their TAG
structures respectively, were also statistically correlated with Ea values, with only TAGs containing three unsaturated fatty acids (TUTAG) negatively affecting Ea values.
4. Conclusions All oil samples exhibited varying DSC thermograms, indicating different thermal oxidation attributes, as affected by their compositions. The Pearson correlation analysis showed that (i) the k value at T s and T p were positively correlated with PUFA while being negatively correlated with MUFA; (ii) the k values at T s and T p were found to have unique positive correlations with TUTAG; and (iii) the Ea values at T s , T on , and T p were not positively correlated with SFA as well as TSTAG, whereas, they were instead positively
B. Qi et al. / Food Chemistry 194 (2016) 601–607
correlated with MSTAG and DSTAG. This study could be used as a possible tool to predicate the thermal oxidation properties of oil based on its FAs and TAGs composition. Acknowledgements The authors would like to acknowledge the support for this study by the National Natural Science Foundation of China (research Grant number: 31430067), the National Key Technology Research and Development Program of the Ministry of Science and Technology of China (research Grant number: 2014BAD22B01-02), Educational Commission of Heilongjiang Province of China (research Grant number: 12531049), Natural Science Foundation of Heilongjiang Province of China (research Grant number: QC2013C014), and the Key Laboratory of Soybean Biology of Chinese Education Ministry. References Adhvaryu, A., Erhan, S. Z., Liu, Z. S., & Perez, J. M. (2000). Oxidation kinetic studies of oils derived from unmodified and genetically modified vegetables using pressurized differential scanning calorimetry and nuclear magnetic resonance spectroscopy. Thermochimica Acta, 364, 87–97. Caponio, F., Chiavaro, E., Paradiso, V. M., Paciulli, M., Summo, C., Cerretani, L., & Gomes, T. (2013). Chemical and thermal evaluation of olive oil refining at different oxidative levels. European Journal of Lipid Science and Technology, 115, 1146–1154. Cerretani, L., Bendini, A., Rinaldi, M., Paciulli, M., Vecchio, S., & Chiavaro, E. (2012). DSC evaluation of extra virgin olive oil stability under accelerated oxidative test: Effect of fatty acid composition and phenol contents. Journal of Oleo Science, 61, 303–309. Cerretani, L., Maggio, R. M., Barnaba, C., Toschi, T. G., & Chiavaro, E. (2011). Application of partial least square regression to differential scanning calorimetry data for fatty acid quantitation in olive oil. Food Chemistry, 127, 1899–1904. Choe, E., & Min, D. B. (2006). Mechanisms and factors for edible oil oxidation. Comprehensive Reviews in Food Science and Food Safety, 5, 169–186. Cunha, S. C., & Oliveira, M. (2006). Discrimination of vegetable oils by triacylglycerols evaluation of profile using HPLC/ELSD. Food Chemistry, 95, 518–524. Dijkstra, A. J., & Segrers, J. C. (2007). Production and refining of oils and fats. In F. D. Gunstone, J. L. Harwood, & A. J. Dijkstra (Eds.), The lipid handbook (3rd ed., pp. 143–251). CRC Press. Flynn, J. H., & Wall, L. A. (1966). A quick direct method for determination of activation energy from thermogravimetric data. Journal of Polymer Science Part B-Polymer Letters, 4, 323–328. Holcapek, M., Lisa, M., Jandera, P., & Kabatova, N. (2005). Quantitation of triacylglycerols in plant oils using HPLC with APCI-MS, evaporative lightscattering, and UV detection. Journal of Separation Science, 28, 1315–1333.
607
ISO, 1990 ISO 5508:1990, Animal and vegetable fats and oils, Analysis by gas chromatography of methyl esters of fatty acids, American National Standards Institute (ANSI), 1990. Kowalski, B., Ratusz, K., Kowalska, D., & Bekas, W. (2004). Determination of the oxidative stability of vegetable oils by differential scanning calorimetry and rancimat measurements. European Journal of Lipid Science and Technology, 106, 165–169. Lerma-Garcia, M. J., Lusardi, R., Chiavaro, E., Cerretani, L., Bendini, A., Ramis-Ramos, G., & Simo-Alfonso, E. F. (2011). Use of triacylglycerol profiles established by high performance liquid chromatography with ultraviolet-visible detection to predict the botanical origin of vegetable oils. Journal of Chromatography A, 1218, 7521–7527. Li, Y., Zhang, Y., Wang, M., Jiang, L., & Sui, X. (2013). Simplex-centroid mixture design applied to the aqueous enzymatic extraction of fatty acid-balanced oil from mixed seeds. Journal of the American Oil Chemists’ Society, 90, 349–357. Maggio, R., Barnaba, C., Cerretani, L., Paciulli, M., & Chiavaro, E. (2014). Study of the influence of triacylglycerol composition on DSC cooling curves of extra virgin olive oil by chemometric data processing. Journal of Thermal Analysis and Calorimetry, 115, 2037–2044. Martinez-Monteagudo, S. I., Saldan, M. D. A., & Kennelly, J. J. (2012). Kinetics of nonisothermal oxidation of anhydrous milk fat rich in conjugated linoleic acid using differential scanning calorimetry. Journal of Thermal Analysis and Calorimetry, 107, 973–981. Nagaraj, G. (2009). Oilseeds: properties, products, processing and procedures. New Delhi: New India Publishing. Ostrowska-Ligeza, E., Bekas, W., Kowalska, D., Lobacz, M., Wroniak, M., & Kowalski, B. (2010). Kinetics of commercial olive oil oxidation: Dynamic differential scanning calorimetry and Rancimat studies. European Journal of Lipid Science and Technology, 112, 268–274. Ozawa, T. (1965). A new method of analyzing thermogravimetric data. Bulletin of the Chemical Society of Japan, 38, 1881–1886. Shahidi, F., & Zhong, Y. (2010). Lipid oxidation and improving the oxidative stability. Chemical Society Reviews, 39, 4067–4079. Tan, C. P., & Man, Y. B. C. (2000). Differential scanning calorimetric analysis of edible oils: Comparison of thermal properties and chemical composition. Journal of the American Oil Chemists Society, 77, 143–155. Tan, C. P., Man, Y. B. C., Selamat, J., & Yusoff, M. S. A. (2001). Application of arrhenius kinetics to evaluate oxidative stability in vegetable oils by isothermal differential scanning calorimetry. Journal of the American Oil Chemists Society, 78, 1133–1138. Tan, C. P., & Nehdi, I. A. (2014). DSC analysis of vegetable oils: Relationship between thermal profiles and chemical composition. In E. Chiavaro (Ed.), Differential scanning calorimetry: Applications in fat and oil technology (pp. 3–26). CRC Press. Vecchio Ciprioti, S., & Chiavaro, E. (2014). Introduction. In E. Chiavaro (Ed.), Differential scanning calorimetry: Applications in fat and oil technology (pp. xv–xvii). CRC Press. Vecchio, S., Cerretani, L., Bendini, A., & Chiavaro, E. (2009). Thermal decomposition study of monovarietal extra virgin olive oil by simultaneous thermogravimetry/ differential scanning calorimetry: Relation with chemical composition. Journal of Agricultural and Food Chemistry, 57, 4793–4800. Wijesundera, C. (2008). The influence of triacylglycerol structure on the oxidative stability of polyunsaturated oils. Lipid Technology, 20, 199–202. Zeb, A. (2012). Triacylglycerols composition, oxidation and oxidation compounds in camellia oil using liquid chromatography–mass spectrometry. Chemistry and Physics of Lipids, 165, 608–614.
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Differential scanning calorimetry study—Assessing the influence of composition of vegetable oils on oxidation Baokun Qi 1, Qiaozhi Zhang 1, Xiaonan Sui, Zhongjiang Wang, Yang Li ⇑, Lianzhou Jiang ⇑ College of Food Science, Northeast Agricultural University, Harbin 150030, China National Research Center of Soybean Engineering and Technology, Harbin 150030, China
a r t i c l e
i n f o
Article history: Received 9 February 2015 Received in revised form 30 July 2015 Accepted 31 July 2015 Available online 6 August 2015 Keywords: Differential scanning calorimetry Vegetable oil Fatty acid Triacylglycerol Lipid oxidation Kinetic analysis
a b s t r a c t The thermal oxidation of eight different vegetable oils was studied using differential scanning calorimetry (DSC) under non-isothermal conditions at five different heating rates (5, 7.5, 10, 12.5, and 15 °C/min), in a temperature range of 100–400 °C. For all oils, the activation energy (Ea ) values at T p were smaller than that at T s and T on . Among all the oils, refined palm oil (RPO) exhibited the highest Ea values, 126.06 kJ/mol at T s , 134.7 kJ/mol at T on , and 91.88 kJ/mol at T p . The Ea and reaction rate constant (k) values at T s , T on , and T p were further correlated with oil compositions (fatty acids and triacylglycerols) using Pearson correlation analysis. The rate constant (k) and Ea of all oils exhibited varying correlations with FAs and TAGs, indicating that the thermal oxidation behaviors were affected by oil compositions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Vegetable oils play an important role in the human diet due to their nutritional and sensory properties. They are also one of the main ingredients in a wide range of food products, such as ice cream, sausages, and margarine (Shahidi & Zhong, 2010). However, oils are susceptible to oxidation reactions during food processing and subsequent storage of food products. The oxidation of oils leads to the development of off-flavor compounds and undesirable chemical changes of foods, resulting in decreased nutritional value of foods and potential food safety problems (Choe & Min, 2006; Dijkstra & Segrers, 2007). Differential scanning calorimetry (DSC), which belongs to the family of thermal analysis, has been applied for the last 50 years in the field of fats and oils for various purposes (Tan & Nehdi, 2014). In particular, DSC evaluates the thermal oxidation behaviors of oils through precise recording of heat flow into and out of an oil sample. The heat flow is reported as a function of either time or temperature and plotted on the DSC thermograms, in which each peak is associated with a specific physical or chemical process (Vecchio Ciprioti & Chiavaro, 2014). As a very well-established ⇑ Corresponding authors at: College of Food Science, Northeast Agricultural University, Harbin 150030, China. E-mail addresses: [email protected] (Y. Li), [email protected] (L. Jiang). 1 The first and second author contributed equally to this work. http://dx.doi.org/10.1016/j.foodchem.2015.07.148 0308-8146/Ó 2015 Elsevier Ltd. All rights reserved.
method, DSC has been widely applied in lipid studies. Cerretani et al. (2012) investigated the effect of fatty acid composition and phenol contents on the stability of extra virgin olive oil under accelerated oxidative test using DSC. Caponio et al. (2013) applied DSC in the evaluation of cooling and heating curves as well as thermal attributes of olive oil during refining process. Cerretani, Maggio, Barnaba, Toschi, and Chiavaro (2011) quantified the fatty acid in olive oils using a novel DSC-PLS (partial least square) method. Maggio, Barnaba, Cerretani, Paciulli, and Chiavaro (2014) developed a faster analytical method based on DSC, named highpressure liquid chromatography–differential scanning calorimetry–partial least square (HPLC–DSC–PLS), to investigate the DSC cooling curves of extra virgin olive oil affected by triacylglycerol composition. Oil oxidation is a complex reaction involving several simultaneous reactions. The different oxidation behaviors among oils may be affected by the various oil compositions. To the best of our knowledge, only one study conducted by Vecchio, Cerretani, Bendini, and Chiavaro (2009) investigated the relationship between oil composition and thermal oxidation properties. They reported that the onset temperatures of the thermal decomposition transition, the maximum heat flow temperatures, and as well as the sum of enthalpy were found to significantly correlate with the components of twelve different extra virgin olive oils. However, no study was found to address the relationship between oil composition and
602
B. Qi et al. / Food Chemistry 194 (2016) 601–607
thermal oxidation properties for other vegetable oils, such as soybean oil, corn oil, and sunflower oil. Therefore, the aim of this study was to evaluate the thermal oxidation properties of vegetable oils using DSC, and further investigate the relationships between oil compositions and thermal oxidation properties. 2. Materials and methods 2.1. Materials Eight different vegetable oils, namely refined palm oil (RPO), olive oil (OeO), grapeseed oil (GsO), sunflower oil (SuO), corn oil (CnO), soybean oil (SoO), safflower oil (SaO), and sesame oil (SeO), were purchased from a local market. Fatty acid methyl esters (FAME) standards were purchased from Sigma–Aldrich (St. Louis, MO, USA). All other chemicals and reagents used were of analytical grade. 2.2. Analysis of fatty acid (FA) by GC–MS The gas chromatography–mass spectrometry (GC–MS) analysis of fatty acid composition was performed using an Agilent GC–MS instrument (7890A-MSD5975C, Agilent Technologies, Palo Alto, CA, USA) equipped with a quadrupole selective mass detector. A CP-Sil 88 capillary column (100 m 0.25 mm i.d., 0.2 lm film thickness, Chrompack, Bridgewater, NJ, USA) was used in the analysis. The fatty-acid methyl esters (FAMEs) were prepared according to (ISO, 1990) method 5508. The injection volume of FAMEs was 1 ml. Helium was used as the carrier gas at a flow rate of 0.8 ml/min. Split ratio was 1:30. The injection temperature was 260 °C, and detector temperature was 280 °C. The GC oven program was as follows: first held at 70 °C for 1 min, and then increased to 100 °C at 5 °C/min and held for 2 min; then increased to 175 °C at 10 °C/min and held for 40 min; and finally increased to 220 °C at 15 °C/min and kept for 30 min. The individual fatty acids
were identified and quantified by comparing their retention times with external standards. 2.3. Identification and quantification of triacylglycerol The triacylglycerol (TAG) analysis was conducted on an Eksigent high performance liquid chromatography (HPLC) system (AB SCIEX, CA, USA) equipped with a tandem quadrupole mass spectrometer (TripleTOF 4600, AB SCIEX, CA, USA) and a 250 4.6 mm C18 column (Waters, Milford, MA, USA). Gradient elution was performed using n-hexane and dichloromethane (1:1, v/v) as mobile phase A and acetone and acetonitrile (5:95, v/v) as mobile phase B at a flow rate of 0.6 ml/min. The elution program used was as follows: from 0 to 20 min, held at 20% A; from 20 to 35 min, increased to 80% A; from 35 to 45 min, decreased back to 20% A. The column temperature was set at 40 °C. TAGs were identified according to their mass spectra data in literatures (Cunha & Oliveira, 2006; Holcapek, Lisa, Jandera, & Kabatova, 2005; Lerma-Garcia et al., 2011; Zeb, 2012). The amount of each TAG was normalized and expressed as a percentage of the total peak area (%). According to the types of fatty acids bonded to glycerol skeleton, the TAG were further grouped as trisaturated (TSTAG), disaturated (DSTAG), monosaturated (MSTAG), and triunsaturated triacylglycerol (TUTAG). 2.4. Differential scanning calorimetry analysis The DSC analysis was conducted on a STA449 F3 simultaneous thermoanalyzer (Netzsch, Bavaria, Germany) using oxygen (99.99%, 80 ml/min) as the purge gas and nitrogen (99.99%, 20 ml/min) as the protective gas. An oil sample of 2.7–3.6 mg was weighed into an aluminum pan and hermetically sealed with a pinhole lip, which can allow for constant interaction between the sample and the oxygen streamoxygen stream. The system was equilibrated at 30 °C for 5 min and then heated linearly to 400 °C at 5, 7.5, 10, 12.5, and 15 °C/min to generate the oxidative profile
Fig. 1. DSC oxidation curve of the eight different oils at the heating rate of 10 °C/min. (a) Sunflower oil; (b) soybean oil; (c) refined palm oil; (d) olive oil; (e) corn oil; (f) safflower oil; (g) sesame oil; (h) grapeseed oil. Top large plots: DSC oxidation curve. Bottom small charts: first derivative of DSC oxidation curve.
B. Qi et al. / Food Chemistry 194 (2016) 601–607
(heat flow against temperature). To precisely determine T s , T on , and T p , first derivative curves were plotted for each oil sample at each heating rate based on their corresponding DSC data (representative first derivative curves are shown as small charts in Fig. 1). The T s , T on , and T p , were located following the above descriptions with the help of the software (NETZSCH-Propeus-6.1) provided by the manufacturer. In our study, the non-isothermal oxidation of oils was conducted in the presence of excess oxygen, such that oil oxidation was independent of the oxygen concentration. This results in the oxidation of oils following a first order rate reaction (Adhvaryu, Erhan, Liu, & Perez, 2000; Martinez-Monteagudo, Saldan, & Kennelly, 2012; Ostrowska-Ligeza et al., 2010; Tan, Man, Selamat, & Yusoff, 2001). The Ozawa–Flynn–Wall method is one of the most commonly used iso-conversional methods to calculate kinetic parameters (Flynn & Wall, 1966; Ostrowska-Ligeza et al., 2010; Ozawa, 1965). Using this method and a set of data (T on and T p ) obtained for different heating rates (b), the activation energy (Ea , kJ/mol) and pre-exponential factor (A) were calculated using Eq. (1).
1 log b ¼ a þ b T
ð1Þ
where b is the heating rate (K/min), and T is the temperature T on or T p (K). The Ea and A were computed directly from the slope and intercept by plotting log b against 1=T using Eqs. (2) and (3).
a ¼ 0:4567
Ea R
Ea b ¼ 2:315 þ log A R
ð2Þ ð3Þ
where a and b are the slope and intercept from Eq. (1), respectively, and R is the universal gas constant (8.314 J/mol K). The activation energy (Ea ), reaction rate constant (k, min1), and pre-exponential factor (A) were then calculated using Eqs. 4–6.
Ea ¼ 2:19R
d log b dT 1
Ea
k ¼ AeRT A¼
bEa eEa =RT RT 2
ð4Þ ð5Þ ð6Þ
2.5. Statistical analysis All data were conducted at least in triplicate and reported as mean values ± standard deviations. One-way analysis of variance (ANOVA) and Duncan’s multiple-range test at a 95% confidence level (p < 0.05) were performed to identify significant differences among values using SPSS (Version 22.0, SPSS Inc., Chicago, IL, USA). Pearson correlation coefficients were calculated between the chemical compositions of oils and thermal kinetic parameters using SPSS (Version 22.0, SPSS Inc., Chicago, IL, USA). The calculated Pearson correlation coefficients (R) were used to generate a heat map, in which color scale indicated the degree of correlation using Tableau (Version 8.3, Tableau Software, Seattle, USA). 3. Results and discussion 3.1. Fatty acid composition analysis The fatty acid composition of the eight oils are shown in Table 1. Overall, all the tested fatty acid compositions were within the range of earlier reports (Tan & Man, 2000; Tan et al., 2001). Among
603
all oil samples, refined palm oil (RPO) had the most abundant palmitic acid (C16:0, 31.26%) and oleic acid (C18:1, 47.08%) content. Consequently, the saturated fatty acid (SFA) content (38.34%) of RPO was the highest, and the mono-unsaturated fatty acid (MUFA) content (47.08%) of RPO ranked the second highest among all the oils. The poly-unsaturated fatty acid (PUFA) content (14.02%) of RPO was significantly lower than other oils except olive oil (OeO). Olive oil contained the highest amount of oleic acid (C18:1, 74.84%). Due to the higher proportion of linoleic acid (C18:2, 65.47%), the PUFA content (65.47%) of grapeseed oil (GsO) ranked the second highest among all the oils. The fatty acid compositions of sunflower oil (SuO) was close to those of GsO, although SuO contained a low concentration of behenic acid (C22:0, 0.82%). Apart from the similar fatty acid compositions of GsO and SuO, corn oil (CnO) contained a small amount of linolenic acid (C18:3, 0.81%) and arachidic acid (C20:0, 0.46%). Soybean oil (SoO) was characterized by a significantly higher value of linolenic acid (C18:3, 9.53%), as observed by Li, Zhang, Wang, Jiang, and Sui (2013). The highest proportion of linoleic acid (C18:2, 76.79%) was observed in safflower oil (SaO), which in turn resulted in SaO having the highest PUFA content (76.80%). SaO was also characterized by having the lowest SFA content (11.21%) and MUFA content (11.60%). Sesame oil (SeO) had a similar amount of oleic acid (C18:1, 39.46%) and linoleic acid (C18:2, 43.69%). 3.2. Triacylglycerol composition analysis The triacylglycerol (TAG) composition of oil samples are shown in Table 1. TAG results showed significant compositional differences among oils, in accordance with those observed for FAs. RPO contained the highest amount of POO (31.81%) and PPO (28.32%). RPO was characterized by a high proportion of TSTAG (5.2%), MSTAG (42.8%) and DSTAG (50.39%), as well as a low content of TUTAG (1.93%). For OeO, due to the presence of a high concentration of oleic acid, the levels of OOO and OOL were remarkably high (18.37% and 38.18%, respectively), compared to the other oils. GsO exhibited a higher content of TUTAG (75.86%) and DSTAG (23.19%), and a lower content of MSTAG (0.87%). The TAG composition of SuO was similar to that of GsO. The proportions of MSTAG (2.79%), DSTAG (29.14%), and TUTAG (67.93%) in CnO were similar to OeO, GsO, and SuO, although the composition of each type of TAG was varied. SoO had a significantly higher amount of LLP (16.69%) among all the oils. SaO showed the highest LLL + OLLn percentage of 53.16% which subsequently contributed to the highest percentage of TUTAG (78.06%). SeO had a significantly higher content of OOO (14.90%). The MSTAG (7.84%) content of SeO was higher than other oils except RPO, and the DSTAG and TUTAG content of SeO were 32.95% and 59.21%, respectively. 3.3. DSC analysis The representative thermograms of the eight vegetable oils at the heating rate of 10 °C/min in the range of 100–400 °C are shown in Fig. 1 and of the corn oil at five heating rates (5, 7.5, 10, 12.5, and 15 °C/min) are shown in Fig. 2. The eight different oils exhibited different DSC thermograms at the heating rate of 10 °C/min, indicating different oxidation profiles. T s is identified at the point when the first derivative curve (red lines in Fig. 1) shows an inflexion point between a maximum and a minimum point of the curve. T on , which is the temperature at which a rapid increase occurs in heat flow, was calculated by extrapolating the tangent drawn on the steepest slope of the DSC thermogram. T p is considered when the heat flow has reached a maximum point on the DSC spectra. T on was reported to be the most suitable parameter for lipid oxidation under non-isothermal conditions as it is closely associated
604
B. Qi et al. / Food Chemistry 194 (2016) 601–607
Table 1 Fatty acid and triacylglycerol composition of the eight different vegetable oils.* RPO
OeO
GsO
SuO
CnO
SoO
SaO
SeO
Fatty acid (%) C10:0 C14:0 C16:0 C16:1 C18:0 C18:1 C18:2 C18:3 C20:0 C22:0
0.40 ± 0.02a 1.15 ± 0.05a 31.26 ± 0.10a nd 5.08 ± 0.13b 47.08 ± 0.11b 14.02 ± 0.04f nd 0.45 ± 0.08b nd
nd nd 11.52 ± 0.15c 0.79 ± 0.12a 3.81 ± 0.04d 74.84 ± 0.18a 6.21 ± 0.05g 0.72 ± 0.07c 0.43 ± 0.01c nd
nd nd 7.91 ± 0.06e nd 4.66 ± 0.09c 21.97 ± 0.14e 65.47 ± 0.21b nd nd nd
nd nd 7.30 ± 0.02e nd 3.87 ± 0.05d 29.24 ± 0.10d 57.87 ± 0.15c nd nd 0.82 ± 0.03a
nd nd 12.64 ± 0.12b nd 2.08 ± 0.07f 29.67 ± 0.22d 53.65 ± 0.10c 0.81 ± 0.05b 0.46 ± 0.02b nd
nd nd 12.14 ± 0.04c nd 5.03 ± 0.03b 22.41 ± 0.19e 49.58 ± 0.14d 9.53 ± 0.08a nd 0.44 ± 0.03b
nd nd 8.15 ± 0.03e nd 3.06 ± 0.04e 11.60 ± 0.13f 76.80 ± 0.08a nd nd nd
nd nd 9.76 ± 0.09d nd 6.05 ± 0.04a 39.46 ± 0.11c 43.69 ± 0.24e 0.40 ± 0.01d 0.63 ± 0.03a nd
SFA MUFA PUFA
38.34 ± 0.08a 47.08 ± 0.11b 14.02 ± 0.04f
15.76 ± 0.07d 75.63 ± 0.16a 6.93 ± 0.05g
12.57 ± 0.08e 21.97 ± 0.14e 65.47 ± 0.21b
11.99 ± 0.04e 29.24 ± 0.10d 57.87 ± 0.15c
15.18 ± 0.06d 29.67 ± 0.22d 54.46 ± 0.08d
17.61 ± 0.03b 22.41 ± 0.19e 59.11 ± 0.10c
11.21 ± 0.03f 11.60 ± 0.13f 76.80 ± 0.08a
16.44 ± 0.05c 39.46 ± 0.11c 44.09 ± 0.17e
Triacylglycerol (%) LLnLn nd LLLn nd MPP 0.53 ± 0.08a LLL + OLLn nd OLL + OLPo nd LLP nd MPL 2.45 ± 0.04a OOL 0.96 ± 0.01h PLO + SLL 9.01 ± 0.02d PPL 8.45 ± 0.05a OOO 0.97 ± 0.03f SLO 5.51 ± 0.08a POO 31.81 ± 0.15a PPO 28.32 ± 0.08a SOO 4.06 ± 0.04a PSO 2.36 ± 0.05a MMM 0.45 ± 0.02a PPP 4.22 ± 0.03a AOO nd SSO 0.90 ± 0.05a OOB nd
nd nd nd 1.72 ± 0.15f 6.68 ± 0.07f 1.51 ± 0.03e nd 18.37 ± 0.12a 6.94 ± 0.08e 1.48 ± 0.06d 38.18 ± 0.15a 2.21 ± 0.09e 15.63 ± 0.03b 1.90 ± 0.05b 2.64 ± 0.08b 1.45 ± 0.04c nd nd 0.35 ± 0.03b 0.78 ± 0.07b nd
nd 1.92 ± 0.10b nd 40.35 ± 0.08b 22.56 ± 0.05b 9.62 ± 0.06c nd 10.15 ± 0.15c 10.79 ± 0.06a 0.47 ± 0.04f 0.88 ± 0.06g 1.17 ± 0.07f 1.61 ± 0.03f 0.40 ± 0.02g nd nd nd nd nd nd nd
nd nd nd 39.58 ± 0.07b 23.37 ± 0.14b 9.81 ± 0.09c nd 5.63 ± 0.11f 9.90 ± 0.06b 0.92 ± 0.04e 3.41 ± 0.07d 3.66 ± 0.06b 1.36 ± 0.03f 1.15 ± 0.08d 0.53 ± 0.10d nd nd nd nd nd 0.68 ± 0.06a
nd 0.98 ± 0.05b nd 28.49 ± 0.13c 25.26 ± 0.06a 14.06 ± 0.04b nd 8.82 ± 0.15d 9.14 ± 0.03c 1.17 ± 0.09d 4.38 ± 0.05c 2.53 ± 0.08e 2.49 ± 0.04d 1.62 ± 0.10c 0.40 ± 0.02d nd nd nd 0.52 ± 0.05a nd nd
1.87 ± 0.03a 9.16 ± 0.06a nd 24.52 ± 0.09d 18.63 ± 0.16c 16.69 ± 0.07a nd 6.96 ± 0.05e 10.61 ± 0.09a 1.84 ± 0.03c 2.32 ± 0.06e 2.91 ± 0.12d 1.73 ± 0.01e 0.93 ± 0.03e 0.68 ± 0.05d 0.65 ± 0.02d nd nd nd nd 0.45 ± 0.03b
nd nd nd 53.16 ± 0.11a 17.91 ± 0.07e 10.14 ± 0.04c nd 3.76 ± 0.05g 7.80 ± 0.08e 0.95 ± 0.03e 3.23 ± 0.10d 1.12 ± 0.05f 1.24 ± 0.06f 0.68 ± 0.03f nd nd nd nd nd nd nd
nd 0.23 ± 0.01c nd 10.17 ± 0.05e 18.19 ± 0.14d 8.77 ± 0.10d nd 15.72 ± 0.05b 8.14 ± 0.05b 5.53 ± 0.02b 14.90 ± 0.07b 3.12 ± 0.15c 11.08 ± 0.09c 0.46 ± 0.02g 1.33 ± 0.05c 1.85 ± 0.03b nd nd 0.51 ± 0.08a nd nd
5.20 ± 0.09a 42.48 ± 0.05a 50.39 ± 0.07a 1.93 ± 0.04g
0.00 5.61 ± 0.09c 29.28 ± 0.10c 64.95 ± 0.04e
0.00 0.87 ± 0.03g 23.19 ± 0.15d 75.86 ± 0.08b
0.00 2.07 ± 0.05e 25.94 ± 0.11d 71.99 ± 0.12c
0.00 2.79 ± 0.03e 29.14 ± 0.07c 67.93 ± 0.05d
0.00 3.42 ± 0.03d 33.07 ± 0.10b 63.46 ± 0.08e
0.00 1.63 ± 0.08f 20.30 ± 0.03e 78.06 ± 0.09a
0.00 7.84 ± 0.05b 32.95 ± 0.09b 59.21 ± 0.09f
TSTAG MSTAG DSTAG TUTAG
Abbreviations: M, myristic acid(C14:0); P, palmitic acid(C16:0); Po, palmitoleic acid(C16:1); S, stearic acid(C18:0); O, oleic acid(C18:1); L, linoleic acid(C18:2); Ln, linolenic acid(C18:3); A, arachidic acid(C20:0); B, behenic acid(C22:0). * Mean values in the same row followed by the same superscript letters are not significantly different (p > 0.05). nd: not detected.
with the formation of peroxides, while T p was indicative of the termination stage (Kowalski, Ratusz, Kowalska, & Bekas, 2004). The DSC thermograms of SuO, SoO, OeO, CnO, SaO, SeO, and GsO exhibited similar flat and short profiles in the initial heating stage (100 to <200 °C), before a sudden increase to maximum heat flow temperatures. For RPO, its DSC spectra began with similar flat, but longer lasting curves (100 to >200 °C), compared to other oils, before gradually increasing to a maximum heat flow temperature. In greater details, the changes in T s , T on and T p for the eight oil samples under each heating rates are shown in Table 2. The T s , T on , and T p shifted towards to higher values as the heating rate increases, which was in agreement with previous studies (Martinez-Monteagudo et al., 2012; Ostrowska-Ligeza et al., 2010). Vecchio et al. (2009) also observed similar increases in peak temperature (T p ) as heating rates were increased to 10 K min1. Although the changes in T s and T on were not reported, DSC profiles were found not to vary at the other heating rates (2.5, 5, and 7.5 K min1). The increase in heating rate from 5 to 15 °C/min led to an increase in T s and T on of around 20 °C for all oil samples, with an even larger magnitude of increment for T p (up to 70 °C for CnO), which indicated that T p was more sensitive than T s and T on . Martinez-Monteagudo et al. also arrived at the same conclusion in their evaluation of the non-isothermal kinetics
of anhydrous milk fat (AMF) rich in conjugated linoleic acid (CLA). They found that T p varied (increased by 35.15 and 36.62 °C for medium and high CLA AMF, respectively) more notably than T s (increased by 22.13 and 22.19 °C for medium and high CLA AMF, respectively) and T on (increased by 25.19 and 28.68 °C for medium and high CLA AMF, respectively), while for low CLA AMF the variance for T s , T on , T p were close, 15.34 for T s , 22.36 for T on , and 21.24 for T p .
3.4. Thermal kinetic analysis The kinetic parameters, namely activation energy (Ea ), preexponential factor (A), and rate constant (k) at 200 °C, which were computed from T s , T on , and T p for each oil, are shown in Table 3. The high correlation coefficients (R2 ) indicated that experimental data was well fitted to the models (Eqs. 4–6). For all oil samples, the Ea values at T p were smaller than that at T s and T on . Similarly, the Ea values at T p were also found to be smaller than that at T s and T on in the work conducted by MartinezMonteagudo et al. (2012). RPO exhibited the highest Ea values, 126.06 kJ/mol at T s , 134.7 kJ/mol at T on , and 91.88 kJ/mol at T p , among all the oils. As expected, the highest Ea values of RPO gave
605
B. Qi et al. / Food Chemistry 194 (2016) 601–607
3.5. Pearson correlation analysis The Pearson correlation coefficients between FA and TAG, and thermal kinetic parameters (Ea and k) at T s , T on , and T p were analyzed. For ease of analysis, the calculated Pearson correlation coefficients (R) were further subjected to create a heat map indicating the degree of correlation (Fig. 3). In general, the k at the three temperatures exhibited correlations with FA and TAG at varying degrees. Positive correlation coefficients were observed between the k at T s and linoleic acid (0.85), PUFA (0.87), and LLL + OLLn (0.93) at the significance level of 0.01, and OLL + OLPo (0.77) and TUTAG (0.81) at the significance level of 0.05. In addition, the k at T s was negatively correlated with oleic acid (0.73), SFA (0.78), MUFA (0.72), SSO (0.76), and MSTAG (0.78) at the significance level of 0.05, and arachidic acid (0.86), PPL (0.85), POO (0.91), SOO (0.90), PSO (0.94), and DSTAG (0.84) at the significance level of 0.01. Both MUFA and PUFA are unsaturated fatty acids despite their different degrees of saturation. However, it is interesting to note that the k value at T s appeared to be positively correlated with PUFA and negatively correlated with MUFA. Furthermore, no significant positive and negative correlations were found between k at T on and FA & TAG. The k at T p showed similar correlation coefficients to k at T s . However, k at T p exhibited a significant positive correlation to LLP, and no negative correlation to arachidic acid (C20:0) and SFA. As for correlations between k and TAGs, the TSTAG did not show any positive correlations with k, which may be due to the low concentrations of TSTAG in oils, except RPO. Our results showed that only fully unsaturated TAG (TUTAG) positively affected the k values at T s and T p , which, for the first time, indicated the interesting relationship between unsaturated TAG and k. The Ea at T s , T on , and T p gave very similar correlations to FA and TAG. In particular, significant positive correlations were found between all the Ea values at T s , T on , and T p , and capric acid (C10:0), myristic acid (C14:0), palmitic acid (C16:0), SFA, MPP, MPL, PPL, SLO, POO, PPO, SOO, PSO, MMM, PPP, TSTAG, MSTAG, and DSTAG. The Ea at T p additionally positively correlated with SSO. Linoleic acid (C18:2), PUFA, and LLL + OLLn were negatively correlated with the Ea at T s and T p , whereas OLL + OLPo were negatively correlated with the Ea at T on and T p , and TUTAG was negatively correlated with the Ea at the three temperatures. Vecchio et al. (2009) evaluated the thermal decomposition of extra virgin
Fig. 2. DSC oxidative profiles of corn oil at five heating rates.
the lowest k values of 0.339 min1 at T s , 0.157 min1 at T on , and 0.001 min1 at T p . The oxidation kinetic parameters of RPO indicated that it was the most stable oil among the tested oil samples. This is not surprising when considering that RPO had the highest percentage of SFA (38.49%), which was more than twice as high compared to the other oils, resulting in RPO being highly stable (Nagaraj, 2009). Due to the abundant amount of PUFA (76.8%, Table 1), which is oxidized very readily (Wijesundera, 2008), SaO showed the lowest Ea values at T s (72.76 kJ/mol) and T p (31.85 kJ/ mol). Besides, the lowest Ea value at T on (69.58 kJ/mol) was found for grapeseed oil. According to the Ea values, the stability scale for the tested oils could be proposed as RPO > SeO > OeO > CnO > SuO > SoO > GsO > SaO at T s , RPO > SoO > SeO > OeO > SaO > SuO > CnO > GsO at T on , and RPO > SeO > SoO > OeO > GsO > SuO > CnO > SaO at T p . The differences observed in the stability of oils may be attributed to their various compositions. To explore the relationship between oil compositions and its thermal attributes, a Pearson correlation analysis was conducted in Section 3.5.
Table 2 The start (T s ) onset (T on ) and maximum heat flow (T p ) temperature of oils under the five heating rates.*
*
Heating rate (°C/min)
RPO
OeO
GsO
SuO
CnO
SoO
SaO
SeO
Start temperature (Ts) 5.0 7.5 10.0 12.5 15.0
203.4 ± 0.3a 206.7 ± 0.8a 211.0 ± 0.1a 214.3 ± 0.2a 219.3 ± 0.8a
178.8 ± 0.6c 187.1 ± 1.1c 189.8 ± 0.1c 195.5 ± 0.7b 199.8 ± 0.4b
161.3 ± 0.1f 168.4 ± 0.4e 172.6 ± 0.3f 178.9 ± 1.0e 182.9 ± 0.6e
167.6 ± 0.8e 173.5 ± 1.3d 179.5 ± 0.6e 182.6 ± 0.4d 174.9 ± 0.6g
172.1 ± 0.3d 176.7 ± 0.1d 183.2 ± 0.4d 188.2 ± 0.9c 190.9 ± 0.2c
167.3 ± 0.7e 171.8 ± 0.8d 177.1 ± 0.5e 183.5 ± 0.2d 186.8 ± 1.1d
159.7 ± 0.5f 162.1 ± 0.1f 171.0 ± 0.6f 177.8 ± 0.7e 179.4 ± 0.4f
184.5 ± 0.2b 190.0 ± 0.4b 197.8 ± 0.9b 194.1 ± 1.4b 202.3 ± 0.5b
Onset temperature (Ton) 5.0 7.5 10.0 12.5 15.0
212.3 ± 0.1a 216.4 ± 0.2a 221.5 ± 0.1a 224.1 ± 0.4a 227.8 ± 0.6a
187.3 ± 0.5b 194.8 ± 0.3b 203.2 ± 0.2b 202.8 ± 0.8b 208.6 ± 1.0b
165.2 ± 0.6d 172.5 ± 0.3d 177.1 ± 1.6e 185.6 ± 0.2e 189.8 ± 0.8d
171.4 ± 1.0c 176.6 ± 1.2c 182.5 ± 0.1d 191.1 ± 0.9d 189.0 ± 0.8d
172.5 ± 0.1c 177.2 ± 0.4c 187.6 ± 0.2c 192.6 ± 0.1d 193.5 ± 0.8c
169.2 ± 0.5c 175.3 ± 0.1c 178.5 ± 0.8e 184.4 ± 0.1e 187.1 ± 0.2e
164.5 ± 0.1d 165.1 ± 0.1e 175.0 ± 0.3f 178.9 ± 0.1f 180.0 ± 0.4f
188.6 ± 0.4b 192.3 ± 0.5b 201.6 ± 0.2b 198.1 ± 0.5c 207.6 ± 0.1b
maximum heat flow temperature (Tp) 5.0 337.1 ± 0.7a 7.5 348.8 ± 0.1a 10.0 354.1 ± 0.4a 12.5 368.3 ± 0.3a 15.0 372.7 ± 0.4a
253.1 ± 0.2b 284.8 ± 0.1b 290.2 ± 0.7b 293.0 ± 0.6b 299.7 ± 0.2b
220.8 ± 0.3e 254.2 ± 0.8c 255.0 ± 0.2d 270.4 ± 1.0d 277.9 ± 0.3d
244.7 ± 0.1c 252.2 ± 0.1d 263.4 ± 0.6c 292.2 ± 0.3e 263.1 ± 0.2e
233.5 ± 0.4d 250.2 ± 0.6d 261.0 ± 0.5c 265.8 ± 0.8e 303.4 ± 0.2b
206.9 ± 0.5f 217.0 ± 0.2f 220.5 ± 0.1e 235.3 ± 0.8f 246.3 ± 0.3f
221.3 ± 0.4e 249.8 ± 0.1d 251.4 ± 0.6d 288.4 ± 0.3c 290.3 ± 0.5c
234.6 ± 0.7d 238.7 ± 0.2e 263.4 ± 0.8c 264.5 ± 0.4e 265.6 ± 0.7e
Mean values in the same row followed by the same superscript letters are not significantly different (p > 0.05).
606
B. Qi et al. / Food Chemistry 194 (2016) 601–607
Table 3 The kinetic parameters (k at 200 °C) calculated from start (T s ) onset (T on ) and peak (T p ) temperature under each heating rate. Ts Ea . RPO OeO GsO SuO CnO SoO SaO SeO
126.06 90.16 79.26 89.16 89.79 85.18 72.76 100.43
Ton A
k 13
2.85 10 8.82 109 1.19 109 1.49 1010 1.23 1010 4.66 109 2.33 108 1.06 1011
0.339 0.970 2.092 2.118 1.487 1.821 2.147 0.857
2
R
Ea .
0.9789 0.9922 0.9949 0.7468 0.9919 0.9865 0.9671 0.9264
134.7 89.23 69.58 82.62 74.23 96.93 86.05 93.21
Tp A
k 14
1.18 10 4.31 109 6.92 107 1.83 109 1.77 105 1.01 1011 7.57 109 1.31 1010
0.157 0.601 1.427 1.37 1.117 1.985 2.366 10.664
2
R
0.9941 0.9800 0.9890 0.9648 0.9777 0.9929 0.9445 0.9199
Ea . 91.88 52.70 44.37 40.81 36.21 52.84 31.85 56.74
A 7
1.85 10 4.82 104 6.85 103 6.01 103 2.32 103 2.05 105 9.77 102 2.38 105
k
R2
0.001 0.073 0.216 0.187 0.232 0.300 0.297 0.129
0.9851 0.9330 0.9635 0.7312 0.9300 0.9616 0.9626 0.9272
Fig. 3. Heat map for the Pearson correlation analysis.
olive oils and they also found positive correlations between Ea calculated from II deconvoluted peak, palmitic acid (C16:0, 0.74), SFA (0.78), and MSTAG (0.65) at the significance level of 0.01. The increased percentage of saturated CH2 carbons in oil molecules would improve the resistance of oil against the initial thermal breakdown, leading to a higher Ea value (Adhvaryu et al., 2000). However, our results showed that the Ea values had no statistical correlation with stearic acid (C18:0), arachidic acid (C20:0), and behenic acid (C22:0), although their contents in oils were close to capric acid (C10:0) and myristic acid (C14:0). Additionally, it is worthy to point out that although k values at T s and T p were negatively correlated with MUFA, the Ea values showed no significant correlation with MUFA. The TSTAG showed a significantly positive correlation with Ea values. However, MSTAG and DSTAG, which contained two and one unsaturated fatty acids in their TAG
structures respectively, were also statistically correlated with Ea values, with only TAGs containing three unsaturated fatty acids (TUTAG) negatively affecting Ea values.
4. Conclusions All oil samples exhibited varying DSC thermograms, indicating different thermal oxidation attributes, as affected by their compositions. The Pearson correlation analysis showed that (i) the k value at T s and T p were positively correlated with PUFA while being negatively correlated with MUFA; (ii) the k values at T s and T p were found to have unique positive correlations with TUTAG; and (iii) the Ea values at T s , T on , and T p were not positively correlated with SFA as well as TSTAG, whereas, they were instead positively
B. Qi et al. / Food Chemistry 194 (2016) 601–607
correlated with MSTAG and DSTAG. This study could be used as a possible tool to predicate the thermal oxidation properties of oil based on its FAs and TAGs composition. Acknowledgements The authors would like to acknowledge the support for this study by the National Natural Science Foundation of China (research Grant number: 31430067), the National Key Technology Research and Development Program of the Ministry of Science and Technology of China (research Grant number: 2014BAD22B01-02), Educational Commission of Heilongjiang Province of China (research Grant number: 12531049), Natural Science Foundation of Heilongjiang Province of China (research Grant number: QC2013C014), and the Key Laboratory of Soybean Biology of Chinese Education Ministry. References Adhvaryu, A., Erhan, S. Z., Liu, Z. S., & Perez, J. M. (2000). Oxidation kinetic studies of oils derived from unmodified and genetically modified vegetables using pressurized differential scanning calorimetry and nuclear magnetic resonance spectroscopy. Thermochimica Acta, 364, 87–97. Caponio, F., Chiavaro, E., Paradiso, V. M., Paciulli, M., Summo, C., Cerretani, L., & Gomes, T. (2013). Chemical and thermal evaluation of olive oil refining at different oxidative levels. European Journal of Lipid Science and Technology, 115, 1146–1154. Cerretani, L., Bendini, A., Rinaldi, M., Paciulli, M., Vecchio, S., & Chiavaro, E. (2012). DSC evaluation of extra virgin olive oil stability under accelerated oxidative test: Effect of fatty acid composition and phenol contents. Journal of Oleo Science, 61, 303–309. Cerretani, L., Maggio, R. M., Barnaba, C., Toschi, T. G., & Chiavaro, E. (2011). Application of partial least square regression to differential scanning calorimetry data for fatty acid quantitation in olive oil. Food Chemistry, 127, 1899–1904. Choe, E., & Min, D. B. (2006). Mechanisms and factors for edible oil oxidation. Comprehensive Reviews in Food Science and Food Safety, 5, 169–186. Cunha, S. C., & Oliveira, M. (2006). Discrimination of vegetable oils by triacylglycerols evaluation of profile using HPLC/ELSD. Food Chemistry, 95, 518–524. Dijkstra, A. J., & Segrers, J. C. (2007). Production and refining of oils and fats. In F. D. Gunstone, J. L. Harwood, & A. J. Dijkstra (Eds.), The lipid handbook (3rd ed., pp. 143–251). CRC Press. Flynn, J. H., & Wall, L. A. (1966). A quick direct method for determination of activation energy from thermogravimetric data. Journal of Polymer Science Part B-Polymer Letters, 4, 323–328. Holcapek, M., Lisa, M., Jandera, P., & Kabatova, N. (2005). Quantitation of triacylglycerols in plant oils using HPLC with APCI-MS, evaporative lightscattering, and UV detection. Journal of Separation Science, 28, 1315–1333.
607
ISO, 1990 ISO 5508:1990, Animal and vegetable fats and oils, Analysis by gas chromatography of methyl esters of fatty acids, American National Standards Institute (ANSI), 1990. Kowalski, B., Ratusz, K., Kowalska, D., & Bekas, W. (2004). Determination of the oxidative stability of vegetable oils by differential scanning calorimetry and rancimat measurements. European Journal of Lipid Science and Technology, 106, 165–169. Lerma-Garcia, M. J., Lusardi, R., Chiavaro, E., Cerretani, L., Bendini, A., Ramis-Ramos, G., & Simo-Alfonso, E. F. (2011). Use of triacylglycerol profiles established by high performance liquid chromatography with ultraviolet-visible detection to predict the botanical origin of vegetable oils. Journal of Chromatography A, 1218, 7521–7527. Li, Y., Zhang, Y., Wang, M., Jiang, L., & Sui, X. (2013). Simplex-centroid mixture design applied to the aqueous enzymatic extraction of fatty acid-balanced oil from mixed seeds. Journal of the American Oil Chemists’ Society, 90, 349–357. Maggio, R., Barnaba, C., Cerretani, L., Paciulli, M., & Chiavaro, E. (2014). Study of the influence of triacylglycerol composition on DSC cooling curves of extra virgin olive oil by chemometric data processing. Journal of Thermal Analysis and Calorimetry, 115, 2037–2044. Martinez-Monteagudo, S. I., Saldan, M. D. A., & Kennelly, J. J. (2012). Kinetics of nonisothermal oxidation of anhydrous milk fat rich in conjugated linoleic acid using differential scanning calorimetry. Journal of Thermal Analysis and Calorimetry, 107, 973–981. Nagaraj, G. (2009). Oilseeds: properties, products, processing and procedures. New Delhi: New India Publishing. Ostrowska-Ligeza, E., Bekas, W., Kowalska, D., Lobacz, M., Wroniak, M., & Kowalski, B. (2010). Kinetics of commercial olive oil oxidation: Dynamic differential scanning calorimetry and Rancimat studies. European Journal of Lipid Science and Technology, 112, 268–274. Ozawa, T. (1965). A new method of analyzing thermogravimetric data. Bulletin of the Chemical Society of Japan, 38, 1881–1886. Shahidi, F., & Zhong, Y. (2010). Lipid oxidation and improving the oxidative stability. Chemical Society Reviews, 39, 4067–4079. Tan, C. P., & Man, Y. B. C. (2000). Differential scanning calorimetric analysis of edible oils: Comparison of thermal properties and chemical composition. Journal of the American Oil Chemists Society, 77, 143–155. Tan, C. P., Man, Y. B. C., Selamat, J., & Yusoff, M. S. A. (2001). Application of arrhenius kinetics to evaluate oxidative stability in vegetable oils by isothermal differential scanning calorimetry. Journal of the American Oil Chemists Society, 78, 1133–1138. Tan, C. P., & Nehdi, I. A. (2014). DSC analysis of vegetable oils: Relationship between thermal profiles and chemical composition. In E. Chiavaro (Ed.), Differential scanning calorimetry: Applications in fat and oil technology (pp. 3–26). CRC Press. Vecchio Ciprioti, S., & Chiavaro, E. (2014). Introduction. In E. Chiavaro (Ed.), Differential scanning calorimetry: Applications in fat and oil technology (pp. xv–xvii). CRC Press. Vecchio, S., Cerretani, L., Bendini, A., & Chiavaro, E. (2009). Thermal decomposition study of monovarietal extra virgin olive oil by simultaneous thermogravimetry/ differential scanning calorimetry: Relation with chemical composition. Journal of Agricultural and Food Chemistry, 57, 4793–4800. Wijesundera, C. (2008). The influence of triacylglycerol structure on the oxidative stability of polyunsaturated oils. Lipid Technology, 20, 199–202. Zeb, A. (2012). Triacylglycerols composition, oxidation and oxidation compounds in camellia oil using liquid chromatography–mass spectrometry. Chemistry and Physics of Lipids, 165, 608–614.