5.1 Foreword
5.1.1 Concept Delineation
When food is consumed, the interaction of taste, odor and textural feeling provides an overall sensation which is best defined by the English word “flavor”. German and some other languages do not have an adequate expression for such a broad and comprehensive term. Flavor results from compounds that are divided into two broad classes: Those responsible for taste and those responsible for odors, the latter often designated as aroma substances. However, there are compounds which provide both sensations.
Compounds responsible for taste are generally nonvolatile at room temperature. Therefore, they interact only with taste receptors located in the taste buds of the tongue. The four important basic taste perceptions are provided by: sour, sweet, bitter and salty compounds. They are covered in separate sections (cf., for example, 8.10, 22.3, 1.2.6, 1.3.3, 4.2.3 and 8.8). Glutamate stimulates the fifth basic taste (cf. 8.6.1).
Aroma substances are volatile compounds which are perceived by the odor receptor sites of the smell organ, i. e. the olfactory tissue of the nasal cavity. They reach the receptors when drawn in through the nose (orthonasal detection) and via the throat after being released by chewing (retronasal detection). The concept of aroma substances, like the concept of taste substances, should be used loosely, since a compound might contribute to the typical odor or taste of one food, while in another food it might cause a faulty odor or taste, or both, resulting in an off-flavor.
5.1.2 Impact Compounds of Natural Aromas The amount of volatile substances present in food is extremely low (ca. 10–15 mg/kg). In general, however, they comprise a large number of
components. Especially foods made by thermal processes, alone (e. g., coffee) or in combination with a fermentation process (e. g., bread, beer, cocoa, or tea), contain more than 800 volatile compounds. A great variety of compounds is often present in fruits and vegetables as well.
All the known volatile compounds are classified according to the food and the class of compounds and published in a tabular compilation (Nijssen, L. M. et al., 1999). A total of 7100 compounds in more than 450 foods are listed in the 1999 edi- tion, which is also available as a database on the internet.
Of all the volatile compounds, only a limited number are important for aroma. Compounds that are considered as aroma substances are prima-
Table 5.1. Examples of key odorants
Compound Aroma Occurrence
(R)-Limonene Citrus-like Orange juice (R)-1-p-Menthene- Grapefruit- Grapefruit juice
8-thiol like
Benzaldehyde Bitter Almonds,
almond-like cherries, plums Neral/geranial Lemon-like Lemons 1-(p-Hydroxy- Raspberry- Raspberries
phenyl)-3-butanone like (raspberry ketone)
(R)-(−)-1-Octen-3-ol Mushroom- Champignons,
like Camembert
cheese (E,Z)-2,6- Cucumber- Cucumbers
Nonadienal like
Geosmin Earthy Beetroot
trans-5-Methyl-2- Nut-like Hazelnuts hepten-4-one
(filbertone)
2-Furfurylthiol Roasted Coffee 4-Hydroxy-2,5- Caramel- Biscuits,
dimethyl-3(2H)- like dark beer,
furanone coffee
2-Acetyl-1-pyrroline Roasted White-bread crust
H.-D. Belitz · W. Grosch · P. Schieberle, Food Chemistry 340
© Springer 2009
5.1 Foreword 341
rily those which are present in food in concen- trations higher than the odor and/or taste thresh- olds (cf. “Aroma Value”, 5.1.4). Compounds with concentrations lower than the odor and/or taste thresholds also contribute to aroma when mix- tures of them exceed these thresholds (for ex- amples of additive effects, see 3.2.1.1, 20.1.7.8, 21.1.3.4).
Among the aroma substances, special attention is paid to those compounds that provide the charac- teristic aroma of the food and are, consequently, called key odorants (character impact aroma com- pounds). Examples are given in Table 5.1.
In the case of important foods, the differentiation between odorants and the remaining volatile com- pounds has greatly progressed. Important find- ings are presented in the section on “Aroma” in the corresponding chapters.
5.1.3 Threshold Value
The lowest concentration of a compound that is just enough for the recognition of its odor is called the odor threshold (recognition threshold).
The detection threshold is lower, i. e., the concen- tration at which the compound is detectable but the aroma quality still cannot be unambiguously established. The threshold values are frequently determined by smelling (orthonasal value) and by tasting the sample (retronasal value). With a few exceptions, only the orthonasal values are given in this chapter. Indeed, the example of the carbonyl compounds shows how large the difference between the ortho- and retronasal thresholds can be (cf. 3.7.2.1.9).
Threshold concentration data allow comparison of the intensity or potency of odorous substances.
The examples in Table 5.2 illustrate that great differences exist between individual aroma com- pounds, with an odor potency range of several or- ders of magnitude.
In an example provided by nootkatone, an es- sential aroma compound of grapefruit peel oil (cf. 18.1.2.6.3), it is obvious that the two enan- tiomers (optical isomers) differ significantly in their aroma intensity (cf. 5.2.5 and 5.3.2.4) and, occasionally, in aroma quality or character.
The threshold concentrations (values) for aroma compounds are dependent on their vapor pres- sure, which is affected by both temperature and
Table 5.2. Odor threshold values in water of some aroma compounds (20◦C)
Compound Threshold value
(mg/l)
Ethanol 100
Maltol 9
Furfural 3.0
Hexanol 2.5
Benzaldehyde 0.35
Vanillin 0.02
Raspberry ketone 0.01
Limonene 0.01
Linalool 0.006
Hexanal 0.0045
2-Phenylethanal 0.004
Methylpropanal 0.001
Ethylbutyrate 0.001
(+)-Nootkatone 0.001
(-)-Nootkatone 1.0
Filbertone 0.00005
Methylthiol 0.00002
2-Isobutyl-3-methoxypyrazine 0.000002 1-p-Menthene-8-thiol 0.00000002
medium. Interactions with other odor-producing substances can result in a strong increase in the odor thresholds. The magnitude of this effect is demonstrated in a model experiment in which the odor thresholds of compounds in water were determined in the presence and absence of 4-hydroxy-2,5-dimethyl-3(2H)-furanone (HD3F).
The results in Table 5.3 show that HD3F does not influence the threshold value of 4-vinylguaiacol.
However, the threshold values of the other odor-
Table 5.3. Influence of 4-hydroxy-2,5-dimethyl-3(2H)- furanone (HD3F) on the odor threshold of aroma sub- stances in water
Compound Threshold value (µg/1) Ratio
Ia IIb II to I
4-Vinylguaiacol 100 90 ≈1
2,3-Butanedione 15 105 7
2,3-Pentanedione 30 150 5
2-Furfurylthiol 0.012 0.25 20
β-Damascenone 2×10−3 0.18 90
aI, odor threshold of the compound in water.
b II, odor threshold of the compound in an aqueous HD3F solution having a concentration (6.75 mg/1, aroma value A= 115) as high as in a coffee drink.
Table 5.4. Comparison of threshold valuesa in water and beer
Compound Threshold (mg/kg) in
Water Beer
n-Butanol 0.5 200
3-Methylbutanol 0.25 70
Dimethylsulfide 0.00033 0.05 (E)-2-Nonenal 0.00008 0.00011
aOdor and taste.
ants increase in the presence of HD3F. This effect is the greatest in the case ofβ-damascenone, the threshold value being increased by a factor of 90.
Other examples in this book which show that the odor threshold of a compound increases when it is influenced by other odor-producing substances are a comparison of the threshold values in water and beer (cf. Table 5.4) as well as in water and in aqueous ethanol (cf. 20.2.6.9).
5.1.4 Aroma Value
As already indicated, compounds with high
“aroma values” may contribute to the aroma of foods. The “aroma value” Axof a compound is calculated according to the definition:
Ax=cx
ax
(5.1) (cx: concentration of compound X in the food, ax: odor threshold (cf. 5.1.3) of compound X in the food). Methods for the identification of the corresponding compounds are described under Section 5.2.2.
The evaluation of volatile compounds on the basis of the aroma value provides only a rough pattern at first. The dependence of the odor intensity on the concentration must also be taken into account.
In accordance with the universally valid law of Stevens for physiological stimuli, it is formulated as follows:
E= k · (S − So)n (5.2)
E: perception intensity, k: constant, S: concentra- tion of stimulant, So: threshold concentration of stimulant.
The examples presented in Fig. 5.1 show that the exponent n and, therefore, the dependency of the odor intensity on the concentration can vary substantially. Within a class of compounds, the range of variations is not very large, e. g., n= 0.50−0.63 for the alkanals C4–C9.
In addition, additive effects that are difficult to assess must also be considered. Examinations of mixtures have provided preliminary information.
They show that although the intensities of com- pounds with a similar aroma note add up, the in- tensity of the mixture is usually lower than the sum of the individual intensities (cf. 3.2.1.1). For substances which clearly differ in their aroma note, however, the odor profile of a mixture is composed of the odor profiles of the components added together, only when the odor intensities are approximately equal. If the concentration ratio is such that the odor intensity of one component pre- dominates, this component then largely or com- pletely determines the odor profile.
Examples are (E)-2-hexenal and (E)-2-decenal which have clearly different odor profiles (cf. Fig.
5.2 a and 5.2 f). If the ratio of the odor intensities is approximately one, the odor notes of both aldehydes can be recognized in the odor profile of the mixture (Fig. 5.2 d). But if the dominating odor intensity is that of the decenal (Fig. 5.2 b), or of the hexenal (Fig. 5.2 e), that particular note determines the odor profile of the mixture.
Fig. 5.1. Relative odor intensity Irel(reference: n-buta- nol) as a function of the stimulant concentration (ac- cording to Dravnieks, 1977).
Air saturated with aroma substance was diluted.•−•−
•α-pinene,◦−◦−◦ 3-methylbutyric acid methyl ester,
− − hexanoic acid, − − 2,4-hexadienal,
− − hexylamine
5.1 Foreword 343
Fig. 5.2. Odor profiles of (E)-2-decenal (D), (E)-2- hexenal (H) and mixtures of both aldehydes (accord- ing to Laing and Willcox, 1983). The following concen- trations (mg/kg) dissolved in di-2-ethylhexyl-phthalate were investigated: 50 (D); 2 (H1); 3.7 (H2); 11 (H3) and 33 (H4).
ID and IH: Odor intensity of each concentration of (E)-2-decenal and (E)-2-hexenal. Odor quality: 1, warm; 2, like clean washing; 3, cardboard; 4, oily, fatty; 5, stale; 6, paint; 7, candle; 8, rancid; 9, stinkbug;
10, fruity; 11, apple; 12, almond; 13, herbal, green;
14, sharp, pungent; 15, sweet; 16, banana; 17, floral.
The broken line separates the aroma qualities of (E)-2- decenal (left side) and (E)-2-hexenal
The mixture in Fig. 5.2, c gives a new odor pro- file because definite features of the decenal (stale, paint-like, rancid) and the hexenal (like apples, al- monds, sweet) can no longer be recognized in it.
The examples show clearly that the aroma profiles of foods containing the same aroma substances can be completely dissimilar owing to quanti- tative differences. For example, changes in the recipe or in the production process which cause alterations in the concentrations of the aroma sub- stances can interfere with the balance in such a way that an aroma profile with unusual char- acteristics is obtained.
5.1.5 Off-Flavors, Food Taints
An off-flavor can arise through foreign aroma substances, that are normally not present in a food, loss of key odorants, or changes in the concentration ratio of individual aroma substances. Figure 5.3 describes the causes for aroma defects in food. In the case of an odorous contaminant, which enters the food via the air or water and then gets enriched, it can be quite difficult to determine its ori- gin if the limiting concentration for odor perception is exceeded only on enrichment.
Examples of some off-flavors that can arise during food processing and storage are listed in Table 5.5. Examples of microbial metabolites wich may be involved in pigsty-like and earthy- muddy off-flavors are skatole (I; faecal-like, 10 µ g/kg∗), 2-methylisoborneol (II; earthy- muddy, 0.03 µ g/kg∗) and geosmin (III; earthy, (−): 0.01 µ g/kg∗;(+): 0.08 µg/kg∗):
(5.3) 2,4,6-Trichloroanisole (IV) with an extremely low odor threshold (mouldy-like: 3.10−5 µ g/kg, water) is an example of an off-flavor substance (cf. 20.2.7) which is produced by fungal degra- dation and methylation of pentachlorophenol fungicides.
To a certain extent, unwanted aroma substances are concealed by typical ones. Therefore, the threshold above which an off-flavor be- comes noticeable can increase considerably in food compared to water as carrier, e. g., up to 0.2 µg/kg 2,4,6-trichloroanisole in dried fruits.
∗Odor threshold in water.
Table 5.5. “Off-flavors” in food products
Food product Off-flavor Cause
Milk Sunlight flavor Photooxidation of methionine to methional (with riboflavin as a sensitizer)
Milk powder Bean-like The level of O3in air too high: ozonolyis of 8,15- and 9,15-isolinoleic acid to
6-trans-nonenal
Milk powder Gluey Degradation of tryptophan to o-amino- acetophenone
Milk fat Metallic Autoxidation of pentaene- and hexaene fatty acids to octa-1,cis-5-dien-3-one
Milk products Malty Faulty fermentation by Streptococcus lactis, var. maltigenes; formation of phenylacetaldehyde and 2-phenylethanol from phenylalanine
Peas, Hay-like Saturated and unsaturated aldehydes,
deep froze octa-3,5-dien-2-one, 3-alkyl-2-methoxypyrazines,
hexanal
Orange juice Grapefruit note Metal-catalyzed oxidation or photooxidation of valencene to nootkatone
Orange juice Terpene note d-Limonene oxidation to carvone
Passion Aroma flattening Oxidation of (6-trans-2- fruit juice during pasteurization trans)-6-(but-2-enyliden)-1,5,5-
trimethylcyclohex-1-ene to 1,1,6-trimethyl-1,2- dihydronaphthalene:
Beer Sunlight flavor Photolysis of humulone: reaction of one degradation product with hydrogen sulfide yielding 3-methyl-2-buten-1-thiol Beer Phenolic note Faulty fermentation: hydrocinnamic acid
decarboxylation by Hafnia protea
5.2 Aroma Analysis 345
Fig. 5.3. The cause of aroma defects in food
5.2 Aroma Analysis
The aroma substances consist of highly di- versified classes of compounds, some of them being highly reactive and are present in food in extremely low concentrations. The diffi- culties usually encountered in qualitative and quantitative analysis of aroma compounds are based on these features. Other difficulties are associated with identification of aroma com- pounds, elucidation of their chemical structure and characterization of sensory properties.
The results of an aroma analysis can serve as an objective guide in food processing for assess- ing the suitability of individual processing steps, and for assessing the quality of raw material, intermediate- and endproducts. In addition, inves- tigation of food aroma broadens the possibility of food flavoring with substances that are prepared synthetically, but are chemically identical to those found in nature, i. e. the so-called “nature identi- cal flavors” (cf. 5.5).
The elucidation of the aroma of any food is car- ried out stepwise; the following instrumental and sensory analyses are conducted:
• Isolation of the volatile compounds
• Differentiation of the aroma substances from the remaining components of the volatile frac- tion by dilution analyses
• Concentration and identification
• Quantification and calculation of aroma values
• Simulation of the aroma on the basis of the analytical results
• Omission experiments
5.2.1 Aroma Isolation
The amount of starting material must be selected to detect even those aroma substances which are present in very low concentrations (ppb range), but contribute considerably to the aroma because of still lower odor thresholds. The volatile com- pounds should be isolated from food using gentle methods because otherwise artifacts can easily be produced by the reactions listed in Table 5.6.
Additional difficulties are encountered in foods which retain fully-active enzymes, which can alter the aroma. For example, during the homoge- nization of fruits and vegetables, hydrolases split the aroma ester constituents, while lipoxygenase, together with hydroperoxide lyase, enrich the aroma with newly-formed volatile compounds.
To avoid such interferences, tissue disintegration is done in the presence of enzyme inhibitors, e. g., CaCl2 or, when possible, by rapid sample preparation. It is useful in some cases to inhibit enzyme-catalyzed reactions by the addition of methanol or ethanol. However, this can result in a change in aroma due to the formation of esters and acetals from acids and aldehydes respectively.
Table 5.6. Possible changes in aromas during the isolation of volatile compounds Reaction
Enzymatic
1. Hydrolysis of esters (cf. 3.7.1)
2. Oxidative cleavage of unsaturated fatty acids (cf. 3.7.2.3) 3. Hydrogenation of aldehydes (cf. 5.3.2.1)
Non-enzymatic
4. Hydrolysis of glycosides (cf. 5.3.2.4 and 3.8.4.4) 5. Lactones from hydroxy acids
6. Cyclization of di-, tri-, and polyols (cf. 5.3.2.4) 7. Dehydration and rearrangement of tert-allyl alcohols 8. Reactions of thiols, amines, and aldehydes (cf. 5.3.1.4) in the
aroma concentrate
9. Reduction of disulfides by reductones from the Maillard reaction 10. Fragmentation of hydroperoxides
At the low pH values prevalent in fruit, non- enzymatic reactions, especially reactions 4–7 shown in Table 5.6, can interfere with the isolation of aroma substances by the formation of artifacts. In the concentration of isolates from heated foods, particularly meat, it cannot be excluded that reactive substances, e. g., thiols, amines and aldehydes, get concentrated to such an extent that they condense to form heterocyclic aroma substances, among other compounds (Reaction 8, Table 5.6).
In the isolation of aroma substances, foods which owe their aroma to the Maillard reaction should not be exposed to temperatures of more than 50◦C. At higher temperatures, odorants are additionally formed, i. e., thiols in the reduction of disulfides by reductones. Fats and oils contain volatile and non-volatile hydroperoxides which fragment even at temperatures around 40◦C.
An additional aspect of aroma isolation not to be neglected is the ability of the aroma substances to bind to the solid food matrix. Such binding ability differs for many aroma constituents (cf. 5.4).
The aroma substances present in the vapor space above the food can be very gently detected by headspace analysis (cf. 5.2.1.3). However, the amounts of substance isolated in this process are so small that important aroma substances, which are present in food in very low con- centrations, give no detector signal after gas chromatographic separation of the sample. These substances can be determined only by sniffing the carrier gas stream. The difference in the
detector sensitivity is clearly shown in Fig. 5.4, taking the aroma substances of the crust of white bread as an example. The gas chromatogram does not show, e. g., 2-acetyl-1-pyrroline and 2-ethyl-3,5-dimethylpyrazine, which are of great importance for aroma due to high FD factors in the FD chromatogram (definition in 5.2.2.1).
These aroma substances can be identified only after concentration from a relatively large amount of the food.
5.2.1.1 Distillation, Extraction
The volatile aroma compounds, together with some water, are removed by vacuum distilla- tion from an aqueous food suspension. The highly volatile compounds are condensed in an efficiently cooled trap. The organic com- pounds contained in the distillate are separated from the water by extraction or by adsorption to a hydrophobic matrix and reversed phase chromatography and then prefractionated.
The apparatus shown in Fig. 5.5 is recommended for the gentle isolation of aroma substances from aqueous foods by means of distillation. In fact, a condensate can be very quickly obtained be- cause of the short distances. As in all distilla- tion processes, the yield of aroma substances de- creases if the food or an extract is fatty (Ta- ble 5.7).
After application of high vacuum (≈5 mPa) the distillation procedure is started by dropping the
5.2 Aroma Analysis 347
Fig. 5.4. Headspace analysis of aroma substances of white-bread crust. a Capillary gas chromatogram (the arrows mark the positions of the odorants), b FD chromatogram. Odorants: 1 methylpropanal, 2 diacetyl, 3 3-methylbutanal, 4 2,3-pentanedione, 5 butyric acid, 6 2-acetyl-1-pyrroline, 7 1-octen-3-one, 8 2-ethyl-3,5- dimethylpyrazine, 9 (E)-2-nonenal (according to Schieberle and Grosch, 1992)
Table 5.7. Yields of aroma substances on distillation under vacuuma
Aroma substance (amount)a Yieldb (%) Model I Model II
3-Methylbutyric acid 91 31
(1.9 µg)
Phenylacetaldehyde 84 21
(4.2 µg)
3-Hydroxy-4,5-dimethyl- 100 3.3 2(5H)-furanone (2.2 µg)
2-Phenylethanol (3.7 µg) 100 10.7 (E,E)-2,4-Decadienal 100 3.4
(1.4 µg)
(E)-β-Damascenone 100 2.8
(0.9 µg)
Vanillin (3.7 µg) 100 0.4
a Amount in the model solution: I in diethylether (50 ml), II in a mixture of diethylether (50 ml) and triglycerides (50 ml)
bDistillation in the apparatus shown in Fig. 5.5 at 35◦C
liquid food or the extract from the funnel (1 in Fig. 5.5) into the distillation flask which is heated to 35–40◦C in a water bath (2). The volatiles in- cluding the solvent vapor are transferred into the distillation head (3). The distillate is condensed by liquid nitrogen in the receiver (4). The Dewar flask (5) protects the vacuum pump (reduced pres- sure 10−3Pa).
Fig. 5.5. Apparatus for the distillation of aroma sub- stances from foods (for explanation, see text. Accord- ing to Engel et al., 1999)
Solid foods are first extracted, the addition of wa- ter may be required to increase the yield of aroma substances.
An extraction combined with distillation can be achieved using an apparatus designed by Likens–
Nickerson (Fig. 5.6).
In this process, low-boiling solvents are usually used to make subsequent concentration of the aroma substances easier. Therefore, this process is carried out at normal pressure or slightly reduced pressure. The resulting thermal treatment of the food can lead to reactions (examples in Table 5.6) that change the aroma composition. The example in Table 5.8 shows the extent to which some aroma substances are released from glycosides during simultaneous distillation/extraction.
Fig. 5.6. Apparatus according to Likens and Nicker- son used for simultaneous extraction and distillation of volatile compounds.
1 Flask with heating bath containing the aqueous sam- ple, 2 flask with heating bath containing the solvent (e. g. pentane), 3 cooler, 4 condensate separator: extract is the upper and water the lower phase
Table 5.8. Isolation of odorants from cherry juice – Comparison of distillation in vacuo (I) with simultan- eous distillation and extraction (II)
Odorant I (µg/1) II
Benzaldehyde 202 5260
Linalool 1.1 188
Table 5.9. Relative retention time (trel) of some com- pounds separated by gas chromatography using Pora- pak Q as stationary phase (Porapak: styrene divinylben- zene polymer; T: 55◦C)
Compound trel Compound trel
Water 1.0 Methylthiol 2.6
Methanol 2.3 Ethylthiol 20.2
Ethanol 8.1 Dimethylsulfide 19.8 Acetaldehyde 2.5 Formic acid
Propanal 15.8 ethyl ester 6.0
5.2.1.2 Gas Extraction
Volatile compounds can be isolated from a solid or liquid food sample by purging the sample with an inert gas (e. g., N2, CO2, He) and adsorb- ing the volatiles on a porous, granulated polymer (Tenax GC, Porapak Q, Chromosorb 105), fol- lowed by recovery of the compounds. Water is retarded to only a negligible extent by these poly- mers (Table 5.9). The desorption of volatiles is usually achieved stepwise in a temperature gradi- ent. At low temperatures, the traces of water are removed by elution, while at elevated tempera- tures, the volatiles are released and flushed out by a carrier gas into a cold trap, usually connected to a gas chromatograph.
5.2.1.3 Headspace Analysis
The headspace analysis procedure is simple: the food is sealed in a container, then brought to the desired temperature and left for a while to estab- lish an equilibrium between volatiles bound to the food matrix and those present in the vapor phase.
A given volume of the headspace is withdrawn with a gas syringe and then injected into a gas chromatograph equipped with a suitable separa- tion column (static headspace analysis). Since the water content and an excessively large volume of the sample substantially reduce the separation ef- ficiency of gas chromatography, only the major volatile compounds are indicated by the detector.
The static headspace analysis makes an important contribution when the positions of the aroma sub-
5.2 Aroma Analysis 349
stances in the chromatogram are determined by olfactometry (cf. 5.2.2.2).
More material is obtained by dynamic head- space analysis or by solid phase microextrac- tion (SPME). In the dynamic procedure the headspace volatiles are flushed out, adsorbed and thus concentrated in a polymer, as outlined in 5.2.1.2. However, it is difficult to obtain a representative sample by this flushing proce- dure, a sample that would match the original headspace composition. A model system assay (Fig. 5.7) might clarify the problems. Samples (e) and (f) were obtained by adsorption on different polymers. They are different from each other and differ from sample (b), which was obtained
Fig. 5.7. A comparison of some methods used for aroma compound isolation (according to Jennings and Filsoof, 1977).
a a Ethanol, b 2-pentanone, c heptane, d pentanol, e hexanol, f hexyl formate, g 2-octanone, h d-limonene, i heptyl acetate and k γ-heptalactone. b Headspace analysis of aroma mixture a. c From aroma mixture 10 µl is dissolved in 100 ml water and the headspace is analyzed. d As in c but the water is saturated with 80% NaCl. e As in c but purged with nitrogen and trapped by Porapak Q. f As in c but purged with ni- trogen and trapped by Tenax GC. g As in e but distilled and extracted (cf. Fig. 5.6)
directly for headspace analysis. The results might agree to a greater extent by varying the gas flush- ing parameters (gas flow, time), but substantial differences would still remain. A comparison of samples (a) and (g) in Fig. 5.7 shows that the results obtained by the distillation-extraction procedure give a relatively good representation of the composition of the starting solution, with the exception of ethanol. However, the formation of artifacts is critical (cf. 5.2.1.1).
SPME is based on the partitioning of compounds between a sample and a coated fiber immersed in it. The odorants are first adsorbed onto the fiber (e. g. nonpolar polydimethylsilo-xane or polar polyacrylate) immersed in a liquid food, a food extract or in the headspace above a food sample for a certain period of time. After adsorption is completed, the compounds are thermally desorbed into a GC injector block for further analysis.
Particularly in food applications headspace SPME is preferred to avoid possible contamination of the headspace system by non-volatile food com- ponents. Also SPME analysis is quite sensitive to experimental conditions. In addition to the stationary phase, sample, volume concentration of odorants, sample matrix and uniformity as well as temperature and time of the adsorption and desorption processes influence the yield.
In quantitative SPME analysis these influences are eliminated by the use of labelled internal standards (cf. 5.2.6.1).
5.2.2 Sensory Relevance
In many earlier studies on the composition of aro- mas, each volatile compound was regarded as an aroma substance. Although lists with hundreds of compounds were obtained for many foods, it was still unclear which of the volatiles were re- ally significant as odorants and to what extent im- portant odorants occurring in very low concentra- tions were detected.
The studies meanwhile concentrate on those com- pounds which significantly contribute to aroma.
The positions of these compounds in the gas chro- matogram are detected with the help of dilution analyses. Here, both of the following methods based on the aroma value concept (cf. 5.1.4) find application.
5.2.2.1 Aroma Extract Dilution Analysis (AEDA)
In AEDA, the concentrate of the odorants obtained by distillation is separated by gas chro- matography on a capillary column. To determine the retention times of the aroma substances, the carrier gas stream is subjected to sniffing detection after leaving the capillary column (GC/olfactometry). The sensory assessment of a single GC run, which is often reported in the literature, is not very meaningful because the perception of aroma substances in the carrier gas stream depends on limiting quantities which have nothing to do with the aroma value, e. g., the amount of food analysed, the degree of concen- tration of the volatile fraction, and the amount of sample separated by gas chromatography.
These limitations are eliminated by the stepwise dilution of the volatile fraction with solvent, fol- lowed by the gas chromatographic/olfactometric analysis of each dilution. The process is contin- ued until no more aroma substance can be de- tected by GC olfactometry. In this way, a dilution factor 2n(n= number of 1+1 dilutions) is deter- mined for each aroma substance that appears in the gas chromatogram. It is designated as the fla- vor dilution factor (FD factor) and indicates the number of parts of solvent required to dilute the aroma extract until the aroma value is reduced to one.
Another more elaborate variant of the dilution analysis requires, in addition, that the duration of each odor impression is recorded by a computer and CHARM values are calculated (CHARM:
acronym for combined hedonic response meas- urement), which are proportional to aroma values.
The result of an AEDA can be represented as a diagram. The FD factor is plotted against the retention time in the form of the retention in- dex (RI) and the diagram is called a FD chro- matogram.
The FD chromatograms of the volatile com- pounds of white bread and French fries are presented in Fig. 5.4 and 5.8, respectively.
The identification experiments concentrate on those aroma substances which appear in the FD chromatogram with higher FD factors. To detect all the important aroma substances, the range of FD factors taken into account must not be too narrowly set at the lower end because
differences in yield shift the concentration ratios.
Labile compounds can undergo substantial losses and when distillation processes are used, the yield decreases with increasing molecular weight of the aroma substances.
In the case of French fries (Fig. 5.8), 19 aroma substances appearing in the FD-factor range 21–27 were identified (cf. legend of Fig. 5.8).
Based on the high FD factors, the first approxi- mation indicates that methional, 2-ethyl-3,5- dimethylpyrazine, 2,3-diethyl-5-methylpyrazine and (E,E)-2,4-decadienal substantially contribute to the aroma of French fries.
5.2.2.2 Headspace GC Olfactometry
In the recovery of samples for AEDA, highly volatile odorants are lost or are covered by the solvent peak in gas chromatography, e. g., methanethiol and acetaldehyde. For this reason, in addition to AEDA, a sample is drawn from the gas space above the food, injected into the gas chromatograph, transported by the carrier gas stream into a cold trap and concentrated there, as shown in Fig. 5.9. After quick evaporation, the sample is flushed into a capillary column by the carrier gas and chromatographed. At the end of the capillary, the experimentor sniffs the carrier gas stream and determines the positions of the chromatogram at which the odorants appear. The gas chromatogram is simultaneously monitored by a detector.
To carry out a dilution analysis, the volume of the headspace sample is reduced stepwise until no odorant is detectable by gas chromatogra- phy/olfactometry. In our example with French fries (Fig. 5.10), e. g., the odors of methanethiol, methylpropanal and dimethyltrisulfide were detectable in the sixth dilution, but only methan- ethiol was detectable in the seventh. The eighth dilution was odorless. Further experiments showed that methanethiol does in fact belong to the key odorants of French fries.
In GC/olfactometry, odor thresholds are consid- erably lower than in solution because the aroma substances are subjectied to sensory assessment in a completely vaporized state. The examples given in Table 5.10 show how great the differ- ences can be when compared to solutions of the aroma substances in water.
5.2 Aroma Analysis 351
Fig. 5.8. FD chromatogram of the volatile fraction of French fries. Ordinate: n, number of 1+ 1 dilutions.
Abscissa: retention index (RI) on the capillary SE-54. The following odorants were identified: 1 methional, 2 2-acetyl-1-pyrroline, 3 dimethyltrisulfide, 4 1-octen-3-one, 5 phenylacetaldehyde, 6 2-ethyl-3,6-dimethyl- pyrazine, 7 2-ethyl-3,5-dimethylpyrazine, 8 nonanal, 9 (Z)-2-nonenal, 10 2,3-diethyl-5-methylpyrazine, 11 (E)- 2-nonenal, 12 2-ethenyl-3-ethyl-5-methylpyrazine, 13 2-isobutyl-3-methoxypyrazine, 14 dimethyltetrasulfide, 15 (E,E)-2,4-nonadienal, 16 (Z)-2-decenal, 17 (E,Z)-2,4-decadienal, 18 (E,E)-2,4-decadienal, 19 trans-4,5-epoxy- (E)-2-decenal (according to Wagner and Grosch, 1997)
Table 5.10. Odor thresholds of aroma substances in air and water
Compound Odor thresholds in Air (a) Water (b) b/a (ng/I) (µg/l)
β-Damascenone 0.003 0.002 6.7 ×102
Methional 0.12 0.2 1.6 ×103
2-Methylisoborneol 0.009 0.03 3.3 ×103 2-Acetyl-1-pyrroline 0.02 0.1 5×103
4-Vinylguaiacol 0.6 5 8.3 ×103
Linalool 0.6 6 1.0 ×104
Vanillin 0.9 20 2.2 ×104
4-Hydroxy-2,5- 1.0 30 3×104
dimethyl-3(2H)- furanone (furaneol)
5.2.3 Enrichment
When an aroma concentrate contains phenols, or- ganic acids or bases, preliminary separation of these compounds from neutral volatiles by extrac- tion with alkali or acids is advantageous.
The acidic, basic and neutral fractions are indi- vidually analyzed. The neutral fraction by itself consists of so many compounds that in most cases not even a gas chromatographic column with the highest resolving power is able to separate them into individual peaks. Thus, separation of the neu- tral fraction is advisable and is usually achieved by liquid chromatography, or preparative gas or high performance liquid chromatography. A pre- liminary separation of aroma extracts is achieved
Fig. 5.9. Apparatus for the gas chromatography–olfactometry of static headspace samples. 1 Sample in ther- mostated glass vessel, 2 septum, 3 gastight syringe, 4 injector, 5 hydrophobed glass tube, 6 carrier gas, e. g., he- lium, 7 purge and trap system, 8 cold trap, 9 gas chromatograph with capillary column, 10 sniffing port, 11 flame ionization detector (according to Guth and Grosch, 1993)
Fig. 5.10. FD chromatogram of static headspace samples of French fries. Ordinate: n, number of 1+ 1 dilutions.
Abscissa: retention index (RI) on the capillary SE-54. The following odorants were identified: 1 methanethiol, 2 methylpropanal, 3 2,3-butanedione, 4 3-methylbutanal, 5 2-methylbutanal, 6 2,3-pentanedione, 7 hexanal, 8 me- thional, 9 2-acetyl-1-pyrroline, 10 dimethyltrisulfide (according to Wagner and Grosch, 1997)
5.2 Aroma Analysis 353
by chromatography on silica gel, as shown in Ta- ble 5.11 for coffee aroma. To localize the aroma substances each of the four fractions is ana- lyzed by gas chromatography and olfactometry.
Some volatile compounds are aroma active in such low concentrations that even enrichment by column chromatography does not allow identifi- cation, e. g., 3-methyl-2-butenethiol (Fraction A in Table 5.11) and the two methoxypyrazines (Fraction B) in coffee. In most cases, further concentration is achieved with the help of multi- dimensional gas chromatography (MGC). The fraction which contains the unknown aroma
Table 5.11. Column chromatographic preliminary separation of an aroma extract of roasted coffee FractionaAroma substance
A 2-Methyl-3-furanthiol, 2-furfurylthiol, bis(2-methyl-3-furyl)disulfide, 3-methyl-2-butenethiol
B 2,3-Butanedione, 3-methylbutanal, 2,3-pentanedione, trimethylthiazole, 3-mercapto-3-methylbutylformiate, 3-isopropyl-2-methoxypyrazine, phenylacetaldehyde,
3-isobuty1-2-methoxypyrazine, 5-methyl-5(H)-cyclopentapyrazine, p-anisaldehyde,
(E)-β-damascenone
C Methional, 2-ethenyl-3,5-dimethylpyrazine, linalool, 2,3-diethyl-5-methylpyrazine, guaiacol, 2-ethenyl-3-ethyl-
5-methylpyrazine,
4-ethylguaiacol, 4-vinylguaiacol
D 2-/3-Methylbutyric acid, trimethylpyrazine, 3-mercapto-3-methyl-1-butanol,
5-ethyl-2,4-dimethylthiazole,
2-ethyl-3,5-dimethylpyrazine, 3,4-dimethy1- 2-cyclopentenol-1-one, 4-hydroxy-2,5- dimethyl-3(2H)-furanone, 5-ethyl-4- hydroxy-2-methyl-3(2H)-furanone, 3-hydroxy-4,5-dimethy1-2(5H)-furanone, 5-ethyl-3-hydroxy-4-methyl-
2(5H)-furanone, vanillin
aChromatography at 10–12◦C on a silica gel column (24× 1 cm, deactivated with 7% water); elution with mixtures of pentane-diethylether (50 ml, 95+ 5, v/v, fraction A; 30 ml, 75× 25, v/v, Fraction B; 30 ml, 1+ 1, v/v, Fraction C) and with diethylether (100 ml, Fraction D).
substance is first subjected to preliminary sepa- ration on a polar capillary. The eluate containing the substance is cut out, rechromatographed on a non-polar capillary and finally analyzed by mass spectrometry. The MGC is also used in quantitative analysis for the preliminary purifica- tion of analyte and internal standard (cf. 5.2.6.1).
5.2.4 Chemical Structure
In the structure elucidation of aroma substances, mass spectrometry has become an indispensable tool because the substance amounts eluted by gas chromatography are generally sufficient for an evaluable spectrum. If the corresponding refer- ence substance is available, identification of the aroma substance is based on agreement of the mass spectrum, retention indices on at least two capillary columns of different polarity, and of odor thresholds, which are compared by gas chro- matography/olfactometry. If the reference sub- stance is not available, the following procedure is suitable for the identification of the odorant:
The analyte is concentrated until a 1H-NMR spectrum and, if necessary, a 13C-spectrum can be measured. An example is the identification of the characteristic odorant of roasted hazelnuts.
The mass spectrum of this substance (Fig. 5.11a) indicates an unsaturated carbonyl compound with a molar mass of 126. In conjunction with the structural elements shown by the1H-NMR spec- trum (Fig. 5.11b), it was proposed that the odor- ant is 5-methyl-(E)-2-hepten-4-one (Filbertone).
It goes without saying that the synthesis of the proposed aroma substance was part of the identi- fication. It was also guaranteed that its chromato- graphic and sensory properties correspond with those of the unknown odorant.
5.2.5 Enantioselective Analysis
In the case of chiral aroma substances, elu- cidation of the absolute configuration and determination of the enantiomeric ratio, which is usually given as the enantiomeric excess (ee), are of especial interest because the enantiomers of a compound can differ considerably in their odor quality and threshold. The compound 3a,4,5,7a-
Fig. 5.11. Instrumental analysis of 5-methyl-(E)-2-hepten-4-one (according to Emberger, 1985) (a) mass spec- trum, (b)1H-NMR spectrum (for discussion, see text)
tetrahydro-3,6-dimethyl-2(3H)-benzofuranone (wine lactone) represents an impressive example which shows how much the odor activity of enantiomers can vary. The four enantiomeric pairs of this compound have been separated by gas chromatography on a chiral phase (Fig. 5.12).
The 3S,3aS,7aR-enantiomer (No. 6 in Table 5.12) has the lowest odor threshold of the eight dia- stereomers. The identification of this substance in wine (cf. 20.2.6.9) led to the name wine
lactone. Two diastereomers (No. 3 and 8) are odorless.
The determination of the ee value can be used to detect aromatization with a synthetic chiral aroma substance because in many cases one enantiomer is preferentially formed in the biosynthesis of chi- ral aroma substances (examples in Table 5.13). In contrast to biosynthesis, chemical synthesis gives the racemate which is usually not separated for economic reasons. The addition of an aroma sub-
5.2 Aroma Analysis 355
Fig. 5.12. Gas chromatogram of the diastereomers of 3a,4,5,7a-tetrahydro-3,6-dimethyl-2(3H)-benzofuranone (wine lactone) on a chiral phase (according to Guth, 1997)
Table 5.12. Odor threshold values of diastereomeric 3a,4,5,7a-tetrahydro-3,6-dimethyl-2(3H)-benzo- furanone
No.a Stereoisomer- Odor threshold conformation (ng/l air)
1 (3S,3aS,7aS) 0.007–0.014
2 (3R,3aR,7aR) 14–28
3 (3R,3aR,7aS) >1000
4 (3R,3aS,7aS) 8–16
5 (3S,3aR,7aR) 0.05–0.2
6 (3S,3aS,7aR) 0.00001–0.00004
7 (3S,3aR,7aS) 80–160
8 (3R,3aS,7aR) >1000
aNumbering as in Fig. 5.12.
stance of this type can be determined by enantio- selective analysis if safe data on the enantiomeric excess of the compound in the particular food are available. It should also be taken into account that the ee value can change during food process-
ing, e. g., that of filbertone decreases during the roasting of hazelnuts (cf. Table 5.13).
Table 5.13. Enantiomeric excess (ee) of chiral aroma substances in some foods
Aroma substance Food ee (%)
R(+)-γ-Decalactone Peach, apricot,
mango, strawberry >80 pineapple, maracuya
R(+)-δ-Decalactone Milk fat 60
R(+)-trans-α-Ionone Raspberry 92.4
Carrot 90.0
Vanilla bean 94.2
R(-)-1-Octen-3-ol Mushroom, >90 chanterelle
S(+)-E-5-Methy1-2- Hazelnut, raw 60–68 hepten-4-one Hazelnut, roasted 40–45 (filbertone)
R-3-Hydroxy-4,5- Sherry ca. 30 dimethyl-2(5H)-
furanone (sotolon)
Fig. 5.13. Enantioselective gas chromatographic analy- sis of trans-α-ionone in aroma extracts of different rasp- berry fruit juice concentrates (according to Werkhoff et al., 1990): a and b samples with nature identical aroma, c natural aroma
The method frequently applied to determine ee values is the enantioselective gas chromato- graphic analysis of the aroma substance on a chiral phase, e. g., peralkylated cyclodextrins.
This method was used, e. g., to test raspberry fruit juice concentrates for unauthorized aromatization with trans-α-ionone. The gas chromatograms of trans-α-ionone from two different samples are shown in Fig. 5.13. The low excesses of the R-enantiomer of ee= 8% (concentrate A) and ee= 24% (B) can probably be put down to the addition of synthetic trans-α-ionone racemate to the fruit juice concentrate because in the natural aroma (C), the ee value is 92.4%.
5.2.6 Quantitative Analysis, Aroma Values 5.2.6.1 Isotopic Dilution Analysis (IDA)* The quantitative analysis of aroma substances us- ing conventional methods often gives incorrect values. The high vapor pressure, the poor ex- tractability especially of polar aroma substances from hydrous foods and the instability of import- ant aroma substances, e. g., thiols, can cause un- foreseeable losses in the purification of the sam- ples and in gas chromatography.
The results of quantitative analyses are exact (standard deviation<10%) and reproducible if the chemical structure of the internal standard is very similar to the structure of the analyte. An
*Most of the quantitative data on aroma substances in this book come from IDAs.
isotopomer of the analyte is the most similar. In this case, the physical and chemical properties of both substances correspond, except for a small isotope effect which can lead to partial separation in capillary gas chromatography.
The examples given in Fig. 5.14 show that for economic reasons, mostly internal standards la- belled with deuterium are synthesized for IDA.
The considerably more expensive carbon isotope 13 is introduced into the odorant (examples are the internal standards No. 11 and 12 in Fig. 5.14) only if a deuterium/protium exchange can occur in the course of analysis. This exchange would falsify the result. Another advantage of this iso- tope is the completely negligible isotope effect compared to deuterium.
It is easy to conduct an IDA because losses of analyte in the distillative recovery (cf. 5.2.1.1) and in purification do not influence the result since the standard suffers the same losses. These advantages of IDA are used in food chemistry for other analytes as well, e. g., pantothenic acid (cf. 6.3.5.2) or for the mycotoxin patulin (cf. 9.2.3).
The quantification of the odorants 2-furfuryl- thiol (FFT), 2-methyl-3-furanthiol (MFT) and 3-mercapto-2-pentanone (3M2P) in boiled meat will be considered as an example. Especially MFT and 3M2P are very instable, so after the addition of the deuterated standards d-FFT, d-MFT and d-3M2P (No. 13 in Fig. 5.14) to the extract, it is advisable to concentrate via a trap- ping reaction for thiols which is performed with p-hydroxymercuribenzoic acid. The analytes and their standards are displaced from the derivatives by cysteine in excess, separated by gas chro- matography, and analyzed by mass spectrometry.
In this process, mass chromatograms for the ions are monitored in which the analyte and its isotopomer differ (Fig. 5.15). After calibration, the mass chromatograms are evaluated via a com- parison of the areas of analyte and standard.
2-Mercapto-3-pentanone (2M3P) is also identi- fied in this analysis. However, this compound is of no importance for the aroma of boiled meat because of its lower concentration and higher odor threshold compared to those of 3M2P.
5.2.6.2 Aroma Values (AV)
To approach the situation in food aroma values (definition cf. 5.1.4) are calculated. It is assumed
5.2 Aroma Analysis 357
Fig. 5.14. Odorants labelled with deuterium () or carbon-13 () as internal standard substances for isotopic dilution analyses of the corresponding unlabelled odorants.
1 2-[α-2H2]furfurylthiol, 2 2-[2H3]methyl-3-furanthiol, 3 3-mercapto-2-[4,5-2H2]pentanone, 4 [4-2H3]methio- nal, 5 2-[2H3]ethyl-3,5-dimethylpyrazine, 6 (Z)-1,5-[5,6-2H2]octadien-3-one, 7 trans-4,5-epoxy-(E)-2-[6,72H4] decenal, 8 1-(2,6,6-[6,6-2H6]trimethyl-1,3-cyclohexadienyl)-2-buten-1-one (β-damascenone), 9 3a,4,5,7atetra- hydro-3,6-[3-2H3]dimethyl-2(3H)-benzofuranone (wine lactone), 10 tetrahydro-4-methyl-2-(2-methylpropenyl)- 2H-[3,4-2H3]pyran (sotolon), 11 4-hydroxy-2,5-[13C2]dimethyl-3(2H)-furanone, 12 3-hydroxy-4,5-[4-13C]di- methyl-2(5H)-[5-13C]furanone (rose oxide)
that the odorants showing higher AVs contribute strongly to the aroma of the food. For this pur- pose, the odor thresholds of the compounds dis- solved in water, in oil or applied to starch are used, depending on which of these materials dom- inates in the food.
An example are the AVs of the odorants of French fries based on their odor thresholds in an oil (Table 5.14). Methanethiol, methional, methyl- propanal and 2-methylbutanal exhibit the highest aroma values. Consequently, they should belong to the most important odorants of French fries.
5.2.7 Aroma Model, Omission Experiments Finally, the identified odorants must actually pro- duce the aroma in question. To test this, the de- termined concentrations of the odorants are dis-
solved in a suitable medium, which is not difficult in the case of liquid foods. The solvent for the re- combination mixture called the aroma model can be adapted to the food. An ethanol/water mixture, for example, is suitable for wine. In the case of solid foods, however, compromises have to be ac- cepted.
The aroma profile of the model is then compared to that of the food. In the example of French fries discussed in detail here, a very good approxima- tion of the original aroma was achieved.
The selection of odorants by dilution analyses (cf. 5.2.2) does not take into account additive (cf. 20.1.7.8) or antagonistic effects (example in Fig. 5.2) because the aroma substances, after separation by gas chromatography, are sniffed individually. Therefore, in view of the last men- tioned effect, the question arises whether all the compounds occurring in the aroma model really contribute to the aroma in question. To answer
Fig. 5.15. Isotopic dilution analysis of 2-furfurylthiol (FFT), 2-methy1-3-furanthiol (MFT) and 3-mercapto-2- pentanone (3M2P).
(a) Gas chromatogram, (b–g) mass chromatograms of the analytes and the deuterated (d) internal standards; traces of the ions shown in brackets were monitored: d-MFT (m/z 118), MFT (m/z 115), d-3M2P (m/z 121), 3M2P and 2M3P (m/z 119), d-FFT (m/z 83), FFT (m/Z 81) (according to Kerscher and Grosch, 1998)
this question, one or several aroma substances are omitted in the model and a triangle test is used to examine which of three samples (two complete and one reduced aroma model) offered to the testers in random order differs in aroma from the others. If a significant number of testers deter- mine a difference in the reduced model, it can be assumed that the odorants lacking in the reduced model contribute to the aroma and, consequently, belong to the key odorants of the food.
Some omission experiments, e. g., conducted with the aroma model for French fries, are shown in Table 5.15.
If methanethiol and the two decadienal isomers are missing in Experiments 1 and 2, the aroma
has no similarity to that of French fries. All five testers were in agreement. The Strecker aldehy- des with the malt odor (Exp. 3), 4,5-epoxydecenal (Exp. 4) and both pyrazines (Exp. 5) are also im- portant for the aroma because their absence was noticed by four of the five testers. 1-Octen-3-one, (Z)-2- and (E)-2-nonenal are of no importance for the aroma (Exp. 6). Surprisingly, this also applies to methional (Exp. 7) although it has the second highest aroma value (cf. Table 5.14) and smells of boiled potatoes. It is obvious that methional is masked by other odorants occurring in the aroma model. In French fries, the odor note “like boiled potatoes” is probably produced by methanethiol in combination with pyrazines.
5.3 Individual Aroma Compounds 359
Table 5.14. Volatile compounds with high aroma values in French friesa
Compound ConcentrationbOdor thresholdcAroma valued (µg/kg) (µg/kg)
Methanethiol 1240 0.06 2×104
Methional 783 0.2 3.9 ×103
Methylpropanal 5912 3.4 1.7 ×103
2-Methylbutanal 10599 10 1.1 ×103
trans-4,5-Epoxy-(E)-2-decenal 771 1.3 592
3-Methylbutanal 2716 5.4 503
(E,Z)-2,4-Decadienal 1533 4 383
4-Hydroxy-2,5-dimethyl-3 2778 25 111
(2H)-furanone
2,3-Diethyl-5-methylpyrazine 41 0.5 83
(E,E)-2,4-Decadienal 6340 180 35
2,3-Butanedione 306 10 31
2-Ethyl-3,5-dimethylpyrazine 42 2.2 19
2-Ethenyl-3-ethyl-5-methylpyrazine 5.4 0.5 11
3-Isobutyl-2-methoxypyrazine 8.6 0.8 11
2-Ethyl-3,6-dimethylpyrazine 592 57 10
aPotato sticks deep-fried in palm oil.
bResults of IDA.
cOdor threshold of the compound dissolved in sunflower oil.
dQuotient of concentration and odor threshold.
Table 5.15. Aroma model for French fries as affected by the absence of one or more odorantsa
Exp. Odorant omitted Numberb
No. in the model
1 Methanethiol 5
2 (E,Z)-2,4-Decadienal and 5 (E,E)-2,4-decadienal
3 Methylpropanal, 2- and 4
3-methylbutanal
4 trans-4,5-Epoxy-(E)-2-decenal 4 5 2-Ethyl-3,5-dimethylpyrazine 4
and 3-ethyl-2,5-dimethylpyrazine 6 1-Octen-3-one, (Z)-2- and 1
(E)-2-nonenal
7 Methional 0
aModels lacking in one or more components were each compared to the model containing the complete set of 19 odorants.
bNumber of the assessors detecting an odor difference in triangle tests, maximum 5.
The instrumental and sensory methods presented in the French fries example have also been suc- cessfully applied in the elucidation of other aro-
mas. The results are presented in the book for some individual foods.
5.3 Individual Aroma Compounds
The results of dilution analyses and of aroma simulation experiments show that only 5% of the more than 7000 volatile compounds identified in foods contribute to aromas. The main reason for the low number of odorants in the volatile fraction is the marked specificity of the sense of smell (for examples, cf. 5.6).
Important odorants grouped according to their formation by nonenzymatic or enzymatic reactions and listed according to classes of compounds are presented in the following sections. Some aroma compounds formed by both enzymatic and nonenzymatic reactions are covered in sections 5.3.1 and 5.3.2. It should be noted that the reaction pathways for each aroma compound are differentially established.
Frequently, they are dealt with by using hypo- thetical reaction pathways which lead from the precursor to the odorant. The reaction steps and the intermediates of the pathway are postulated
Table 5.16. Some Strecker degradation aldehydesa
Amino acid Strecker-aldehyde Odor
precursor threshold value
Name Structure Aroma (µg/l; water)
description
Gly Formaldehyde CH2O Mouse-urine,
ester-like
50×103
Ala Ethanal Sharp,
penetrating, fruity
10
Val Methylpropanal Malty 1
Leu 3-Methylbutanal Malty 0.2
Ile 2-Methylbutanal Malty 4
Phe 2-Phenylethanal Flowery,
honey-like 4
aMethional will be described in 5.3.1.4.
by using the general knowledge of organic chemistry or biochemistry. For an increasing number of odorants, the proposed formation pathway can be based on the results of model experiments. Postulated intermediates have also been confirmed by identification in a numbers of cases. However, studies on the formation of odorants are especially difficult since they involve, in most cases, elucidation of the side pathways occurring in chemical or biochemical reactions, which quantitatively are often not much more than negligible.
5.3.1 Nonenzymatic Reactions
The question of which odorants are formed in which amounts when food is heated depends on the usual parameters of a chemical reaction.
These are the chemical structure and concentra- tion of the precursors, temperature, time and envi- ronment, e. g., pH value, entry of oxygen and the water content. Whether the amounts formed are really sufficient for the volatiles to assert them- selves in the aroma depend on their odor thresh- olds and on interactions with other odorants.
Aroma changes at room temperature caused by nonenzymatic reactions are observed only after prolonged storage of food. Lipid peroxidation (cf. 3.7.2.1), the Maillard reaction and the related Strecker degradation of amino acids (cf. 4.2.4.4.7) all play a part. These processes are greatly accelerated during heat treatment of food.
The diversity of aroma is enriched at the higher temperatures used during roasting or frying. The food surface dries out and pyrolysis of carbo- hydrates, proteins, lipids and other constituents, e. g., phenolic acids, takes place generating odor- ants, among other compounds.
The large number of volatile compounds formed by the degradation of only one or two constituents is characteristic of nonenzymatic reactions. For example, 41 sulfur-containing compounds, including 20 thiazoles, 11 thiophenes, 2 dithi- olanes and 1 dimethyltrithiolane, are obtained by heating cysteine and xylose in tributyrin at 200◦C. Nevertheless, it should not be overlooked that even under these drastic conditions, most of the volatile compounds are only formed in concentrations which are far less than the often relatively high odor thresholds (cf. 5.6). For this reason, only a small fraction of the many volatile compounds formed in heated foods is aroma active.
