Results and Discussion
The dependence of 100% BCS phase selectivity on elution temperature may be exploited to precisely tailor the separation of all FAME that need to be quantified for nutritional labeling purposes [10]. Raising the elution temperature increases the relative retention of FAME with more DB relative to those with fewer DB. Also, with a lower magnitude, it increases the retention of FAME with DB in cis configuration relative to their analogs with same DB but in transconfiguration. PHO and RVO are preferentially analyzed by isothermal elution at 180°C because it provides suitable resolution oft 13/t 14-18:1 from c 9-18:1, and the elution ofc 11-20:1 in between t 9,c 12,c 15-18:3 and 18:3n-3 [4], but the separation of SCFA is sacrificed. Lowering the elution temperature to separate SCFA would cause a progressively stronger overlap between the trans -18:1 and cis -18:1 and increase retention of c 11-20:1 with respect to 18:3. Golayet al. [7,8] developed a temperature gradient method capable of simultaneously separating SCFA and 18:1 TFA contained in milk products, infant formula, and adult/pediatric nutritional formula. In this study, the applicability of this method to the analysis of PHO and RVO was tested using a 2:1 blend of PHO and soybean oil. Starting the temperature program at 60°C provided the separation of SCFA, then the rapid temperature ramp to 165°C followed by a single ramp to 225°C provided the elution of trans -18:1 ahead of c 9-18:1 (Fig. 1A). However, the elution of all C18 FAME with a single ramp to 225°C as prescribed in the original method caused the overlap oft 9,c 12,c 15-18:3 with c 11-20:1. This inconvenience may be resolved by tailoring the temperature program more precisely. The addition of a temperature plateau at 184°C (method 1) rephased the elution of c 11-20:1 in exchange for a minor loss of separation of t 13/t 14-18:1 from c 9-18:1 (Fig. 1B). The temperature of this plateau (184°C) may be finely tuned to account for minor column to column variability: the elution of c 11-20:1 may be moved toward t 9,c 12,c 15-18:3 by increasing this temperature, or toward 18:3n-3 by lowering it.
If a greater separation of t 13/t 14- and c9-18:1 is necessary, the temperature of the plateau may be increased 190°C (Fig 1C). This change improved the separation of trans -18:1, but it caused the partial overlap of t 9,c 12,c 15-18:3 withc 11-20:1. This co-elution may be resolved by adding a negative temperature gradient to 175°C after the elution of trans -18:1 (Fig. 1D, method 2). If necessary, the time that the column is maintained at 190°C may be slightly adjusted to optimize the separation of c 11-20:1 in between t 9,c 12,c 15-18:3 and 18:3n-3 (i.e., c 11-20:1 may be eluted closer tot 9,c 12,c 15-18:3 by increasing the time at 190°C). After the elution of 18:3n-3 the elution temperature is increased to 225°C to elute the remaining analytes. The retention time of 21:0 was used as reference point to start the final temperature gradient to 225°C, which provided the elution of all FAME contained in the Supelco 37 mix in 71 minutes. To increase the reproducibility of separations the samples were injected back-to-back with the column re-equilibration time set to 5 min, and the column was conditioned 25 min at 225°C prior to the first injection.
The negative temperature program modification (method 2) was applied to the analysis of the Supelco 37 FAME mix, a canola oil, a PHO, a shortening blend (containing PHO and palm oil), and butter (BCR 632A) (Fig 2, 3, 4). The proposed modifications provided the separation of all PUFA contained in butter (Fig 4); however, the final ramp may be more precisely tuned if the resolution of other specific PUFA eluting after 18:3n-3 is necessary. While this adaptation of the Golay et al.method adds some complexity to the temperature program, the resolution of the overlap between c 11-20:1 andt 9,c 12,c 15-18:3 provides a remarkable gain in method versatility by providing accurate measurement of all TFA occurring in PHO and RVO. If the GC is not capable of negative temperature gradients, or more simplicity is desired, the introduction of the 184°C plateau is a suitable compromise.