Journal of Oleo Science Copyright 2007 by Japan Oil Chemists Society A Rapid Method for Trans-Fatty Acid Determination Using a Single Capillary GC Seiichi Shirasawa, Akiko Sasaki, Yasue Saida and Chiemi Satoh Division of Analytical Science, Research Laboratory, The Nisshin OilliO Group, Ltd. (1-Banchi, Shinmei-cho, Yokosuka-shi, Kanagawa, 239-0832, JAPAN) Abstract: A rapid analytical method for trans-fatty acid determination using a single capillary column was developed, and an isothermal condition for separating cis/trans-isomers with 30- to 60-m columns depending on the contents of the trans-fatty acids was established. Under the established conditions, analysis of trans-fatty acids was completed in 20 min for non-hydrogenated oils low in trans-fatty acids and in 40 min for hydrogenated oils rich in trans-fatty acids. The results were virtually the same as those obtained by the AOCS official method with a 100-m column. By correcting with molecular weights of fatty acid methyl esters and free fatty acids, it was confirmed that the analytical data were the same as those obtained by quantitative analysis using an internal standard. It is anticipated that this proposed method can be applied in a similar way to the AOCS official method, particularly in quality control processes. Key words: trans-fatty acid, fatty acid analysis, capillary column, gas chromatography, stationary phase, hydrogenated oil 1 INTRODUCTION Trans-fatty acids are known to increase serum LDLcholesterol and decrease HDL-cholesterol 1,2). Consequently, excessive intake of trans-fatty acids increases the risk of cardiovascular disease 1,2). Therefore, much attention is focused on the consumption of trans-fatty acids worldwide, and regulations or compulsory claims for trans-fatty acids are instituted in Denmark 3), the USA 4), and Canada 5). This raised awareness of trans-fatty acids requires precise and convenient methods for analysis of commercial products. Analytical methods currently available for transfatty acids include gas chromatography (GC), infrared spectroscopy, and AgNO 3 -TLC-GC 6). Among these methods, GC has recently been used predominantly because of its convenience and sensitivity. Bis cyanopropyl polysiloxane or bis cyanopropylsiloxane polysilphenylene stationary phases are usually employed in GC for analyzing transfatty acids in correspondence with the need for high resolution among fatty acid methyl ester (FAME) peaks 7,8). : GC, gas chromatograph; FAME, fatty acid methyl ester; MS, mass spectrometer; AOCS, the American Oil Chemists Society; AOAC, the Association of Official Agricultural Chemists; TLC, thin layer chromatography; ECL, equivalent chain length There are a number of official analytical methods using these stationary phases, such as the AOAC 9) and AOCS methods 10). The AOCS official GC method has been revised in conjunction with the enforcement of compulsory indication in the USA 11). Although this method is highly precise, it has problems such as a longer analysis time because of its column length of 100 m, difficulty in peak identification due to the bis cyanopropyl polysiloxane stationary phase, and greater economic burden due mainly to use of an expensive standard, triheneicosanoin. Thus, a more convenient GC method is desirable, particularly for the purpose of quality control. Therefore, an analytical GC method using a shorter column without the standard reagent was developed, and its accuracy was confirmed by comparison with the AOCS method. 2 EXPERIMENTAL 2 1 All test samples (soybean oil, rapeseed oil, palm oil, hydrogenated soybean oil, hydrogenated palm oil, and hydrogenated cottonseed oil) were obtained as commercially available purified edible oils. C18:2 (Cat. No. 47791) and Correspondence to: Seiichi Shirasawa, Division of Analytical Science, Research Laboratory, The Nisshin OilliO Group, Ltd., 1-Banchi, Shinmei-cho, Yokosuka-shi, Kanagawa 239-0832, JAPAN E-mail: s-shirasawa@nisshin-oillio.com Accepted September 19, 2006 (received for review July 24, 2006) Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 online http://jos.jstage.jst.go.jp/en/ 53
S. Shirasawa, A. Sasaki, Y. Saida et al. C18:3 (Cat. No. 47792) FAME isomer mix reagents for identification of the retention time were purchased from Supelco Inc. (Bellefonte, PA). Heptadecanoic acid (Tokyo Kasei, Tokyo) was used as an internal standard. Other reagents used were all of reagent grade purchased from Wako Pure Chemical Industries, Ltd., Tokyo. 2 2 Analyses of FAME were carried out on an HP6890 gas chromatographic system equipped with a flame-ionization detector (FID) and a split injector (Agilent Technologies, Palo Alto, CA). Three types of capillary column were used for comparison: two TC70 TM capillary columns (30 m and 60 m 0.25 mm, 0.25 mm thickness; GL-Science, Tokyo), and a SP2560 TM capillary column (100 m 0.25 mm, 0.20 mm thickness; Supelco Inc. Bellefonte, PA). The GC-MS device used for peak identification was an HP5973 MSD (Agilent Technologies, Palo Alto, CA). 2 3 Methyl-esterification of samples used in the analyses was performed by the BF 3 -MeOH method 12) after alkaline hydrolysis. To 20 mg of sample oils were added 2 ml of 0.5 mol/l NaOH-methanol solution, and the mixture was heated at 100 for 7 min. After cooling, 3 ml of 14% BF 3 - MeOH reagent was added, and the vessel was sealed and heated at 100 for 5 min. After cooling, 2 ml of hexane and 7 ml of saturated NaCl solution were added, followed by a thorough shaking. The resulting hexane layer was used as a sample solution for GC. Two milligrams of internal standard was added as a chloroform solution before esterification, and the solvent was removed under nitrogen. 2 4 FAME peaks were identified by comparing their retention times with those of predetermined FAME standards and those of the AgNO 3 -TLC fraction of hydrogenated soybean oil. GC-MS analyses were performed under conditions identical to that of GC-FID when necessary. MS spectra were obtained at an ionization energy of 70eV, range width m/z 29-700, interface temperature 240, and ion source temperature 230. The contents of trans-fatty acids were calculated both by composition analysis using an integrator and quantitative analysis using heptadecanoic acid as an internal standard. The following FAME peaks were identified as trans-isomers: C16:1t, C18:1t, C18:2t, C18:3t, and C20:1t. Trans-fatty acids were expressed as percentages rounded to the first decimal place. 3 RESULTS The GC conditions were set based on analysis at constant temperature that allowed a high peak resolution. shows a comparison of the GC conditions applied. With respect to TC70 TM columns, the GC conditions were set based on the best resolution for C18:1 isomers by comparison with chromatograms of the AgNO 3 -TLC fraction 13) of hydrogenated soybean oil obtained under multiple oven temperature conditions and optimal carrier flow rates as determined from the Van Deemter curve 14). The analyses Table 1 Methods Examined. Method 1 Method 2 Method 3 1) Column TC-70 TM (GL-Science) SP-2560 TM (Supelco Inc.) Stationary phase bis cyanopropylsiloxane polysilphenylene bis cyanopropyl polysiloxane Length 30 m 60 m 100 m Internal diameter 0.25 mm 0.25 mm Film thickness 0.25 mm 0.20 mm Temperature 190 180 2) Inj/Det 250 /260 250 /250 Carrier gas 1 ml/min (Helium) 1 ml/min (Helium) Split ratio 100:1 100:1 Inject volume 3) 1 ml 1 ml Runtime 18 min 40 min 74 min 1) Control test : AOCS Ce1h-05 2) 180 (60 min) - (10 /min) - 220 (10min) 3) 1% oil in hexane 54
GC Analysis of Trans-fatty Acid were continued until C24:1 methyl ester was eluted. Analysis times required for elution of C24:1 methyl ester were 20 and 40 min for TC70 TM (30 m) and TC70 TM (60 m), respectively. Analyses with SP2560 TM were performed according to AOCS Ce1h-05 11). However, SP2560 TM required as long as 95 min for elution under isothermal GC conditions. Therefore, a temperature program was introduced after isothermal analysis (180 ) for 45 min. FAME peaks were primarily identified by comparing the retention times with those of the FAME standards analyzed concomitantly using GC-MS. In confirmation with GC-MS, carbon number and unsaturation degree were determined from molecular ions, and cis/trans geometrical isomers were identified using the characteristic elution patterns of the stationary phase for reference. shows the chromatograms obtained under the three GC conditions. summarizes the experimental ECL values of peaks shown in. ECL values were determined using the retention times of C16:0 (peak a) and C18:0 (peak d) in each method. In methods 1 and 2, which used the same stationary phase, ECL values were nearly equal. However, because peak elution was fast, C18:3c (peak j) and C20:0 (peak k) overlapped in Method 1. In Method 3, C18:3t (peak i) and C20:1t (peak l) overlap was confirmed by GC-MS analyses. shows the result of analyses repeated six times for rapeseed oil with a low trans-fatty acid content and for hydrogenated soybean oil with a high trans-fatty acid content. All three GC conditions showed a good coefficient of variation (C.V.; 0.1-1.7). This result indicated that peak separation and integration were both satisfactorily. shows the results of GC analyses of the transfatty acid contents of seven edible oils with different transfatty acid contents. For the sample oils A, B and C, equivalent values were obtained among the three methods. It was clear that the same value was reproducible using the TC70 TM (30 m) column, which allowed rapid analysis (Method 1), as that obtained using the SP2560 TM column, which required a long analytical time (Method 3), when samples were non-hydrogenated oils with a low trans-fatty acid content and a narrow distribution range of positional Fig. 1 The C16 - C20 Regions of Gas Chromatograms of FAMEs from Rapeseed Oil (Solid Line) and Hydrogenated Soybean Oil (Dotted Line). Peak Identification: a, C16:0; b, C16:1t; c, C16:1c; d, C18:0; e, C18:1t; f, C18:1c; g, C18:2t; h, C18:2c; i, C18:3t; j, C18:3c; k, C20:0; l, C20:1t; and m, C20:1c. Table 2 Equivalent Chain Length (ECL) Values of FAMEs. ECL Peak 1) FAME Method 1 Method 2 Method 3 a 16:0 16.00 16.00 16.00 b 16:1t 16.22 16.27-16.41 16.34-16.60 c 16:1c 16.50 16.49 16.72 d 18:0 18.00 18.00 18.00 e 18:1t 18.15-18.38 18.15-18.46 18.32-18.72 f 18:1c 18.55-18.87 18.53-18.93 18.80-19.24 g 18:2t 19.00-19.24 18.98-19.23 19.33-19.86 h 18:2c 19.42-19.52 19.36-19.51 19.99-20.15 i 18:3t 19.87-20.28 19.82-20.29 20.84-21.30 j 18:3c 20.44 20.43 21.47 k 20:0 20.48 20.52 20.51 l 20:1t 20.70-20.83 20.75-20.91 20.84-21.30 m 20:1c 21.08 21.10 21.36 1) Peak a - m were indicated in Fig.1. 55
S. Shirasawa, A. Sasaki, Y. Saida et al. Rapeseed Oil Hydrogenated Oil Table 3 All data were described as area%, n=6. 1) C.V., Coefficient of variation. Precision Data. Method 1 Method 2 Method 3 Average C.V. 1) Average C.V. 1) Average C.V. 1) C16:1t 0.1 4.3 0.1 9.5 - - C18:1t - - 0.1 8.9 - - C18:2t 0.3 2.4 0.3 4.3 0.3 5.2 C18:3t 1.8 0.4 1.7 0.9 1.8 1.2 C20:1t - - - - - - Total 2.2 0.6 2.2 0.5 2.1 1.7 Method 1 Method 2 Method 3 Average C.V. 1) Average C.V. 1) Average C.V. 1) C16:1t 0.2 2.0 0.1 7.0 - - C18:1t 44.0 0.4 45.3 0.2 48.5 0.1 C18:2t 0.7 21.9 0.8 1.1 1.2 7.4 C18:3t 0.1 22.0 0.1 12.8 - - C20:1t 0.2 7.0 0.2 10.0 - - Total 45.2 0.7 46.5 0.2 49.7 0.1 Table 4 Trans-fatty Acid Contents by Three Methods. Sample A Sample B Sample C Method 1 Method 2 Method 3 Method 1 Method 2 Method 3 Method 1 Method 2 Method 3 C16:1t 0.1 0.1-0.1 0.1-0.1 0.2 - C18:1t 0.1 0.1-0.1 0.1-0.4 0.4 0.5 C18:2t 0.3 0.3 0.4 1.0 1.0 1.0 0.4 0.4 0.4 C18:3t 1.8 1.8 1.8 1.4 1.4 1.3-0.1 - C20:1t - - - - - - - - - Total 2.3 2.3 2.2 2.6 2.6 2.3 0.9 1.1 0.9 Sample D Sample E Sample F Sample G Method 1 Method 2 Method 3 Method 1 Method 2 Method 3 Method 1 Method 2 Method 3 Method 1 Method 2 Method 3 C16:1t 0.1 0.2-0.3 0.3 0.1 0.1 0.1-0.1 0.1 - C18:1t 16.9 16.9 18.6 26.8 28.9 30.7 35.9 40.2 39.1 45.9 45.4 48.4 C18:2t 0.3 0.3 0.1 1.9 2.0 2.1 3.1 1.1 3.7 0.7 0.3 0.8 C18:3t - - - - 0.1-0.1 0.1 - - - - C20:1t - - - - - - 0.3 0.1-0.2 0.2 - Total 17.3 17.4 18.7 29.0 31.3 32.9 39.5 41.6 42.8 46.9 46.0 49.2 All data were described as area %. Sample A, rapeseed oil; B, soybean oil; C, palm oil; D, hydrogenated palm oil; E, hydrogenated cottonseed oil; F and G, hydrogenated soybean oil. 56
GC Analysis of Trans-fatty Acid Table 5 Quantitative Analysis and Composition Analysis. Quantitation 1) (g/100g) Composition analysis 2) (%) Corrected 3) (g/100g) Soybean oil 2.3 2.6 2.5 Rapeseed oil 2.2 2.3 2.2 Hydrogenated Oil A 16.3 17.4 16.6 Hydrogenated Oil B 29.1 31.2 29.7 Hydrogenated Oil C 37.5 40.1 38.2 Hydrogenated Oil D 43.7 46.7 44.5 GLC condition of Method 2 was applied for this examination. Hydrogenated Oil A, hydrogenated palm oil; B, hydrogenated cottonseed oil; C and D, hydrogenated soybean oil. 1)Quantitative analysis using C17:0 FFA internal standard. Calculation : sum of trans area counts weight of internal standard/c17:0 area count/weight of test sample 100. 2)Composition analysis using FAME peak area size on chromatogram (area%) 3)Corrected by a factor for conversion of FAME to FFA (area% FFA MW/FAME MW). Example: Conversion factor for calculation of C18:1t is, C18:1 FFA MW, 282.5/C18:1 FAME MW, 296.5 = 0.953. isomers. Some degrees of deviation were found for the sample oils D to G, which contained higher levels of trans-fatty acids (all samples were hydrogenated). Particularly, the deviations became greater with an increasing trans-fatty acid content, as shown in Methods 1 and 3. However, the extent of deviation between Methods 2 and 3 was less than 10%. Based on these results, it was clear that a column of 60 m or longer is required for analysis of hydrogenated oils, in contrast to the case of non-hydrogenated oils, for which a column of approximately 30 m is satisfactorily applicable. As shown in, in the analysis of sample oils F and G, C20:1t was detected using both Methods 1 and 2, and C18:3t and C20:1t were independently detected when the TC70 TM column was employed. However, these peaks were not detected by Method 3. It seems likely that C18:3t and C20:1t were indistinct on the SP2560 TM column, because these peaks and C20:1c substantially overlapped, and/or there is a possibility that these peaks became broad due to the length of the column, 100 m in this case. Quantitative results obtained using the standard substance (heptadecanoic acid) and results from the established analysis using Method 2 were compared ( ), and it was found that the former were lower than the latter. This was attributed to the difference in molecular weight between free fatty acids and FAMEs; the quantitative analytical data were expressed as free fatty acids, whereas the analytical data were expressed as FAMEs. Therefore, correction for molecular weight was performed by multiplying the peak areas from each FAME detected in the chromatogram by a conversion factor (molecular weight of free fatty acid/molecular weight of FAME). As a result of this correction, the values from the currently established method and those from quantitative analyses were virtually equivalent in all samples. 4 DISCUSSION In this study, experiments were performed primarily using columns with a bis cyanopropylsiloxane polysilphenylene stationary phase. This stationary phase characteristically elutes each FAME in the order C18:3t, C18:3c, C20:0 and C20:1 7). Therefore, the disadvantage of possibly mistaking C18:3t for C20:1, which might occur when the bis cyanopropyl polysiloxane stationary phase is used, can be eliminated, and the presence of C20:1t can be definitively ascertained. The most ideal analytical method would be one that has both convenient operating conditions and easy data interpretation, so that variations in operator skill can be eliminated. In this regard, the analytical conditions employed in our new method are considered to be useful. Nevertheless, the peaks of C20:0 and C18:3c may sometimes overlap depending on the column and operating conditions using this stationary phase. Even so, however, calculation of the trans-fatty acid content does not come into question. Since confirmation of saturated fatty acids is as 57
S. Shirasawa, A. Sasaki, Y. Saida et al. important as that of trans-fatty acids, it is very likely that simultaneous analysis will be required. It is therefore necessary to ensure separation of these peaks using appropriate reference standards. In this study we compared the results obtained by compositional and quantitative analyses and showed that the deviations between the two procedures could be minimized after molecular weight adjustment. Quantitative analysis is basically employed for labeling the ingredient contents of products, but a more convenient and reliable method is required for routine process control or product management. The present results suggest that it is desirable to introduce composition analysis in addition to quantitative analysis without complex operations and difficulties in managing quantitative standards. The method we have established in this study appears to satisfy these requirements from the standpoint of quality control. ACKNOWLEDGMENT We are grateful to Reiko Ogawado for technical assistance in this study. 1. Zock, P.L.; Katan, M.B. Trans Fatty Acids, Lipoproteins, and Coronary Risk. Can. J. Physiol. Pharmacol. 75, 211-216 (1997). 2. Mozaffarian, D.; Katan, M.B.; Ascherio, A.; Stampfer, M.J.; Willett, W.C. Trans Fatty Acids and Cardiovascular Disease. N. Engl. J. Med. 354, 1601-1613 (2006). 3. Stender, S.; Dyerberg, J. The Influence of Trans Fatty Acids on Health. 4th edn. The Danish Nutrition Council. pp. 45-51 (2003). 4. Department of Health and Human Services, Food and Drug Administration, Food Labeling: Trans Fatty Acids in Nutrition Labeling, Nutrient Content Claims, and Health Claims. Federal Register. 68, No. 133. July 11. 2003; 21 CFR Part 101. 5. Canadian Food Inspection Agency, 2003 Guide to Food Labelling and Advertising. http://www.inspection.gc.ca/english/fssa/labeti/ guide/toce.shtml 6. Ratnayake, W.M.N. Analysis of Dietary Trans Fatty Acids. J. Oleo Sci. 50, 339-352 (2001). 7. Duchateau, G.S.M.J.E.; Oosten, H.J.V.; Vasconcellos, M.A. Analysis of cis- and trans-fatty Acid Isomers in Hydrogenated and Refined Vegetable Oils by Capillary Gas-Liquid Chromatography. J. Am. Oil Chem. Soc. 73, 275-282 (1996). 8. Ratnayake, W.M.N.; Plouffe, L.J.; Pasquier, E.; Gagnon, C. Temperature-Sensitive Resolution of cis- and trans- Fatty Acid Isomers of Partially Hydrogenated Vegetable Oils on SP2560 and CP-Sil 88 Capillary Columns. J. AOAC Int. 85, 1112-1118 (2002). 9. Fat (Total, Saturated, and Unsaturated) in Food. AOAC Official Method 996.06, (2001). 10. Determination of cis- and trans- Fatty Acids in Hydrogenated and Refined Oils and Fats by Capillary GLC. AOCS Official Method Ce 1f-96 (2002). 11. Determination of cis-, trans-, Saturated, Monounsaturated and Polyunsaturated Fatty Acids in Vegetable or Non-ruminant Animal Oils and Fats by Capillary GLC, AOCS Official Method Ce 1h-05 (2005). 12. Preparations of Methyl Esters of Fatty Acids. AOCS Official Method Ce 2-66 (1997). 13. Mono-trans Fatty Acids (Silver-Ion Thin Layer Chromatography-Gas Chromatography). Standard Methods for the Analysis of Fats, Oils and Related Materials. (Japan Oil Chem. Soc. ed.) zan8-2003 (2003). 14. Ingraham, D.F.; Shoemaker, C.F.; Jennings, W. Computer Comparisons of Variables in Capillary Gas Chromatography. J. HRC & CC. 5, 227-235 (1982). 58