Chem. Senses 26: 533-545,
2001
© Oxford University Press 2001
Reviews |
Evaluation of the Key Odorants of Foods by Dilution Experiments, Aroma Models and Omission
Deutsche Forschungsanstalt für Lebensmittelchemie, Lichtenbergstraße 4, D-85748 Garching, Germany
Correspondence to be sent to: W. Grosch, Deutsche Forschungsanstalt für Lebensmittelchemie, Lichtenbergstraße 4, D-85748 Garching, Germany. E-mail: werner.grosch{at}lrz.tum.de
 |
Abstract |
|---|
 Top Abstract IntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
The state of the art in aroma analysis is reviewed with emphasis on aroma-recombination studies using synthetic blends of odorants (aroma models) which have been prepared on the basis of analytical data. The model that matches the original aroma is the starting material for omission experiments which are performed to establish the odorants that actually contribute to the aroma. These experiments are discussed in detail for the aromas of two wine varieties, three olive oils of different provenance, French fries, boiled beef and coffee. The results indicate that odorants with higher odour activity values (OAV, the ratio of the concentration to the odour threshold) are frequently essential for the aroma. However, there are exceptions where odorants with high OAVs are suppressed in the aroma and compounds with lower OAVs are important contributors. These findings are discussed in the light of model experiments which have been reported in the literature to obtain an insight into the perceptual interactions of odorants in mixtures.
|
Introduction |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
During the first period of aroma research, analytical investigations were performed under the assumption that all the volatiles occurring in food contribute to its aroma. Consequently, the analytical procedure, mostly a combination of gas chromatography and mass spectrometry, was confined to the identification of the volatiles appearing as peaks in the gas chromatogram. Rijkens and Boelens (Rijkens and Boelens, 1975
) calculated that 2600 substances had been identified up to the year 1974 and predicted the occurrence of a total number of 10 000 odorants in foods. Indeed, 
8000 volatiles were reported up to 1997 (Nijssen et al., 1997). However, Rijkens and Boelens (Rijkens and Boelens, 1975) were also beginning to doubt whether all of these volatiles actually contributed to the aromas. These doubts increased further as researchers began to calculate odour activity values (OAVs, synonymous with ‘odour unit’ and ‘odour value’) of volatile compounds in terms of the ratio of the concentration of the odorant in the material to the odour threshold (Rothe and Thomas, 1963). For example, Buttery (Buttery, 1993) found that of the 400 volatiles identified in tomatoes, only 16 reached an OAV of one or more.
The knowledge that not all of the volatiles occurring in a food contribute to its aroma was the reason for changing the methodology of analysis. Since 1984, when the procedure for charm analysis was published (Acree et al., 1984), techniques were developed that focused on the identification of compounds with higher OAVs contributing to the aroma. The new methods were successful as the results allowed the preparation of synthetic blends whose aromas were very close to those of the original foods. Furthermore, the results led to the conclusion that <5 % of the volatiles identified in foods contributed to aromas (Grosch, 2000).
The aim of the following review is to present examples in which the aroma was simulated on the basis of instrumental and sensory analyses. In particular, omission experiments showing character impact odorants of foods will be discussed in some detail. However, the state of the art in the methodology of aroma analysis (Table 1) first needs to be considered.
|
|
Aroma analysis |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
Screening for potent odorants
The analytical procedure starts with dilution experiments (Table 1). As reviewed by Acree (Acree, 1993, 1997) and Grosch (Grosch, 1993, 1994), two techniques—charm analysis and aroma extract dilution analysis (AEDA)—are used in most cases to screen the potent, medium and lower volatile odorants on which the identification experiments are then focused. In both procedures, an extract obtained from the food is diluted, usually as a series of 1:1 or 1:2 dilutions, and each dilution is analysed by gas chromatography–olfactometry (GCO).
In the case of AEDA, the result is expressed as flavour dilution (FD) factor (Grosch, 1993), which is the ratio of the concentration of the odorant in the initial extract to its concentration in the most dilute extract in which the odour is still detectable by GCO. Consequently, the FD factor is a relative measure and is proportional to the OAV of the compound in air. As an example, Figure 1 shows a plot of the FD factors of the odorants of boiled beef versus their retention indices (RI) on a gas chromatography (GC) capillary; this plot is termed an FD chromatogram. The highest FD factors were found for 2-methyl-3-furanthiol (No. 2), 2-furfurylthiol (No. 3), 1-octen-3-one (No. 8), 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HD3F, No. 12) and (E)-2-nonenal (No. 15).
|
Charm analysis constructs chromatographic peaks, the areas of which are proportional to the amount of the chemical in the extract (Acree, 1993
). The primary difference between the two methods is that charm analysis measures the dilution value over the entire time the compounds elute, whereas AEDA simply determines the maximum dilution value detected (Acree, 1993).
Charm values, FD factors and OAVs are functions of the odorant’s concentration in the extract, but are not psychophysical measures for perceived odour intensity (Frijters, 1978). To take this criticism into account, Acree’s group defined the odour spectrum value (OSV) which is the normalized charm value modified with an approximate Steven’s law exponent (Ong and Acree, 1998; Ong et al., 1998).
GCO techniques which are rarely applied are the Osme method (McDaniel et al., 1990; Miranda-Lopez et al., 1992) and olfactometry global analysis or OGA (van Ruth and Roozen, 1994; van Ruth et al., 1995). In the Osme method (Osme is derived from the Greek word meaning smell) the non-diluted aroma extract is examined by GCO (Miranda-Lopez et al. 1992). The odour intensities perceived in replicates by several assessors are averaged, yielding a consensus aromagram. Osme takes into consideration Stevens’ law of psychophysics, as it measures the response to odorants on a scale of time-intensity. However, in a study on the aroma of apples (Plotto et al., 1998) the ranking of aroma-active esters according to Osme resulted in an order comparable with ranking based on OAV. This indicates that the results were not significantly affected when Stevens’ law was not taken into account, as in the case of OAV calculations. OGA is similar to Osme, but each peak in the aromagram is not related to the odour intensity of the volatile, but to its detection frequency. Recently, a first critical comparison of AEDA, Osme and OGA in the case of a food product has indicated that the three methods agree in the evaluation of the most potent odorants of cooked mussels (LeGuen et al., 2000).
A further method—aroma extract concentration analysis (AECA)—is the reverse of AEDA. It starts with the undiluted extract which then is concentrated stepwise and after each step an aliquot is analysed by GCO (Kerscher and Grosch, 1997). This procedure avoids an underestimation of the FD factors of unstable odorants which are degraded in AEDA.
In most cases, the screening methods discussed above are applied to aroma extracts. The highly volatile odorants are not evaluated, as they are lost during the concentration of the aroma extract or are masked in the gas chromatogram by the solvent peak. To overcome this limitation, the screening has to be completed by GCO of decreasing headspace volumes—step II in Table 1 (Holscher and Steinhart, 1992; Guth and Grosch, 1993). In the example of boiled beef (Table 2), the analysis was started with a headspace volume of 20 ml, in which GCO of the headspace (GCOH) revealed 16 odorants. Then the headspace volume was reduced in a series of steps to find the most potent, highly volatile odorants. GCOH of volumes of 10 and 5 ml indicated only 14 and 11 odorants, respectively (Table 2). GCOH of a volume of 2.5 ml still showed five odorants (Nos 1–5 in Table 2); after reduction to 0.5 ml, only methanethiol (No. 1) was found. According to this method, compound No. 1 was the most potent, highly volatile odorant of boiled beef.
|
Identification
The results obtained in the screening procedures are not corrected for losses of odorants during the extraction and concentration procedures. For this reason, the identification experiments should be focused not only on compounds with the highest FD factors, but also on those perceived at lower dilutions (Grosch, 1993). Furthermore, it is necessary to enrich odorants which do not appear as peaks in the gas chromatogram, or are concealed by the large peaks of odourless volatiles. Enrichment (step III in Table 1) is performed by column chromatography (Schieberle and Grosch, 1987; Blank et al., 1992a) or with the new technique of multidimensional gas chromatography or MDGC (Weber et al., 1995). In MDGC the aroma extract is separated on a polar pre-column, then a section of the effluent containing the analyte is cryofocused with liquid nitrogen and subsequently transferred to an unpolar main column which is combined with a mass spectrometer (MS) and a sniffing port.
In the identification experiment it is necessary to compare by GCO the odour quality of the analyte with that of an authentic sample at approximately equal levels. Only when, in addition to GC and MS data, there is an agreement in the sensorial properties, is the analyte, which has been perceived by GCO in the volatile fraction, correctly identified. All of the odorants reported here were identified in this way.
Quantification and calculation of OAVs
Due to the complexity of the volatile fraction and the large differences in concentration, volatility and reactivity of the odorants, it is not possible in most cases to quantify the odorants precisely (error < 10 %) by using conventional methods (Schieberle and Grosch, 1987; Grosch, 1993; Schieberle, 1995; Blank et al., 1999). Not only losses during the extraction and the cleaning-up of the analyte, but even those caused by adsorption during GC (Blank et al., 1992a), which are often overlooked, may lead to incorrect results. However, precise quantitative measurements of the odorants can be performed by the use of stable isotopomers of the analytes as internal standards (examples in Figure 2) in the so-called ‘stable isotope dilution assays’ (SIDA). Each assay consists of the followings steps (Schieberle, 1995; Maarse and Grosch, 1996): after the food sample or its extract has been spiked by the addition of known amounts of the corresponding labelled odorants, the volatile fraction is distilled off; the volatiles are then enriched by column chromatography or MDGC; finally, the subfractions containing the mixture of the unlabelled analytes and their isotopomers are analysed by capillary GC in combination with MS. To differentiate between the labelled and the unlabelled odorants, the traces of distinct ions are recorded. As an example, the mass chromatograms obtained for quantification of the levels of 2-methyl-3-furanthiol, 3-mercapto-2-pentanone, 2-mercapto-3-pentanone and 2-furfurylthiol in cooked beef are shown in Figure 3.
|
) or carbon-13 (
) used as internal standards in stable isotope dilution assays of the corresponding unlabelled analytes. (1) 2-[
-2H2]furfurylthiol; (2) 2-[2H3]methyl-3-furanthiol; (3) 3-mercapto-2-[4,5-2H2]pentanone; (4) [4-2H3]methional; (5) 2-[2H3]ethyl-3,5-dimethylpyrazine; (6) (Z)-1,5-[5,6-2H2]octadien-3-one; (7) trans-4,5-epoxy-(E)-2-[6,7-2H4]decenal; (8) 1-(2,6,6-[6,6-2H6]trimethyl-1,3-cyclohexadienyl)-2-buten-1-one (d-ß-damascenone); (9) 3a,4,5,7a-tetrahydro-3,6-[3-2H3]dimethyl-2(3H)-benzofuranone, (d-wine lactone); (10) tetrahydro-4-methyl-2-(2-methylpropenyl)-2H-[3,4-2H3]pyran (d-rose oxide); (11) 4-hydroxy-2,5-[13C2]dimethyl-3(2H)-furanone; (12) 3-hydroxy-4,5-[4-13C]dimethyl-2(5H)-[5-13C]furanone.
|
The precision of SIDA has been confirmed in model experiments (Schieberle and Grosch, 1987
; Guth and Grosch, 1990). Although after cleaning-up, the yields of some analytes were <10 %, the results of quantification were correct as the standards showed equal losses. Unless otherwise stated, the quantitative data reported here were obtained by SIDA.
To approach the situation in food, OAVs of the odorants are calculated (step IV in Table 1). For this purpose, odour threshold values of these volatiles are estimated in a medium that predominates in the food, e.g. water, oil, starch. As an example, the OAVs of the potent odorants of tomato paste, calculated as ratio of concentration to odour threshold value in water are listed in Table 3. The highest OAVs were found for ß-damascenone, followed by dimethyl sulphide and methional. In agreement with the suggestion of Frijters (Frijters, 1978) that the odour activity concept can give some guidance in flavour research, it is assumed that the odorants showing higher OAVs contribute strongly to the aroma of tomato paste (Guth and Grosch, 1999).
|
Aroma models
Flavour chemists have always wanted to prove that the results of aroma analyses are correct. Therefore, synthetic blends of odorants (aroma models) were prepared on the basis of the obtained analytical data and their aromas were compared with those of the originals (step V in Table 1). Early experiments, e.g. to reconstitute the aroma of olive oil (Flath et al., 1973) and that of orange juice (Dürr and Schobinger, 1981), were not very successful. Lack of odorants which are only detectable in the food by GCO and not by the GC detector as well as incorrect quantitative data are the main reasons that the original and the corresponding aroma model disagree.
Progress was made by Buttery (Buttery, 1993) and Buttery et al. (Buttery et al., 1990, 1994, 1995) in a study on the aromas of fresh ripe tomatoes and tomato paste. The authors identified >400 volatiles and calculated OAVs for 40 odorants which had been quantified with conventional analytical methods. On the basis of these results and supported by GCO, they prepared aroma models for fresh tomatoes and tomato paste. Sensory panel studies showed these mixtures to have aromas very similar to the originals.
Recently, the study of the aroma of tomato paste was repeated using AEDA, GCOH and SIDA (Guth and Grosch, 1999). The composition of the resulting aroma model II is compared in Table 3 with model I obtained by Buttery (Buttery, 1993).
The concentrations of the first six odorants were in the same range in the two models. 1-Nitro-2-phenylethane was lacking in model II because it was not perceived in AEDA. Due to relatively low OAVs, linalool and 2-phenylethanol were not added to model I. HD3F was not taken into consideration, since addition of 1 mg/l of this compound to model I did not confer any improved tomato paste aroma (Buttery et al., 1994, 1995). Also, the aroma of model II was very close to that of the original (Guth and Grosch, 1999). The authors had the impression that of the compounds lacking in model I, 3-hydroxy-4,5-dimethyl-2(5H)-furanone (HD2F) belonged to the key odorants of tomato paste.
The preparation of aroma models is simple for liquid foods, as it is easy to obtain a homogenous blend of odorants. Difficulties arise, however, in the case of solid foods, because in general it is not possible to reproduce the composition and distribution of the non-volatile components. As a compromise simple materials, e.g. oil/water mixtures, cellulose and starch, are used as the base (Czerny et al., 1999). However, the aroma model for a solid food can be improved when it is not affected by interactions of the odorants with the material used as the base. This would be the case when the model is identical with the composition of the odorants in the air above the food. An approach to gain data for the preparation of such models was developed and successfully applied to reconstitute the aroma of bread crust (Zehentbauer and Grosch, 1997) and roasted coffee (Grosch and Mayer, 2000).
|
Omission experiments |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
Aroma models based on the results of dilution experiments have been prepared for the foods listed in Table 4. In most cases the aroma of the model matched very well that of the original. However, in the dilution experiments, the odour impact of the volatiles was evaluated after separation by GC (example in Figure 1). Perceptual interactions of odorants, which were characterized by inhibition and suppression (Laing, 1988a
, b; Acree, 1993), are abolished. Therefore, the question of which compound among the potent odorants actually contributes to the aroma has to be answered by omission experiments (step VI in Table 1).
|
To answer this question, the impact of components of the aroma model was evaluated by omission experiments for those foods which are listed in Table 4. Some details of the results obtained in studies of the aromas of wine, olive oil, French fries and boiled beef will be discussed in the following.
Wine
In a study on the aroma of the two wine varieties, Gewürztraminer and Scheurebe (Guth 1997a, 1998), OAVs were calculated on the basis of odour threshold values in water/ ethanol (9 + 1, w/w). According to Table 5, esters, which are responsible for the fruity note in the aroma, showed high OAVs in the two wines. Characteristic differences in the OAVs were found for cis-rose oxide and 4-mercapto-4-methylpentan-2-one. The former was highly aroma-active in Gewürztraminer and the latter in Scheurebe, although its concentration amounted only to 0.4 µg/l (Table 5).
|
An aroma model for Gewürztraminer containing nine odorants with OAV
_10 (Table 5) was not satisfactory. However, when the model was completed by the odorants showing OAVs of 1–9 (Table 6), the aroma matched very well that of the original wine. In contrast, addition of a further 13 odorants with OAVs <1 did not affect the aroma of the model. Also, the aroma of the Scheurebe wine was imitated by a synthetic blend of the odorants with OAVs 1. When cis-rose oxide and 4-mercapto-4-methylpentan-2-one were absent in the models for Gewürztraminer and Scheurebe, respectively, the aromas changed drastically and did not resemble those found in the original samples. In contrast, omission of e.g. acetaldehyde, ß-damascenone and geraniol, had only a small effect. It might be of interest that neither of the odorants listed in Tables 5 5 and 6 occur only in wine. cis-Rose oxide, for example, contributes to the aroma of linden honey (Blank et al., 1989) and 4-mercapto-4-methyl-2-pentanone to that of basil (Guth and Murgoci, 1997) and grapefruit (Schieberle and Büttner, 2000).
|
To investigate the effect of ethanol on the aroma and the taste of Gewürztraminer, an aqueous ethanolic solution was investigated. The solution contained, in addition to the odorants listed in Tables 5 and 6, the taste compounds screened and quantified in previous experiments (Guth, 1998
). Then the ethanol concentration was lowered in the flavour model (Table 7). Reduction from nearly 10 % to 9 % ethanol did not change the flavour. However, further reduction to a concentration of 7 % affected the flavour as the intensities of the fruity, flowery odour and the acid taste increased. This change increased with the decrease of the ethanol concentration. In the model containing only 3 % ethanol, the characteristic flavour of the wine was completely destroyed.
|
The drastic change in the flavour of the model might be caused by an increase in the partial pressure of the odorants due to the reduction of the ethanol concentration. A model experiment (Guth, 1997b
) indicates that doubling of the odorant concentrations in the gas phase was observed at the most when ethanol was lacking. Then, the experiments summarized in Table 8 were performed to clarify whether ethanol affects the odour-activity of aroma compounds. The odour threshold values of some components of the wine flavour were determined, both in the absence and in the presence of ethanol in the gas phase. The values listed in Table 8 indicate a strong increase in the odour threshold when ethanol was present. The increase varied between factors of 10 and 312 which were found for ethyl hexanoate and methylpropanol, respectively. The results suggest that the inhibition of the odour activity of the wine volatiles by ethanol is much larger than the reduction of their partial pressure.
|
Olive oil
The different aromas of olive oils originating from Italy, Morocco and Spain were reconstituted by solutions of 25, 14 and 21 odorants, respectively, in a refined plant oil (Reiners and Grosch, 1998). Omission experiments indicated that the green smelling (Z)-3-hexenal and (Z)-3-hexenol had the greatest odour impact in the oil from Italy. This corresponds to their relatively higher OAVs. On the other hand, acetaldehyde, which showed the highest OAV of all potent odorants occurring in the oil from Italy, was only of minor importance.
The aroma of the oil from Morocco differed from that of the other two oil samples in a special fruity note which was caused by ethyl cyclohexyl carboxylate. This ester, showing the highest OAV (550) of all odorants, is formed by fermentation of the olive fruits which takes place in the traditional process used for oil production in Morocco. The omission experiments revealed in addition that guaiacol, despite of its lower OAV of 3, was essential for the aroma of this oil variety.
Apart from acetaldehyde, 4-methoxy-2-methyl-2-butanethiol appeared with the highest OAV in the oil from Spain. This blackcurrant-like smelling thiol, which has also been found in the essential oil of blackcurrant buds (Rigaud et al., 1986), was indeed the key odorant of this oil variety.
French fries
A model with aroma very similar to that of French fries processed in palm oil (Wagner and Grosch, 1998) contained 19 odorants dissolved in sunflower oil. The composition of this solution was varied to clarify whether all of the components took part in the generation of the aroma. An extract of the results is shown in Table 9.
|
Methanethiol, the odorant with the highest OAV, was missed in exp. 1 by each member of the panel because a boiled-potato-like note was lacking in the aroma profile of French fries. In exp. 2 the judges agreed that the decadienal stereoisomers had a great impact on the aroma and were responsible for the deep fried, fatty odour note. Of the four pyrazines, the absence of 2-ethyl-3,5-dimethylpyrazine and 3-ethyl-2,5-dimethylpyrazine was clearly recognized in exp. 3, but the absence of the two pyrazines, 2,3-diethyl-5-methyl- and 2-ethenyl-3-ethyl-5-methylpyrazine, was not noted in exp. 4. This means that 2,3-diethyl-5-methyl-pyrazine did not play a role in the aroma of French fries, although its OAV was much higher than the OAV of the other three pyrazines. Absence of the three Strecker aldehydes and of trans-4,5-epoxy-(E)-2-decenal changed the aroma of the model in exps 5 and 6, respectively, but the deviation in the odour intensity and quality was only small. In accordance with their low OAV, the carbonyl compounds which were lacking in exp. 7 did not significantly contribute to the aroma. The result of exp. 8 was very surprising. Although methional was the odorant showing the second highest OAV of all of the volatiles occurring in French fries (Table 9), the aroma of the model in which this odorant was lacking did not differ from that of the complete model. This result was also remarkable for the reason that the odour quality of methional is reminiscent of boiled potatoes. Obviously, other odorants, e.g. methanethiol and the pyrazines lacking in exp. 3, might cause the boiled-potato-like note in the odour profile of French fries.
Boiled beef
The aroma model of a boiled beef sample with a relatively higher fat content consisted of 16 volatiles (Table 10). Their selection based on the results of AEDA and GCOH is shown in Figure 1 and Table 2. Omission experiments (Table 10) revealed that nine odorants contributed significantly to the aroma of boiled beef. 2-Furfurylthiol, HD3F, methylfuranthiol, but also mercaptopentanone and methional belonged to this group. In a further triangle test, the seven odorants whose contribution to the overall odour was not significantly established in exps 10–16 (Table 10) were omitted. When compared with the model containing the complete set of 16 odorants, only 6 of the 11 assessors identified the reduced model (Grosch et al., 2000). This means that the aroma of the reduced model does not differ significantly from that of the complete model. Therefore, it was concluded that compounds Nos 1–9 in Table 10 are the key odorants of boiled beef. 12-Methyltridecanal did not significantly contribute to the aroma of the model, although its OAV was very high (Table 10). Obviously, it did not successfully compete with octanal, nonanal and decadienal for the stimulation of the fatty note in the aroma profile of boiled beef. However, it belonged to the key odorants when lean beef was cooked because it originates from the membranes of the muscle tissue and not from the fat depot (Guth and Grosch, 1994; Kerscher et al., 2000).
|
Roasted coffee
In a further experiment, a model which imitates the odour profile of a medium-roasted coffee sample from Colombia was prepared using 27 volatiles (Czerny et al., 1999). Then 20 omission experiments were carried out. An extract of the results is shown in Table 11.
|
A comparison of exps 1 and 2 indicated that 2-furfurylthiol was the outstanding odorant in the group of sulphur compounds. Its absence in exp. 1 lowered so strongly the sulphurous/roasty note odour of the model that 15 of 20 answers were correct. The important role established in these omission experiments for 2-furfurylthiol confirmed the assumption of Reichstein and Staudinger (Reichstein and Staudinger, 1955
) that it was a key component of coffee flavour. However, its impact was much lower in the coffee beverage due to its lower yield in the brewing process (Mayer et al., 2000). Absence of the four furanones in exp. 3 (Table 3) was clearly recognized. However, when HD3F and 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-furanone (EHM3F) were lacking in exp. 4, or HD2F and EHM2F in exp. 5, the aromas of the complete and the reduced model were not significantly different. Obviously, two of the four furanones were sufficient for the coffee aroma.
Experiments 6 and 7 indicated that the aroma of the model was significantly changed by the absence of the alkyl pyrazines PI, PII and PIII. Methoxypyrazine PIV which, in contrast to exp. 6, was present in exp. 7 did not affect the aroma of the model (Grosch et al., 2001). This was confirmed in omission experiments for the coffee brew aroma (Mayer et al., 2000). This result is of interest as PIV is the most aroma-active constituent of raw coffee and is stable during roasting (Czerny and Grosch, 2000). The authors assumed that roasting of the raw material not only produces the pleasant aroma which is characteristic for coffee but, in addition, generates odorants which suppress the peasy odour note caused by PIV.
|
Final discussion |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
The sensory studies reviewed here indicate that complex odours, such as those of different types of wine and olive oil, French fries, boiled beef and coffee, can be reconstituted by aroma models that contain only a small fraction of the great number of volatiles occurring in these foods. Omission experiments are then performed to identify the character impact odorants.
The ability of human subjects to detect changes in odour mixtures after omission of one compound is poor when the odour activities of the components are similar (Laska and Hudson, 1992). However, in the experiments reported here, the odour activities (expressed as OAV) of the food volatiles are different, and this fact facilitates the identification of a mixture with a single component omitted. The results obtained for boiled beef, for example, indicate that omission of 2-furfurylthiol or HD3F or 3-mercapto-2-pentanone or methanethiol in the aroma model was recognized by a highly significant number of assessors.
The omission experiments reviewed here have revealed that some of the odorants with higher OAVs cause the characteristic note of the aroma, e.g. cis-rose oxide and 4-mercapto-4-methylpentan-2-one in Gewürztraminer and Scheurebe wines, respectively, or 4-methoxy-2-methyl-2-butanethiol in an olive oil variety and 2-furfurylthiol in roasted coffee. But these experiments have also shown negative perceptual interactions of odorants. For example, methional and 12-methyltridecanal, although appearing with high OAVs, are suppressed in the aroma model for French fries and boiled beef, respectively. A further impressive example, as already mentioned above, is the masking of the peasy smell of methoxypyrazines in coffee by odorants formed during the roasting process.
In simple experiments with food-relevant odorants, antagonistic effects have been reported for mixtures of (Z)-3-hexenal and (E,Z)-2,4-heptadienal (Meijboom, 1964), as well as for (E)-2-decenal and (E)-2-hexenal (Laing and Willcox, 1983).
The perceptual interaction of odorants was also investigated using complex mixtures. A large study was performed by Meilgaard (Meilgaard, 1975, 1982) to evaluate the odorants contributing to the aroma of beer. He found that the odour threshold values of most of the volatiles are higher in beer than in water, e.g. that of butanol, 3-methylbutanol and dimethyl sulphide, by factors of 400, 280 and 152, respectively. (E)-2-nonenal is an exception, since its threshold is approximately as low in beer as in water. In 20 separate experiments the odour threshold values of mixtures containing up to five alcohols, fatty acids, esters or carbonyl compounds, as well as alcohols and esters, were estimated using beer as the solvent. A comparison of the OAV of the mixture with the sum of the OAVs of the components led to the general conclusion that most of the beer aroma compounds exhibit partial addition: 0.5 OAV + 0.5 OAV produces 0.8 ± 0.2 OAV. No obvious examples of synergism or antagonism were observed in these studies.
A strict additive effect was shown by Guadagni et al. (Guadagni et al., 1963, 1966) for mixtures containing ten compounds, each at 1/10 of their threshold in aqueous solution. This was found with 20 different combinations of sub-threshold concentrations and was independent of the chemical nature of the compound. Indications for a synergistic effect of odorants at sub-threshold concentrations were reported by Laska and Hudson (Laska and Hudson, 1991). Their suggestion was derived from the observation that the threshold of an odorant in the mixed state is lower than that in the unmixed state. The greatest difference was reported for decyl acetate. Its threshold decreases by five orders of magnitude when it is a component of a mixture consisting of cineole, linalool, (–)-carvone, tert-butylcyclohexyl acetate and 2-pentanone. However, as discussed here, aroma models containing all compounds with OAVs >1 imitate very well the corresponding food aromas. There are no indications that in these cases additional volatiles with OAV <1 might play an important role. In contrast, as mentioned above, antagonistic effects causing a suppression of odorants showing high OAVs have been ascertained.
Another aspect of the perceptual interactions of odorants was investigated by Laing (Laing, 1994). He found that humans can only discriminate and identify up to three or four odorants in a mixture. When more than four odorants are mixed, as shown here for the aroma models, then they produce on overall aroma that differs from the individual odours. Laing (Laing, 1994) suggested that the new aroma is the result of a large loss of the odour properties of the individual constituents. This agrees with the observation that omission of one or two odorants in the aroma model often causes only small changes in odour quality, which are difficult to describe and hardly affect the odour intensity of the model. Laing (Laing, 1994) proposed two mechanisms— spatial and temporal filtering—to explain the perceptual interactions of odorants in mixtures. The spatial filtering hypothesis starts with the observation that in the olfactory epithelium an odorant interacts with a substantial number of receptor sites and cells which are often widely distributed. Consequently, stimulation by an odorant produces a pattern of responsive cells. However, when a mixture of two odorants enters the nose, information about one or both odorants can be lost through competition for receptor sites or cells. Consequently, the greater the number of constituents in a mixture, the greater is the chance of an overlapping of response pattern (Laing, 1994).
The concept of temporal filtering is based on the observation that different odorants may differ greatly in the reaction time for the recognition of odour quality (Laing and MacLeod, 1992; Laing 1994). This time was termed ‘latency’, and it was shown that it is inversely proportional to the odorant concentration. When, for example, two odorants compete for receptor sites and cells, it is proposed that the ‘faster’ odorant will be the first to occupy receptor sites and thus to reduce the chance of stimulation by the ‘slower’ odorant, or might act as an antagonist (Laing, 1994).
Application of these mechanisms, for example to the results of the omission experiments which have been discussed before for the boiled beef aroma model, might lead to the following proposal: 12-methyltridecanal does not successfully compete with octanal, nonanal and decadienal for the receptor sites and cells, and/or its latency is longer than those of the aldehydes having a shorter carbon chain.
|
Conclusion |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
It has been proven by aroma models that the odorants occurring in complex aromas can be identified by a multi-step analytical procedure based on the odour activity concept. As this procedure has been successfully applied to the majority of important foods, most of the volatiles which might play a role in aromas have been now identified. Omission experiments have revealed that odorants with higher OAVs are frequently essential for the aroma. However, there are also exceptions where odorants with high OAVs are suppressed in the aroma and compounds with lower OAVs are important contributors. These results are in agreement with the observation of Laing (Laing, 1994
) that there is a large loss of sensory properties about the individual constituents when odorants are mixed. To understand the perceptual interactions of odorants which result in well-known aromas, corresponding experiments should be performed with those volatiles that have been identified as odour-active components of foods.
|
References |
|---|
|
TopAbstractIntroductionAroma analysisOmission experimentsFinal discussionConclusionReferences |
|---|
Acree, T.E. (1993) Bioassays for Flavor. In Acree, T.E. and Teranishi, R. (eds), Flavor Science. Sensible Principles and Techniques, ACS Professional Reference Book. American Chemical Society, Washington, DC, pp. 1–20.
Acree, T.E. (1997) GC/olfactometry. Anal. Chem. News Features, 69, 170A–175A.
Acree, T.H., Barnard, J. and Cunningham, D.G. (1984) A procedure for the sensory analysis of gas chromatographic effluents. Food Chem., 14, 273–286.
Blank, I., Fischer, K.-H. and Grosch, W. (1989) Intensive neutral odourants of linden honey. Differences from honeys of botanical origin. Z. Lebensm. Unters. Forsch., 189, 426–433.
Blank, I., Sen, A. and Grosch, W. (1992a) Potent odorants of the roasted powder and brew of Arabica coffee. Z. Lebensm. Unters. Forsch., 195, 239–245.
Blank, I., Sen, A. and Grosch, W. (1992b) Sensory study on the character impact flavour compounds of dill herb (Anethum graveolens L.). Food Chem., 43, 337–343.
Blank, I., Milo, C., Lin, J. and Fay, L.B. (1999) Quantification of aroma-impact components by isotope dilution assay. Recent developments. In Teranishi, R., Wick, E.L. and Hornstein, I. (eds) Flavor Chemistry. Thirty Years of Progress. Kluwer Academic/Plenum, New York, pp. 63–74.
Buttery, R.G. (1993) Quantitative and sensory aspects of flavor of tomato and other vegetables and fruits. In Acree, T.E. and Teranishi, R. (eds), Flavor Science. Sensible Principles and Techniques, ACS Professional Reference Book. American Chemical Society, Washington, DC, pp. 259–286.
Buttery, R.G., Teranishi, R., Ling, L.C. and Turnbaugh, J.G. (1990) Quantitative and sensory studies on aroma paste volatiles. J. Agric. Food Chem., 38, 336–340.
Buttery, R.G., Takeoka, G.R., Krammer, G.E. and Ling, L.C. (1994) Identification of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (furaneol) and 5-methyl-4-hydroxy-3(2H)-furanone in fresh and processed tomato. Lebensm. Wiss. Technol., 27, 592–594.
Buttery, R.G., Takeoka, G.R. and Ling, L.C. (1995) Furaneol: odor threshold and importance to tomato aroma. J. Agric. Food Chem., 43, 1638–1640.
Czerny, M. and Grosch, W. (2000) Potent odorants of raw Arabica coffee. Their changes during roasting. J. Agric. Food Chem., 48, 868–872.[Web of Science][Medline]
Czerny, M., Mayer, F. and Grosch, W. (1999) Sensory study on the character impact odorants of roasted Arabica coffee. J. Agric. Food Chem., 47, 695–699.[Web of Science][Medline]
Dürr, P. and Schobinger, U. (1981) The contribution of some volatiles to the sensory quality of apple and orange juice odour. In Schreier, P. (ed.), Flavour ’81. Walter de Gruyter, Berlin, pp. 179–193.
Flath, R.A., Forrey, R.R. and Guadagni, D.G. (1973) Aroma components of olive oil. J. Agric. Food Chem., 21, 948–952.
Frijters, J.E.R. (1978) A critical analysis of the odour unit number and its use. Chem. Senses, 3, 227–233.
Grosch, W. (1993) Detection of potent odorants in foods by aroma extract dilution analysis. Trends Food Sci. Technol., 4, 68–73.
Grosch, W. (1994) Determination of potent odorants in foods by aroma extract dilution analysis (AEDA) and calculation of odour activity values (OAVs).Flavour Fragrance J., 9, 147–158.
Grosch, W. (2000) Specificity of the human nose in perceiving food odorants. In Schieberle, P. and Engel, K.-H. (eds), Frontiers of Flavour Science, Proceedings of the Ninth Weurman Flavour Research Symposium. Deutsche Forschungsanstalt für Lebensmittelchemie, Garching, pp. 213–219.
Grosch, W. and Mayer, F. (2000) Release of odorants from roasted coffee. In Roberts, D. and Taylor, A. (eds) Flavor Release, ACS Symposium Series No. 763. American Chemical Society, Washington, DC, pp. 430–438.
Grosch, W., Czerny, M., Mayer, F. and Moors, A. (2000) Sensory studies on the key odorants of roasted coffee. In Parliment, T.H., Ho, C.-T. and Schieberle, P. (eds), Caffeinated Beverages. Health Benefits, Physiological Effects, and Chemistry, ACS Symposium Series No. 754. American Chemical Society, Washington, DC, pp. 292–298.
Grosch, W., Kerscher, R., Kubickowa, J. and Jagella, T. (2001) Aroma extract dilution analysis (AEDA) versus aroma extract concentration analysis (AECA). In Leland, J.V., Büttner, A., Schieberle, P. and Acree, T.E. (eds), Gas Chromatography-Olfactometry: The State of the Art, ACS Symposium Series No. 782. American Chemical Society, Washington, DC, in press.
Guadagni, D.G., Buttery, R.G., Okano, S. and Burr, H.K. (1963) Additive effect of sub-threshold concentrations of some organic compounds associated with food aromas. Nature, 200, 1288–1289.
Guadagni, D.G., Buttery, R.G. and Harris, J. (1966) Odour intensities of components. J. Sci. Food Agric., 17, 142–144.[Web of Science][Medline]
Guth, H. (1997a) Quantitation and sensory studies of character impact odorants of different white wine varieties. J. Agric. Food Chem., 45, 3027–3032.
Guth, H. (1997b) Objectification of white wine aromas. Thesis, Technical University, München (in German).
Guth, H. (1998) Comparison of different white wine varieties in odor profiles by instrumental analysis and sensory studies. In Waterhouse, A.L. and Ebeler, S.E. (eds) Chemistry of Wine Flavor, ACS Symposium Series No. 714. American Chemical Society, Washington, DC, pp. 39–53.
Guth, H. and Grosch, W. (1990) Deterioration of soya-bean oil: quantification of primary flavour compounds using a stable isotope dilution assay. Lebensm. Wiss. Technol., 23, 513–522.
Guth, H. and Grosch, W. (1993) Identification of potent odourants in static headspace samples of green and black tea powders on the basis of aroma extract dilution analysis (AEDA). Flavour Fragrance J., 8, 173–178.
Guth, H. and Grosch, W. (1994) Identification of the character impact odorants of stewed beef juice by instrumental analyses and sensory studies. J. Agric. Food Chem., 42, 2862–2866.
Guth, H. and Grosch, W. (1999) Evaluation of important odorants in foods by dilution techniques. In Teranishi, R., Wick, E.L. and Hornstein, I. (eds), Flavor Chemistry. Thirty Years of Progress. Kluwer Academic/Plenum, New York, pp. 377–386.
Guth, H. and Murgoci, A.M. (1997) Identification of the key odorants of basil (Ocimum basilicum L.)—effect of different drying procedures on the overall flavour. In Kruse, H.P. and Rothe, M. (eds), Flavour Perception—Aroma Evaluation. Eigenverlag Universität Potsdam, Potsdam, pp. 233–242.
Holscher, W. and Steinhart, H. (1992) Investigation of roasted coffee freshness with an improved headspace technique. Z. Lebensm. Unters. Forsch., 195, 33–38.
Jagella, T. and Grosch, W. (1999) Flavour and off-flavour compounds of black and white pepper (Piper nigrum L.). Eur. Food Res. Technol., 209, 16–21, 22–26, 27–31.
Kerscher, R. and Grosch, W. (1997) Comparative evaluation of potent odorants of boiled beef by aroma extract dilution and concentration analysis. Z. Lebensm. Unters. Forsch., 204, 3–6.
Kerscher, R. and Grosch, W. (1998) Quantification of 2-methyl-3-furanthiol, 2-furfurylthiol, 3-mercapto-2-pentanone, and 2-mercapto-3-pentanone in heated meat. J. Agric. Food Chem., 46, 1954–1958.
Kerscher, R. and Grosch, W. (2000) Comparison of the aromas of cooked beef, pork and chicken. In Schieberle, P. and Engel, K.-H. (eds), Frontiers of Flavour Science. Proceedings of the Ninth Weurman Flavour Research Symposium. Deutsche Forschungsanstalt für Lebensmittelchemie, Garching, pp. 17–20.
Kerscher, R., Nürnberg, K., Voigt, J., Schieberle, P. and Grosch, W. (2000) Occurrence of 12-methyltridecanal in microorganisms and physiological samples isolated from beef. J. Agric. Food Chem., 48, 2387–2390.[Web of Science][Medline]
Kubickova, J. and Grosch, W. (1998) Quantification of potent odorants in Camembert cheese and calculation of their odour activity values. Int. Dairy J., 8, 17–23.
Laing, D.G. (1988a) Coding of chemosensory stimulus mixtures. Ann. NY Acad. Sci., 510, 61–66.[Web of Science][Medline]
Laing, D.G. (1988b) Relationship between differential adsorption of odorants by the olfactory mucus and perception of odour mixtures. Chem. Senses, 13, 463–471.
Laing, D.G. (1994) Perceptual odour interactions and objective mixture analyses. Food Qual. Pref., 5, 75–80.
Laing, D.G. and MacLeod, P. (1992) Reaction time for the recognition of odour quality. Chem. Senses, 17, 337–346.
Laing D. G. and Willcox, M.E. (1983) Perception of components in binary odour mixtures. Chem. Senses, 7, 249–264.
Laska, M. and Hudson R. (1991) A comparison of the detection thresholds of odour mixtures and their components. Chem. Senses, 16, 651–662.
Laska, M. and Hudson, R. (1992) Ability to discriminate between related odour mixtures. Chem. Senses, 17, 403–415.
Le Guen, S., Prost, C. and Demaimay, M. (2000) Critical comparison of three olfactometric methods for the identification of the most potent odorants in cooked mussels (Mytilus edulis). J. Agric. Food Chem., 48, 1307–1314.[Web of Science][Medline]
Maarse, H. and Grosch, W. (1996) Analysis of taints and off-flavours. In Saxby, M.J. (ed.), Food Taints and Off-flavours. Blackie Academic & Professional, London, pp. 72–106.
Masanetz, C. and Grosch, W. (1998) Key odorants of parsley leaves (Petroselinum crispum [Mill.] Nym. Ssp. crispum) by odour activity values. Flavour Fragrance J., 13, 115–124.
Mayer, F., Czerny, M. and Grosch, W. (2000) Sensory study of the character impact aroma compounds of a coffee beverage. Euro Food Res. Technol., 211, 272–276.
McDaniel, M.R., Miranda-Lopez, R., Watson, B.T., Micheals, N.J. and Libbey, L.M. (1990) Pinot noir aroma: a sensory/gas chromatographic approach. In Charalambous, G. (ed.), Flavours and Off-flavours. Elsevier Science, Amsterdam, pp. 23–25.
Meijboom, P.W. (1964) Relationship between molecular structure and flavor perceptibility of aliphatic aldehydes. J. Am. Oil Chem. Soc., 41, 326–328.
Meilgaard, M.C. (1975) Aroma volatiles in beer: purification, flavour, threshold and interaction. In Drawert, F. (ed.), Geruch und Geschmackstoffe. Verlag Hans Carl, Nürnberg, pp. 211–254.
Meilgaard, M.C. (1982) Prediction of flavor differences between beers from their chemical composition. J. Agric. Food Chem., 30, 1009–1017.
Miranda-Lopez, R., Libbey, L.M., Watson, B.T. and McDaniel, M.R. (1992) Odor analysis of Pinot noir wines from grapes of different maturities by a gas chromatography-olfactometry technique (Osme). J. Food Sci., 57, 985–993, 1019.
Nijssen, L.M., Visscher, C.A., Maarse, H., Willemsens, L.C. and Boelens, M.H. (1997) Volatile compounds in food, 7th edn and supplement 1. TNO Nutrition and Food Research Institute, Zeist, The Netherlands.
Ong, P.K.C. and Acree, T.E. (1998) Gas chromatography/olfactometry analysis of lichee (Litchi chinensis Sonn.). J. Agric. Food Chem., 46, 2282–2286.
Ong, P.K.C., Acree, T.E. and Lavin, E.H. (1998) Characterization of volatiles in rambutan fruit (Nephelium lappaceum L.). J. Agric. Food Chem., 46, 611–615.[Web of Science][Medline]
Plotto, A., Mattheis, J.P., Lundahl, D.S., and McDaniel, M.R. (1998) Validation of gas chromatography olfactometry results for ‘Gala’ apples by evaluation of aroma-active compound mixtures. In Mussinan, C.J. and Morello, M.J. (eds), Flavor Analysis. Developments in Isolation and Characterization, ACS Symposium Series No. 705.. American Chemical Society, Washington, DC, pp. 290–302.
Preininger, M., Warmke, R. and Grosch, W. (1996) Identification of the character impact compounds of Swiss cheese by sensory studies on models. Z. Lebensm. Unters. Forsch., 202, 30–34.
Reichstein, T. and Staudinger, H. (1955) The aroma of coffee. Parfum. Essent. Oil Rec., 46, 86–88.
Reiners, J. and Grosch, W. (1998) Odorants of virgin olive oils with different flavor profiles. J. Agric. Food Chem., 46, 2754–2763.
Rigaud, J., Etievant, P., Henry, R. and Latrasse, A. (1986) 4-Methoxy-2-methyl-2-butanethiol a major constituent of the aroma of the black currant bud (Ribes nigrum L.). Sci. Aliments, 6, 213–220 (in French).
Rijkens, F. and Boelens, M.H. (1975) The future of aroma research. In Maarse, H. and Groenen, P.J. (eds), Aroma Research. Proceedings of the International Symposium on Aroma Research. Centre for Agricultural Publishing and Documentation, Wageningen, pp. 203–220.
Rothe, M. and Thomas, B. (1963) Aroma of bread. Evaluation of chemical taste analyses with the aid of threshold value. Z. Lebensm. Unters. Forsch., 119, 302–310 (in German).
Schieberle, P. (1995) New developments on methods for analysis of volatile flavor compounds and their precursors. In Gaonkar, A.G. (ed.), Characterization of Food-emerging Methods. Elsevier, Amsterdam, pp. 403–433.
Schieberle, P. and Büttner, A. (2000) Comparison of the key odorants in fresh, handsqueezed juices of orange (Valencia late) and grapefruit (White marsh). In Schieberle, P. and Engel, K.-H. (eds), Frontiers of Flavour Science. Proceedings of the Ninth Weurman Flavour Research Symposium. Deutsche Forschungsanstalt für Lebensmittelchemie, Garching, pp. 10–16.
Schieberle, P. and Grosch, W. (1987) Quantitative analysis of aroma compounds in wheat and rye bread crusts using a stable isotope dilution assay. J. Agric. Food Chem., 35, 252–257.
Schieberle, P. and Hofmann, T. (1997) Evaluation of the character impact odorants in fresh strawberry juice by quantitative measurements and sensory studies on model mixtures. J. Agric. Food Chem., 45, 227–232.
Schieberle, P., Gassenmeier, K., Guth, H., Sen, A. and Grosch, W. (1993) Character impact odour compounds of different kinds of butter. Lebensm. Wiss. Technol., 26, 347–356.
Semmelroch, P. and Grosch, W. (1996) Studies on character impact odorants of coffee brews. J. Agric. Food Chem., 44, 537–543.
Van Ruth, S.M. and Roozen, J.P. (1994) Gas chromatography/sniffing port analysis and sensory evaluation of commercially dried bell peppers (Capsicum annuum) after rehydration. Food Chem., 51, 165–170.
Van Ruth, S.M., Roozen, J.P. and Posthumus, M.A. (1995) Instrumental and sensory evaluation of the flavour of dried French beans (Phaseolus vulgaris) influenced by storage conditions. J. Sci. Food Agric., 69, 393–401.
Wagner, R. and Grosch, W. (1998) Key odorants of French fries. J. Am. Oil Chem. Soc., 75, 1385–1392.
Weber, B., Maas, B. and Mosandl, A. (1995) Stereoisomeric distribution of some chiral sulfur-containing trace components of yellow passion fruits. J. Agric. Food Chem., 42, 2438–2441.
Zehentbauer, G. and Grosch, W. (1997) Apparatus for quantitative headspace analysis of characteristic odorants of baguettes. Z. Lebensm. Unters. Forsch., 205, 262–267.
Zehentbauer, G. and Grosch, W. (1998) Crust aroma of baguettes. I. Key odorants of baguettes prepared in two different ways. J. Cereal Sci., 28, 81–92.
Accepted December 17, 2000
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S.-J. Lee and A. C. Noble Use of Partial Least Squares Regression and Multidimensional Scaling on Aroma Models of California Chardonnay Wines Am. J. Enol. Vitic., September 1, 2006; 57(3): 363 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Bonaiti, F. Irlinger, H. E. Spinnler, and E. Engel An Iterative Sensory Procedure to Select Odor-Active Associations in Complex Consortia of Microorganisms: Application to the Construction of a Cheese Model J Dairy Sci, May 1, 2005; 88(5): 1671 - 1684. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||