Comparison of reversed-phase liquid chromatography-mass spectrometry with electrospray and atmospheric pressure chemical ionization for analysis of dietary tocopherols.

ESI and APCI ionization techniques in both negative and positive ion modes were evaluated for simultaneous LC-MS analysis of the four tocopherol homologues (alpha, beta, gamma and delta). The ESI and APCI ionization of tocopherols in positive ion mode was not efficient and proceeded via two competitive mechanisms, with the formation of protonated pseudo-molecular ions and molecular ions, which adversely influenced the repeatability of MS signal. Ionization in negative ion mode in both ESI and APCI was more efficient as it only produced target deprotonated pseudo-molecular ions. The APCI in negative ion mode showed larger linearity range, lower detection limits and was less sensitive to the differences in chemical structure of analytes and nature of applied solvents than negative ion ESI. Negative ion APCI was, therefore, chosen for the development of LC-MS method for simultaneous determination of the four tocopherols in foods. A baseline separation of the tocopherols was achieved on novel pentafluorophenyl silica-based column Fluophase PFP. The use of methanol-water (95:5, v/v) as a mobile phase was preferable to the use of acetonitrile-water due to considerable gain in MS signal. The limits of quantifications were 9 ng/mL for alpha-tocopherol, 8 ng/mL for beta- and gamma- and 7.5 ng/mL for delta-tocopherol when 2 microL was injected. This method was successfully applied to determination of tocopherols in sunflower oil and milk.

Applicability of alkyl-bonded ultra-pure silica stationary phases for gradient reversed-phase HPLC of folates with conventional and volatile buffers under highly aqueous conditions.

Applicability of several alkyl-bonded silica stationary phases was tested for gradient RP-HPLC of folates under highly aqueous conditions. High retention of folates was achieved on alternative phases with enhanced polarity and classical phases with higher carbon content. Phases exhibiting polar secondary interactions were found to provide better selectivity for late-eluting folates, whereas selectivity for early-eluting folates was mostly dependent on hydrophobic interactions. Best selectivity in phosphate buffered mobile phase was achieved on polar-endcapped silica phases (Aquasil C18 and HyPurity Aquastar) followed by alternative Atlantis dC18. Classical phases exhibited poorer separation of 10-formyl-folic acid and 5-formyl-tetrahydrofolate, but it could be considerably improved by increasing the buffer pH. Strong secondary interactions of ion-exchange character on polar-embedded phases resulted in marked peak deterioration, loss of recovery and dramatic changes in retention behaviour for early- and late-eluting folates when changing the mobile phase composition and pH. Therefore, polar-embedded phases such as HyPurity Advance were found to be unsuitable for separating folates. Stationary phases exhibited peak deterioration when using volatile buffer of low ionic strength. Better results were obtained with classical phases, whereas alternative phases showed not only peak deterioration but also a decrease in recovery and poorer selectivity due to increased secondary interactions in volatile buffer.

A computationally feasible quantum chemical model for 13C NMR chemical shifts of PCB-derived carboxylic acids.

Two quantum chemical models have been derived for the prediction of 13C NMR chemical shifts of novel PCB acids obtained from PCBs by catalytic carbonylation. 13C isotropic shielding constants were calculated employing the GIAO (gauge-independent atomic orbital) method with density functional theory (DFT). The best results were obtained by cluster calculations, which took the solvent effects into account properly. In this approach, a solvent molecule (acetone) was attached by a hydrogen bond to every hydrogen atom present in a PCB acid, and the geometry of the molecular cluster was optimized employing the AM1 method. For 158 chemical shifts, the cross-validated standard error was 2.8 ppm and the cross-validated correlation coefficient was 0.94.