2. SURVEY ON GAS CHROMATOGRAPHY–MASS
2.1. Gas Chromatography–Mass Spectrometry of Flavors
Several techniques for the removal of volatile compounds associated with flavor from foods have been reported: dynamic headspace (DHS) and purge and trap (PT) [7–9], solvent extraction [10], high-vacuum distillation and steam distilla- tion [11], simultaneous steam distillation–extraction (SDE) [8,11,12], supercriti- cal fluid extraction (SFE) [13–15], and solid-phase microextraction (SPME) [16,17].
The aroma compounds of dry-cured Parma ham have been analyzed by thermal-desorption GC–MS after extraction by means of the DHS technique [7].
Using GC–MS, 122 substances, including hydrocarbons, aldehydes, alcohols and esters, were identified in the volatile fraction. The same research group used DHS and SDE techniques to isolate aroma components from Parmesan cheese [8]. By means of GC–MS, more than 100 substances were characterized in the extracts.
This application is discussed later in this chapter in section 3.
Volatile compounds present in 14 commercial olive oils and cooking oils (corn, vegetable, and rapeseed oils) have been analyzed by GC–MS [9]; aroma sampling was carried out by PT. This method proved to be effective for the quality control of these oils during production and may give a chromatographic pattern as a useful fingerprint for determination of origin.
Using GC–MS, interesting data on the chemistry of Australian honey vola- tiles have been obtained by D’Arcy et al. [10]. The aroma compounds of Austra- lian blue gum (Eucalyptus leucoxylon) and yellow box (Eucalyptus melliodora) honey samples were isolated by solvent (ethyl acetate) extraction. This chemical fingerprint procedure was demonstrated to be promising for identifying com- pounds that are useful for authenticating the floral origin of honeys. In fact, a variety of distinctive nor-isoprenoids, monoterpenes, benzene derivatives, ali- phatic compounds, and Maillard reaction products were identified among the nat- ural honey volatiles, some of these being floral source descriptors for Australian honeys.
Three different well-established sampling techniques, i.e., high-vacuum distillation steam distillation and SDE, have been investigated for the isolation of volatile compounds derived from raw and roasted earth-almond [11]. Using GC–MS, it was possible to identify the indicator compounds for evaluation of the degree of earth-almond roasting. The main volatiles characterizing raw earth- almond were alcohols, whereas in the roasted earth-almond aroma, furans, pyra- zines and pyrrols prevailed, as formed by the Maillard reaction during the roasting process.
412 Careri and Mangia
In a study to obtain information on the effect ofγ-irradiation on aroma of fresh mushrooms, the SDE technique followed by GC and GC–MS was applied to qualitatively and quantitatively analyze volatile compounds of nonirradiated mushrooms and samples that had beenγ-irradiated with doses of 1, 2, and 5 kGy [12]. The important finding of this investigation was the large reduction of the content of total volatiles, primarily of eight-carbon compounds, due to the irradia- tion process, resulting in a flavor loss.
Supercritical fluid extraction as a sample preparation technique in food analysis has become increasingly popular in recent years. The usefulness of using supercritical fluids to investigate wine aroma has been demonstrated by both off- line SFE and coupled SFE–GC [13].
A fast and efficient method based on the use of SFE and GC–MS for the analysis of flavor compounds from roasted peanuts has been recently developed [14]. A group of substances known to be related to roasted flavor were identified and quantified using the full-scan mode. Total ion chromatograms of extracts of two peanut samples roasted under different conditions are shown in Figure 1.
The GC–MS analysis confirmed the presence of pyrazine compounds and other flavor substances, such as methylpyrrole, that are associated with the roasting conditions and sensory perceptions of a taste panel.
Supercritical fluid extraction combined with high-resolution GC–MS proved to be a powerful tool for the analysis of the virgin olive oil aroma [15].
The volatiles identified were compared with those obtained by using the DHS method. Different aromatic profiles were obtained by applying the two extraction procedures. The profiles obtained by DHS–GC–MS corresponded to a genuine extra-virgin olive oil sample in accordance with previous findings [18]. The pres- ence of off-flavors was not detected. In the SFE extracts, however, markers of oxidation processes were identified, since this technique is also suitable for the extraction of semivolatile compounds. These were volatile compounds related to oxidation of linoleic, linolenic, and oleic acids, and in particular aldehydes and acids, which had been previously found in oxidized olive oil samples [19].
Solid-phase microextraction has been investigated for the analysis of 2,4,6- trichloroanisole, a cork taint compound, in wine samples [16]. This solvent-free procedure was coupled to GC–MS under selective ion monitoring (SIM) condi- tions using a fully deuterated internal standard ([2H5]trichloroanisole) for quanti- tative purposes. The SPME–GC–MS method was demonstrated to be selective, precise, and sensitive with a 5 ng L⫺1 limit of quantification.
Solid-phase microextraction has been applied to the analysis of volatile substances in apples. Compounds were separated by time-compressed GC using time-of-flight MS for detection [17]. Time-compressed GC was proposed for re- duction of the time required for separation without loss in analytical performance.
In fact, unknown compounds, even when not well separated chromatographically, could be characterized by their mass spectra. Time-compressed GC combined
Analysis of Flavors and Fragrances 413
(a)
(b)
Figure 1 Peak profile for extract from roasted peanuts: (a) mild conditions: 145°C, 3 minutes, and (b) severe conditions: 170°C, 17 minutes. Peak notation: 1, methylpyrrole;
2, hexanol; 3, hexanal; 4, methylpyrazine; 5, 2,6-dimethylpyrazine; 6, furancarboxalde- hyde; 7, 2,3,5-trimethylpyrazine; 8, 2-ethyl-5-methylpyrazine and 2-ethyl-6-methylpyra- zine; 9, 3-ethyl-2,5-dimethylpyrazine. (From Ref. 14.)
with time-of-flight MS allows analysis of dozens of components in very few minutes owing to the extremely fast spectral generation rates (up to 500 spectra/
sec) [20]. The potential of this technique is illustrated in Figure 2. Using time- compressed GC, apple aroma compounds containing 1 to 15 carbons were eluted in about 2 minutes (Table 1). Performing traditional PT–GC analysis, the separa-
414 Careri and Mangia
Figure 2 Demonstration of high speed spectral generation (40 spectra/sec) enabling the detection and quantification of coeluting compounds by using GC–TOF-MS. The solid line represents the reconstructed total ion current (RTIC). Retention times differ by ap- proximately 0.2 sec. (From Ref. 17.)
tion of the collected volatiles would have required more than 40 minutes, i.e., 20 times longer. In addition, only 5 of the 29 components identified using SPME–
GC–time-of-flight-MS were detected in the PT–GC–flame ionization detection (FID) profile. From these findings, it can be concluded that the method developed proved to satisfy the requirements of analytical methods for fruit aroma research, since it is rapid, solvent-free, inexpensive, and amenable to automation. In addi- tion, SPME provided good linearity for compounds ranging in concentration from ppb to ppm without influence from matrix changes.
Regarding quantitation of aroma substances by GC–MS, a quantitative assay of ultratrace compounds is particularly demanding. Gas chromatography–
mass spectrometry using deuterium-labeled methoxypyrazines allowed identifi- cation and quantitation of naturally occurring grape methoxypyrazines (Scheme 1) in a variety of red wines, at the low-ng level [21]. These compounds are potent flavorants contributing to the aroma of natural products such as peas and bell peppers [22], and are responsible for the herbaceous/vegetative aroma of wine.
Stable isotope dilution GC–MS under chemical ionization (CI) conditions was used to analyze 2-methoxy-3-(2-methylpropyl)pyrazine in nine Australian and three New Zealand Cabernet Sauvignon wines (mean concentration 19.4 ng/L) and in six blended Bordeaux wines (mean concentration 9.8 ng/L). In all of the
Analysis of Flavors and Fragrances 415
Table 1 Volatile Compounds in Golden Delicious Apples, Sampled with SPME (PDMS, 100àm) and Identified by Time-of-Flight MS
Retention time (min)
Peak Volatile compound SPME Purge and trap
1 Pentane 1.082
2 Acetone 1.132
3 1-Butanol 1.322
4 Propyl acetate 1.494
5 Propyl propanoate 1.524
6 Butyl acetate 1.619 13.45
7 Ethyl 2-methylbutanoate 1.639 13.80
8 2-Methylbutyl acetate 1.753 15.80
9 Propyl butanoate 1.802
10 Butyl propyrate 1.932
11 Pentyl acetate 2.005
12 Butyl 2-methyl butanoate (?) 2.105
13 Butyl butanoate 2.164
14 Hexyl acetate 2.396 31.34
15 Butyl 2-methyl butanoate (?) 2.437
16 Pentyl butanoate 2.460
17 Hexyl propyrate 2.492
18 Propyl 2-methyl-2-butenoate 2.650
19 Hexyl 2-methylpropyrate 2.855
20 Not identified 2.893
21 Hexyl butanoate 3.183
22 Butyl hexanoate 3.202
23 4-Methoxyallylbenzene 3.223
24 Hexyl 2-methylbutanoate 3.264 32.76
25 2-Methylbutyl hexanoate 3.395
26 Hexyl pentanoate 3.633
27 2-Methylpropyl 2-methylbutanoate 3.754
28 Hexyl hexanoate 4.002 38.82
29 α-farnesene 4.453 41.75
Source: Modified from Ref. 17.
wines considered, the determined concentrations of this pyrazine were above the sensory detection threshold in water (about 2 ng/L), thus proving its contribution to wine flavor. As for its isomer, 2-methoxy-3-(1-methylpropyl)pyrazine, its level was less than 2 ng/L, the sensory detection threshold being about 1 ng/L in water, suggesting that this component will seldom significantly influence the wine aroma.
416 Careri and Mangia
Scheme 1