Progress toward production of poultry products having stable and optimum flavor depends considerably on obtaining increased insight into the f undamental causes of flavor and off flayor of the products. To provide part of this basic information, studies on composition of the volatile fraction of cooked chicken have been in progress at this laboratory. A previous paper (12) reported properties of volatile sulfur and nitrogen compounds. The investigation reported here describes a method of isolating volatile carbonyl compounds from cooked chicken, as the 2,4-dinitrophenylhydrazone derivatives, and the subsequent chromatographic separation of the mixed hydrazones into fractions which enabled identification of many of the carbonyl compounds.
EXPERIMENTALIsolation of volatile carbonyl compounds from cooked chicken. Throughout the study, 2,4-dinitrophenylhydrazine ( 2 g. per liter of 2 N HCl) reagent was used to detect presence of carbonyl compounds. Three sets of conditions for isolating volatile carbonyl compounds were considered. These comprised recovery of volatiles evolved from chicken cooked under (a) normal cooking conditions, (b) oxidation-inhibiting conditions, and (c) oxidation-favoring conditions. For condition (a), chickens were simmered in water for 3 hours and distillate of the resulting broth was tested for presence of carbonyl compounds. For conditions (b) and (c), a stream of nitrogen and air, respectively, were passed through each system during cooking, not only to provide an inert and oxidative atmosphere, but also to entrain and sweep volatile components into the detecting reagent. A volatile carbonyl fraction was detected for all 3 isolation methods. Preliminary chromatography of the hydrazones on silicic acid-Celite columns, as described below, showed that regardless of the isolation method a complex mixture of hydrazones was obtained. Of the 3 methods, the one favoring oxidation gave the greatest yield of carbonyl while the one utilizing a nitrogen atmosphere gave the lowest yield. In some respects method (a), based on a normal cooking procedure and utilizing chicken broth distillate, which has been shown (12) to contain detectable flavor, would be a logical starting point for this investigation. However it was decided to study the qualitative nature of the carbonyl fraction isolated by both methods (a) and (c). Such a comparison would show whether a qualitative as well as a quantitative difference exists in the carbonyl fraction isolated and might provide some clues about the nature of carbonyl compounds which are associated with both normal and oxidative off flavors. I t was decided to work first on carbonyl compounds isolated by the atypical oxidation-favoring cooking conditions described below, because this method provided a yield of carbonyl compounds sufficient to make identification of minor constituents seem feasible.The apparatus used for isolation of carbonyl compounds as 2,4-dinitrophenylhydrazones is shown in Figure 1.