Flavor reproducibility

Figure 2. Preparation and design of experiment to compare the effect of heme and non-heme iron on the development of warmed-over flavor. (Reproduced with permission from Ref. 16.)...

The development of precise and reproducible methods of sensory analysis is prerequisite to the determination of what causes flavor, or the study of flavor chemistry. Knowing what chemical compounds are responsible for flavor allows the development of analytical techniques using chemistry rather than human subjects to characterize flavor (38,39). Routine analysis in most food production for the quaUty control of flavor is rare (40). Once standards for each flavor quaUty have been synthesized or isolated, they can also be used to train people to do more rigorous descriptive analyses. 

Fig. 4. Shelf-life of orange oils (as measured by limonene epoxide formation) as influenced by flavor carrier. (Reproduced with permission from ref. 46. Copyright 1986 Institute of Food Technology.)...

While the exact recovery depends on the volatility of the flavor compounds, most compounds can be detected with this method when present at ppb (mg/liter) concentrations. Reproducibility (CV) is between 5% and 10%. To quantify the amount trapped, an internal standard curve can be made by adding the standards in solvent directly to the trap just before thermal desorption on the side of gas entry during thermal desorption. For liquid homogeneous samples, quantification of the amount in the matrix can be done by a standard addition methodology. 

Use of Model Systems. From time to time it may be desirable to utilize a model or artificial citrus juice system as an aid in studying certain flavor attributes. A big advantage, of course, is the resultant standard "juice" being completely reproducible at any time. However, the big disadvantage is that no matter how well a citrus product is simulated in a model juice system, many taste panelists apparently cannot feel really comfortable when evaluating a purely artificial product. 

These had to be pretreated prior to their use in browning reactions. The latest development in this field is the utilization of fully natural starting materials which permit to reproduce traditionally known food flavors of greatest perfection. 

The availability of and improvement in membranes has rekindled some interest in dialysis in aroma research. Benkler and Reineccius (19, 20) initially published studies on the use of Nafion (Dupont) membranes for the separation of fat from flavor isolates. This would permit solvent extraction to be used in the isolation of aroma compounds from fat containing foods. Chang and Reineccius (21) later used a continuous tubular counter current flow system to accomplish this fat/aroma separation more efficiently. These membranes can be obtained commercially and have been improved in terms of membrane thickness and purity. While the aroma isolate obtained using this membrane may not perfectly reproduce the aroma being studied, this is an alternate technique for aroma isolation. 

Samples were reheated at full power for 1 minute in a microwave oven prior to proceeding with flavor volatile analysis (60 C internal temperature). A reproducibility study was carried out on 5 identical, 100 g samples that had been stored for 3 days after cooking, except that they were not reheated in the microwave prior to analysis. An ad hoc panel convened for these experiments consisted of two trained meat flavor panelists who scored the samples for characterization of MFD according to descriptive sensory methods described by Johnsen and Civille (12) and Love (13). The panelists were also active members of a twelve member descriptive sensory panel at the Center. Two duplicate repetitions were carried out for each experiment (4 samples studied). 

Standards of typical meat flavor sulfur aliphatics and heterocyclics were made from 5 ng/pl to 500 ng/pl in hexane to determine response factors as well as reproducibility in the flame photometric detector. Background sulfur compounds were checked in concentrated reagent blanks. 

From the literature, 19 commonly reported beef flavor aroma compounds containing sulfur were chosen as representative of the capillaiy gas chromatographic spectrum of substances active in the FPD and are noted in Table I. The reproducibility of the instrument was demonstrated by repeated injections to identify the range of sulfur volatiles in beef samples. The internal standard chosen, benzothiophene is also listed in Table I. 

An obvious deficiency in the present marker study is that practically none of the other sulfur-containing molecules, known to be positive contributors to cooked-beef flavor, are detectable by this extraction and FPD gas chromatographic method (2-11. 20). The fate of more of these other contributors to the blend of cooked-beef flavor should also be studied to gain a better perspective. However, this extraction method is simple to use and is recommended for its reproducibility and high recovery of useful markers. 

Take et al. (38) who surveyed the taste-active components of the snow crab, Chionoecetes opilio, in the same way as described for the shrimp and prawn in the preceding section, found that the amino acids were the most important flavor components, and that a synthetic extract (Table VI) prepared by simulating the crab extract reproduced its taste fairly well. 

It was also sound in a supplementary test that glycine betaine serves to produce a delicate flavor. Proline, taurine, and TMAO, although their concentrations are remarkably high, contribute little to the taste, as do the other minor components. It was thereby confirmed that a synthetic extract containing the above twelve components could reproduce the crab-like taste, although it is weaker than that of the mixture containing all the constituents listed in Table VII. 

Despite the above difficulties, several specific chemicals have been associated with specific off-flavors in dairy products. Forss et al. (1955a,b) reported that -hexanal, 2-octenal, 2-nonenal, 2,4-heptadienal and 2,4-nonadienal are the principal carbonyls contributing to the copper-induced cardboard off-flavor in milk. Hall and Lingnert (1986) associated this flavor defect with n-hexanal in spray-dried whole milk. l-Octen-3-one has been associated with a metallic off-flavor in dairy products (Stark and Forss, 1962), the metallic off-flavor being reproduced by addition of l-octen-3-one to milk or cream (Bassette et al., 1986). l-Octen-3-one has a threshold concentration of 1 pg/kg in butterfat (Shipe et al., 1978). 

It was reported that q-glut amyl peptides, particularly peptides with hydrophilic amino acids, such as Glu—Asp, Glu-Thr, Glu-Ser, and Glu-Glu, elicited an umami flavor236 These peptides were isolated from the umami constituents in the enzymatically hydrolyzed products of soybean proteins. The same author also reported that tripeptides such as Glu-Gly-Ser also elicited the umami taste. The threshold value (0.15%) of these peptides is greater than that of MSG. The flavor of meat extract can be reproduced using these peptides with MSG and IMP. Recently,. Y-lacroyl glutamic acid, which is a condensation product of lactic acid with glutamic acid, was shown to elicit a weak umami taste, similar to MSG.237... 

Figure 1. Extract from flavored potato chips. Sugars separated as anions by 1on exchange chromatography with a sodium hydroxide eluent and a pulsed amperometrlc detector. Lactose concentration in the extract Is 70 ppm. "Reproduced with permission from Ref. 13 Copyright 1983, Marcel Dekker. "...

A method for the automated analysis of volatile flavor compounds in foods is described. Volatile compounds are removed from the sample and concentrated via the dynamic headspace technique, with subsequent separation and detection by capillary column gas chromatography. With this method, detection limits of low ppb levels are obtainable with good reproducibility. This method has experienced rapid growth in recent years, and is now in routine use in a number of laboratories. 

Sensory evaluation of lipid oxidation has been conducted by many researchers (98-100). However, as a subjective method, the reproducibility of sensory analysis is generally considered worse than that of chemical or instrumental methods. More recently, use of an electronic nose to monitor the formation of volatile compounds associated with off-flavors from hpid oxidation has been proposed to supplement information from human sensory panels (101). 

The process for cholesterol removal from anhydrous milkfat was patented by General Mills (41). Fractionment Tirtiaux also disclosed the development of a vacuum steam distillation system called the LAN cylinder (38). The steam distillation process (Figure 2) was commercialized, producing a 90-95% cholesterol reduction in anhydrous milkfat with a 95% yield that was reconstituted into 2% fat fluid milk (42). The major disadvantage to the process is that it strips or removes most all volatile flavor components from the fat. These flavor components must be captured (i.e., vacreation) before the distillation process to attempt to reproduce the delicate flavors so desired for reconstitution into a butter product. 

Regardless of the official specifications for soybean oil and its products, the ultimate proof of the pudding is in the eating that is, sensory evaluation of the odors and flavors of soybean oil and its products is the ultimate method to assess oil quality and stability. Sensory evaluation cannot be replaced fully by any chemical or instrumental analysis, although some methods can correlate fairly well with sensory results. Sensory evaluation of oils usually is done by a panel of experts or a trained panel, and often the method recommended by the American Oil Chemists Society (300) is used. During the evaluation, the panel is asked to score the overall flavor quality, as well as the intensity of many individual off-fiavors. Although chemical and physical tests are more reproducible and less time consuming than sensory evaluations, oxidative rancidity and off-flavor evaluation of soybean oils are best done by sensory tests. Correlations established between sensory evaluation scores and... 

Of the 102 process flavor samples analyzed, 90 samples had undetectable levels of PHAs. Two samples contained one PHA each at levels barely above the 50 ppb limit of detection established for this method one sample contained IQ at 67 ppb and the other contained MelQx at 59 ppb. Ten other samples were found to contain one or two PHAs per sample in trace amounts below the 50 ppb limit of detection. Overall, it was reported that this assay demonstrated good sensitivity and reproducibility for detection of PHAs in process flavors and a limit of detection was established to be 50 ppb. 

---