Oxidation of quercetin by air oxygen takes place in water and aqueous ethanol solutions under mild conditions, namely in moderately-basic media (pH ∼ 8-10) at ambient temperature and in the absence of any radical initiators, without enzymatic catalysis or irradiation of the reaction media by light. The principal reaction products are typical of other oxidative degradation processes of quercetin, namely 3,4-dihydroxy-benzoic (protocatechuic) and 2,4,6-trihydroxybenzoic (phloroglucinic) acids, as well as the decarboxylation product of the latter -1,3,5-trihydroxybenzene (phloroglucinol). In accordance with the literature data, this process involves the cleavage of the γ-pyrone fragment (ring C) of the quercetin molecule by oxygen, with primary formation of 4,6-dihydroxy-2-(3,4-dihydroxybenzoyloxy)benzoic acid (depside). However under such mild conditions the accepted mechanism of this reaction (oxidative decarbonylation with formation of carbon monoxide, CO) should be reconsidered as preferably an oxidative decarboxylation with formation of carbon dioxide, CO 2 . Direct head-space analysis of the gaseous components formed during quercetin oxidation in aqueous solution at ambient temperature indicates that the ratio of carbon dioxide/carbon monoxide in the gas phase after acidification of the reaction media is ca. 96:4 %. Oxidation under these mild conditi- Molecules, 2007, 12 655ons is typical for other flavonols having OH groups at C 3 (e.g., kaempferol), but it is completely suppressed if this hydroxyl group is substituted by a glycoside fragment (as in rutin), or a methyl substituent. An alternative oxidation mechanism involving the direct cleavage of the C 2 -C 3 bond in the diketo-tautomer of quercetin is proposed.

A comparison of the results of gas-chromatographic analysis performed for the same samples with the use of various instruments with flame-ionization detectors and different (packed and capillary) columns indicates statistically significant differences in the results of data processing by the internal normalization method. Such variations in relative peak areas were detected on the same instrument on a comparison between data obtained under isothermal conditions and with temperature programming. Conceivable reasons for the observed changes in relative peak areas and procedures for the evaluation of errors due to this effect in the analysis of various samples on particular instruments are considered.

No abstract

No abstract

A complete chromatographic analysis usually implies the identification and quantitative determination of the concentrations of all or only particular target components in test samples. To solve the latter problem, various quantitative methods of analysis are used; these differ in sample preparation and data processing techniques [1,2]. The most accurate versions of determination imply the availability of certified reference materials, which are required for the preparation of calibration mixtures (absolute calibration and external standard techniques); the direct addition to test samples (standard addition); or the predetermination of the calibration coefficients ( f i ) expt of the target components with reference to standard substances chosen (internal standard and internal normalization). In the absence of reference samples, quantitative analysis is usually restricted to the use of only external and internal standards or internal normalization on the condition that the chromatographic relative sensitivity coefficients ( f i ) calcd for various compounds were theoretically precalculated. In the presentation of analytical results obtained for chemically similar substances by the internal normalization method with the use of flame-ionization detectors (FIDs), it is often assumed that f i ≡ const. The same approach (the calculation of relative peak areas) is used in processing analytical results obtained by chromatography-mass spectrometry in multicomponent mixtures, although the ionization cross sections of various organic compounds are dramatically different [3].Two objective causes are responsible for such a reduction in the requirements imposed on the results of the quantitative determination of the composition of complex samples. First is the problem of finding analyte reference samples, long considered as insignificant and completely irrelevant to quantitative analysis. However, the attitude to this problem has changed in recent years because the inappropriately great consumption of time at this stage or the high cost of these samples can make unreasonable the performance of the analysis in general [4][5][6][7][8]. Second, various procedures for the theoretical evaluation of relative detector sensitivity coefficients to various compounds based on the composition (or, more rarely, the structure) of analytes [1, vol. 2, p. 173; 7] give only estimated values of ( f i ) calcd , usually without estimated errors; in particular, they do not reflect the known effects of analytical conditions on the values of ( f i ) expt .The effects of conditions (detector geometry, flame temperature, carrier gas, etc.) were characterized in the greatest detail for the absolute characteristics of FID sensitivity [1]. However, variations in the relative values ( f i ) expt were usually considered to be comparatively small, and they were ascribed to the random component of determination errors. For this reason, specific features of data processing by the internal normalization method were not related to the temperature conditions of...