Volatiles in oat meal

Document Sample
Volatiles in oat meal Powered By Docstoc
					J. Sci. Food Agric. 1999 (submitted)

                   Contribution of Volatiles to the Flavour of Oatmeal
                       Meixue Zhou1, Kevin Robards1, Malcolm Glennie-Holmes2, and Stuart Helliwell1
    School of Science and Technology, Charles Sturt University, PO Box 588, Wagga Wagga, NSW 2678, Australia
    NSW Agriculture, Agricultural Research Institute, Wagga Wagga, NSW 2650, Australia

Sensory quality of 12 oatmeals as assessed by a trained panel was related to both variety and growing conditions with
variety being the major controlling factor. The volatile profile of the oatmeals was determined by solid phase
microextraction using headspace and gas chromatography. Volatile compounds assessed in this manner were correlated
with the sensory data and chromatographic peak areas accounted for 43 to 94% variations of sensory attributes.
Examination of groats demonstrated that most of the volatiles in oatmeals were induced during heat processing.

Key words: oat; volatiles; SPME, aroma.
                                                                  Mortlock and Yarran were also grown in Cowra and
INTRODUCTION                                                      Parkes in NSW in 1996.
Oats are a significant world crop and growing recognition of
the nutritional value of the oat grain has led to an expanded     Preparation of Rolled Oats
usage for food purposes. Oats are consumed as a rolled,           Samples of each variety (10 kg) were dehulled and then
flaked breakfast cereal throughout the world. Flavour has         stored at 8C for the two days which had to elapse before
been shown to be an important parameter for the                   processing could proceed. Samples (3.5 kg) were processed
acceptance of oat products to consumers (Heydanek and             following normal commercial practice using small scale
McGorrin 1986). Flavour perception depends upon a                 steaming, kilning and rolling machines in Uncle Tobys
multifaceted series of sensory responses as described by          Research & Development Centre (Wahgunyah, Victoria,
Labows and Cagan (1993). However, the most important              Australia). The conditions were: steaming for 9 min, kilning
contribution to flavour arises from the volatile constituents     at 100C for 45 min and 65C for 15 min, cooling to room
of a food.. In oats, Hrdlicka and Janicek (1964a, b)              temperature, resteaming for 5 min and rolling to a thickness
described the presence of carbonyls, bases and an alcohol in      of 55 m.
toasted oat flakes and regarded carbonyls as one of the
factors relating to the nutlike flavour. The basic fraction of    Sensory testing
the volatiles had strong nutty, earthy, toasted, slightly harsh
                                                                  Sensory assessment of rolled oats was performed by
and burnt flavours, whereas the neutral fraction was
                                                                  the trained taste panel at Uncle Tobys R&D Centre,
described as burnt grain, browned, heavy and not oaty
                                                                  Wahgunyah, Victoria. For sensory assessment, all
(Heydanek and McGorrin 1986).
                                                                  samples of rolled oats were prepared as per
                                                                  microwave directions for Traditional Oats with 2/3
Oats lack flavour as collected from the field and the
                                                                  cup oats and 1 1/3 cup cold water (filtered), then
aroma/flavour that is associated with oat products requires
                                                                  microwaved on Medium/High setting for 2.5 min
the intervention of a heat process to develop (Heydanek and
                                                                  with a lid and 2 min without a lid. Mean sensory
McGorrin 1981a and b; Fors and Schlich 1989).
                                                                  scores were calculated for each sample according to
Nevertheless, steam treatment alone is not effective for
                                                                  attribute determined by 8 panellists (minimum two
flavour development in oats which requires the direct heat
                                                                  years experience) in a round table discussion.
in the kiln without which oat products retain a flat, green
                                                                  Panellists were trained in twelve (one hour) sessions
taste, i.e., raw and slightly bitter (Gansmann and Vorwerck
                                                                  covering basic tastes and taste threshold levels, fruit
1995). Heydanek and McGorrin (1986) found different
                                                                  flavours, aromas, savoury flavours, herbs and spices,
processing procedures resulted in different amounts and/or
                                                                  essences and texture followed by a further twelve
types of volatile compounds.
                                                                  (one hour) sessions of Quantitative Descriptive
                                                                  Analysis (QDA) practice with on-going panel
Relatively few studies have reported varietal effects on oat
                                                                  involvement evaluation. Samples were not ranked for
flavour (Molteberg et al 1996). Our previous study (Zhou et
                                                                  acceptability (hedonic scale) but for definable
al 1999) reported that a new Australian oat variety, Yarran,
                                                                  flavour attributes. Attributes were scored on a linear
was unacceptable to consumers because of its poor flavour.
                                                                  scale of 0 - 100.
As part of a series of studies on oat quality for human food
use, this paper will report the effect of variety and growing
conditions on the flavour of individual oat varieties and the Extraction and measurement
correlations between volatile compounds in oatmeal and the Oatmeal (20 g) was mixed with water (100 mL) in a glass
results of sensory analysis.                                  flask, which was then gently heated in an electric heating
                                                              mantle until boiling. Collection of volatiles was commenced
EXPERIMENTAL                                                  immediately and performed by headspace solid phase
                                                              microextraction (SPME) using a polydimethylsiloxane non-
                                                              bonded fibre (100 m film) (Supelco) for routine analyses.
Groat samples                                                 Collection time was controlled at 20 min and samples were
Eight varieties of Australian oats (Bimbil, Carralup, maintained at a temperature which kept the porridge bubbling
Cooba, Echidna, Euro, Mortlock, Pallinup & Yarran) slowly during the collection stage. A polyacrylate fibre (85
were grown in Condobolin in New South Wales in 1996;          m film) (Supelco) was also examined.

                                                                (ROF). Varieties were separated into two groups in respect
Volatiles absorbed on the fibre of the microextractor were      to Starch Flavour (SF) and Aftertaste Starch Flavour (ASF).
desorbed in the inlet chamber of a HP 5890 Series II Gas        Bimbil, Carrolup, Echidna, Euro and Yarran lacked both
Chromatograph equipped with a flame-ionisation detector         SF and ASF while other samples presented significantly
and open tubular column (SE30, 30 m x 0.25 mm, film             stronger SF and ASF. The variations of Aftertaste Metallic
thickness 0.25 m) for 1.5 min. The column temperature          Flavour (AMF) among varieties were similar to that of MF
program was: 50C held for 1 min, a ramp from 50 to 300C       with Echidna the strongest and Cooba the weakest.
at 10C per min and then held at 300C for 4 min. Injector
temperature was set at 220C and detector temperature was       Mortlock and Yarran were also grown in Parkes and Cowra
maintained at 250C. Splitless injection was employed.          and, as shown in Table 1, growing site had no significant
                                                                effect on OAA, OA, SA, OAF, OF and CF of these two
For identification of these compounds, volatiles absorbed by    varieties whereas MF, AMF, BF, SF and ASF were affected
the fibre of the microextractor were desorbed in the inlet      significantly by growing site. Yarran was particularly
chamber of a VG Quattro GC/MS equipped with a DB-Wax            sensitive to growing conditions with the sample from
                                                                Condobolin having significantly stronger MF and AMF and
(60 m x 0.32 mm) column. The ion source was 200C, 70 eV
ionisation energy. The column temperature programme was:        weaker BF, SF and ASF than samples from Cowra and
                                                                Parkes. Compared to Yarran, Mortlock was less affected by
35C held for 5 min, a ramp from 35 to 220C at 3C per min
                                                                growing conditions with only the SPF of Mortlock from
and then held at 220C for 15 min. Injector and detector        Parkes being significantly stronger than Mortlock from
temperatures were 220C and 250C, respectively.                other growing sites.

Statistical analysis                                            Analysis of volatiles
All analyses were performed at least in duplicate and results   The analysis of thermally generated aromas by gas
recorded as the mean and standard deviation. The data were      chromatography or gas chromatography-mass spectrometry
analysed by single factor or multifactor analysis of variance   invariably requires a preliminary isolation of the volatile or
following Snedecor (1967). A principal components analysis      aroma compounds from the food. Many techniques have
using SAS software program PROC PRINCOMP and an                 been reported for this step including solvent extraction,
oblique factor analysis using SAS software PROC VARCLUS         steam distillation, simultaneous steam distillation-solvent
were performed to cluster the major peaks into groups of        extraction, liquid-liquid extraction (dialysis) and both static
related peaks based on the correlation matrix derived from      and dynamic headspace methods. The concentration of
the means. Both principal components analysis and oblique       volatiles in oatmeal was very low and preliminary results
factor analysis were carried out by Dr. M. Zack in National     showed that both solvent (pentane) extraction and solvent
Centre for Chronic Disease Prevention and Health                collection following water distillation were not suitable for
Promotion, Centers for Disease Control and Prevention,          detecting flavour compounds. With solvent extraction, the
Atlanta, Georgia, USA.                                          large amount of co-extracted lipid complicated subsequent
                                                                analysis and/or mandated further sample clean-up.
RESULTS and DISCUSSION                                          Similarly, distillation/solvent extraction was also found to be
                                                                unsuitable. In this case, co-extracted materials were
                                                                eliminated but the level of volatile components in the oat
Sensory analysis of oatmeals                                    samples (20 g) was below the detection limit using flame
Oatmeals were assessed by panellists on a non-hedonic           ionisation detection. In both cases, peaks in the
scale in which individual characters such as metallic,          chromatograms were attributed to solvent and the effects of
creamy or bitter flavour were assessed on a 0 - 100 scale.      pre-concentration of large volumes of solvent. Problems
Results of these tests are shown in Table 1. Three aroma        related to interference could not be eliminated even with
characteristics and eleven flavour characteristics were         high quality solvents. In contrast, solid phase
measured and most of these characteristics differed             microextraction (SPME) provided a rapid, solvent-free and
significantly between varieties with the exception of Overall   economical approach to recovery of volatiles from small
Aroma (OAA), Overall Flavour (OAF), Soapy Flavour               samples of oats.
(SPF) and Aftertaste Oaty Flavour (ATF).
                                                                Several variants of the technique were investigated,
Of the eight varieties grown at Condobolin, Carrolup,           including immersion sampling from the slurry of oatmeal
Cooba and Bimbil had the strongest Oat Aroma (OA),              and water, and collection of the headspace above the
whereas Mortlock and Yarran had the weakest OA. Starch          oatmeal. With immersion sampling of heated oats, the fibre
Aroma (SA) of Yarran was the strongest, followed by             became coated with starchy materials which could not be
Mortlock, Pallinup and Cooba. No SA was found in Bimbil,        removed by washing of the fibre prior to injection into the
Carrolup, Echidna or Euro.                                      GC. Headspace provided a viable alternative but even here
                                                                volatile compounds were below the limits of detection in all
The difference in Metallic Flavour (MF) between varieties       cases using unheated oats. The large quantities of volatiles
was also significant with Echidna having the highest value      found in uncooked oat groats by Heydanek and McGorrin
and Cooba the lowest. Oat flavour (OF) was regarded by          (1981a) can be attributed to the large sample size (8 kg) and
panellists as one of the most important contributors to         the complicated extraction procedure incorporating a pre-
acceptability. Cooba and Mortlock had significantly             concentration step on Tenax GC. In contrast, the present
stronger OF than other varieties, while Yarran and Bimbil       study used small samples (20 g) with a relatively simple
had the lowest value of OF. Most samples presented low          extraction. Nevertheless, the production of an oat product
Bitter Flavour (BF) while Echidna had relatively higher         suitable for cooking requires steaming, kilning and rolling
values for BF. Creamy Flavour (CF) is another important         and, more importantly, oatmeal is consumed after cooking.
contributor to acceptability for human consumption with a       Each step from raw groat to consumer product is expected
high CF being desired by consumers. Bimbil and Echidna          to develop flavour components (Heydanek and McGorrin
had significantly lower CF than all other varieties. Bimbil     1986) and it is reasonable to examine the volatiles from a
was the only variety that had very strong Raw-Oat Flavour       cooked product particularly where these are to be related to

results of sensory analysis.                                   measured by panellists and the predicted value from
                                                               independent peak areas as shown in Table 4 which were
Two types of fibre were examined for their suitability in      selected by regression analysis. Since the peak areas were
extracting volatiles from the headspace of cooked meals. It    standardised, a larger value in Table 4 means a greater
was found that the non-polar polydimethylsiloxane coating      contribution of the peak to the flavour measurement. For
was more suitable which agrees with previous observations      example, the value of OA was given by a multiple linear
(Coleman 1997) in which a non-polar polydimethylsiloxane       regression equation of the following form: OA = 37.36 +
coated fibre was found to be very effective in the             5.49 Cluster C (P2) - 3.41 Cluster B (P8) - 2.89 Cluster E
qualitative analysis of volatile and semivolatile Mailard      (P14) + 1.96 Cluster D (P20) + 7.87 Cluster F (P21) - 7.72
reaction products. Many volatiles were detected in cooked      Cluster A (P27) where P2, P8 ... P27 are the standard peak
oat samples (see Materials and Methods) and, as predicted,     areas. From the magnitude of the multipliers (7.87 and
the number and intensity increased in going from a groat       7.72), Cluster A and F had the greatest contribution to OA,
sample to oatmeal (Figure 1). Some varietal differences        followed by Clusters C, B, E and D.
were also observed as illustrated in Figure 2 and discussed
below.                                                         All six peak clusters showed significant effect on aroma.
                                                               Cluster B (P6), C (P7) and F (P13) presented significant but
Relationships between volatiles and flavour attributes         minor contributions to OAA. Cluster F (P21) and A (P27)
There were more than 50 volatile compounds detected in were found to be most important contributors to OA and
cooked oat porridge (Figure 2). The 46 most abundant and SA. Cluster C (P2), D (P20) and F (P21) positively
easily separated peaks were correlated (as raw peak areas contributed to OA but negatively to SA. In contrast, Cluster
without any correction for detector response) with the A (P27), B (P8) and E (P14) was negatively contributed to
flavour data as evaluated by taste panellists. Table 2 shows OA but positively to SA.
the ANOVA (F values) of all the selected peaks. F values
larger than 2.8 or 4.4 indicate a significant (P < 0.05) or Both aroma and flavour of the porridge were also attributed
very significant (P < 0.01) difference between varieties. to all the six clusters with cluster A (mostly P27) and cluster
Varieties differed significantly in all peak areas except F (mostly P21) having the greatest contributions. Cluster A
peaks 5 and 22 perhaps due to lower signal to noise ratios. contributed positively to OF, SF, SPF, and ASF and
There were strong co-correlations between the peaks. For negatively to ROF, AMF and AOF while cluster F showed
example, peak 1 (P1) was significantly correlated with P4, the opposite effect on these measurements. Cluster B (P1,
P5, P6, P8, P10, P11 and P16; P2 was significantly P5 or P8) showed a positive effect on CF, SF and ASF and a
correlated with P3, P7, P9 and P31 (Table 3). Thus, negative effect on MF, OF, AMF and AOF while cluster C
regression analysis was inappropriate to identify the peaks (P2, P7 or P15) had the opposite contribution to these
significantly affecting flavour because of multicolinearity flavours. Cluster D effected MF, SPF positively and OF, SF,
between the peaks themselves. From the correlation matrix AMF and ASF negatively. Similar to cluster D, cluster E
(selected data shown in Table 3) based on the means, contributed positively to MF and negatively to OF, SF and
principal components and oblique factor analyses were AOF.
performed to cluster the 46 peaks into six groups of related
peaks that might then be used as independent variables in Identification of volatile compounds
developing a multiple regression correlation between peak The identity of peaks is not essential to an oat breeding
area and flavour. This procedure yielded six peak clusters as program since most flavour attributes were predictable from
follows: Cluster A: P23 - 32, P34 - 38, P41 and P46; Cluster peak areas alone. However, identification of the volatiles is,
B: P1, P4 - 6, P8, P10 - 12 and P16; Cluster C: P2, P3, P7, nevertheless, desirable and essential to studying changes
P9 and P15 where P15 is negatively associated with other during processing and storage. Kovàts Indices of the 46
peaks in the cluster; Cluster D: P19, P20 and P22 where peaks seen in chromatograms of oatmeal (Figure 2) on SE30
P22 is negatively associated with other peaks in the cluster; are given in Table 5.
Cluster E: P14, P33 and P42 - 45; and Cluster F: P13, P17,
P18, P21, P39 and P40.                                         In groats, relatively few peaks were detected (Figure 1) with
                                                               the major compounds being identified from retention data as
This indicates that the original 46 peaks ca be represented palmitic acid and eicosane. Smaller peaks included hexanal
by the 6 clusters as independent variables in the multiple and nonanal and these assignments were supported by mass
regression. One of the peaks from each of these clusters, spectral data. After processing of the groats into oatmeal,
which had a relatively high correlation or high variation more volatile compounds were produced (Figure 1) and
between varieties, was then selected as an independent from the retention times, the major compounds included
variable. In order to get comparable coefficients (the larger hydrocarbons (alkanes) (decane, hexadecane, heptadecane,
the coefficients, the more important the variables), octadecane, nonadecane and eicosane), alcohols (1-
regression analysis was performed after standardising all the heptanol, 3-octanol, 2-octanol and 1-octanol), acids
peak areas to a mean of zero and a variance of 1.00 by (heptanoic acid, octanoic acid, nonanoic acid, dodecanoic
subtracting the mean of the twelve means from each acid and palmitic acid) and aldehydes (hexanal, nonanal and
individual mean and dividing by the standard deviation of decanal) (Table 5) and these assignments were also
these twelve means.                                            supported by mass spectral data. The most pronounced
                                                               flavour effect is expected to occur during the processing of
Regression analysis showed that except for OAF and BF, oat products due to heat treatment. Most of the aldehydes
other flavour measurements such as OA, SA and SF could are products of reactions logically expected to take place in
be predicted by a suitable linear combination of the area of the boiling water heat treatment (Heydanek and McGorrin
the representative peak from each group with from 43% in 1986). The hydrocarbons are derived from oxidative
the case of CF to 94% in the case of OA (adjusted R- decomposition of lipids which are believed to make no
squared statistic, percent variation determined by significant significant flavour contribution to foods (Shahidi et al
peaks, Table 4) of the variations being explained by the 1997). Similarly, some alcohols and aldehydes could also be
independent variables (uncorrelated peak areas). Figure 3 derived from the oxidative decomposition of unsaturated
illustrates the correlations between the actual value fatty acids such as oleic, linoleic and linolenic acids

(Heydanek and McGorrin 1986).                                     mass spectroscopy identification of volatiles from rancid
                                                                  oat groats. J Agric Food Chem 29 1093-1095.
                                                               Heydanek M G, McGorrin R J 1986 Oat flavor chemistry:
CONCLUSIONS                                                       principles and prospects. In: Oats: Chemistry and
Sensory testing requires well-trained panellists and is           Technology, ed Webster F H. American Association of
expensive, time consuming and difficult. Limited samples          Cereal Chemists, St. Paul, MINN.
can be tested which makes sensory testing of progeny from      Hrdlicka J, Janicek G 1964a Carbonyl compounds in toasted
oat breeders an inappropriate method for selecting a good         oat flakes. Nature 201 1223.
quality variety. This study showed that 43 to 94% of the       Hrdlicka J, Janicek G 1964b Volatile amines as components
variations in most flavour parameters could be predicted          of toasted oat flakes. Nature 204 1201.
from less than six peak areas from six independent peak        Labows J N, Cagan R H 1993 Complexity of flavor
groups in the headspace chromatograms. This makes it              recognition and transduction. In: Food Flavor and
possible for a breeder to identify the flavour profiles of a      Safety. eds Spanier A M, Okai H, Tamura M. American
large number of lines in earlier generations since the            Chemical Society, Washington, DC.
measurement of the volatiles can be easily performed on        Molteberg E L, Solheim R, Dimberg L H, Frølich W 1996
relatively small samples.                                         Variation in oat groats due to variety, storage and heat
                                                                  treatment. II. Sensory Quality. J Cereal Sci 24 273-282.
                                                               Shahidi F, Aishima T, Abou-Gharbia H A, Youssef M,
                                                                  Shehata A A Y 1997 Effect of processing on flavor
                                                                  precursor amino acids and volatiles of sesame paste
Acknowledgements                                                  (Tehina). J Am Oil Chem Soc 74 667-678.
The authors gratefully acknowledge the assistance of Uncle     Zhou M X, Robards K, Glennie-Holmes M, Helliwell S 1999
Tobys Ltd, and the support of a grant from the Grains             Effects of oat lipids on groat flour pasting properties. J
Research and Development Corporation of Australia.                Sci Food Agric (In press).

REFERENCES                                                     Captions for Figures
Coleman W M 1997 A study of the behavior of polar and
    nonpolar solid-phase microextraction fibers for the        Figure 1 Chromatograms comparing the volatile compounds
    extraction of Maillard reaction products. J Chromatogr     in the headspace of Yarran groats (a) and Yarran oatmeal
    Sci 35 245-258.                                            (b). Vertical scale is the same in (a) and (b)
Fors S M, Schlich P 1989 Flavour composition of oil obtained
    from crude and roasted oats. ACS Symposium Series 409      Figure 2 Chromatograms comparing the volatile profile of
    121-131.                                                   cooked oatmeals of cv. Mortlock (a), Yarran (b) and Euro
Gansmann W, Vorwerck K 1995 Oat milling, processing            (c). The major peaks are identified by number as peak 1 to
           and storage. In: The Oat Crop, ed Welch R W.        46.
           Chapman and Hall, London, pp 369-408.
Heydanek M G, McGorrin R J 1981 Gas chromatography-            Figure 3 Correlations of flavour measurement between the
    mass spectroscopy investigations on the flavor chemistry   value scored by panellists and the value predicted from
    of oat groats. J Agric Food Chem 29 950-954.               chromatographic peak areas of volatiles.
Heydanek M G, McGorrin R J 1981 Gas chromatography-

Table 1 Sensory scores of oatmeal samples as determined by a trained panel

                                Aroma                      Flavour
                      OAA       OA         SA      OAF     MF            OF       BF          CF       ROF SF                  SPF     AMF      A
    Bimbil            46.7       41.8a     0.0d    53.7     20.8cd       34.6bc   8.125abc 5.6d        41a     0.0d            0.0b    14.8b    25
    Carrolup          47.1       43.2a     0.0d    51.8     26.7bc       37.0abc  11.1abc     25.5abc 0.0b     0.0d            0.0b    19.1ab   20
    Cooba             48.5       41.2a     15.7c 54.2       16.5d        40.1ab   3.3c        26.5ab 6.2b      20.5c           0.8b    11.5b    22
    Echidna           44.0       38.8ab 0.0d       55.0     36.2a        37.0abc  17.7a       3.1d     2.8b    0.0d            0.0b    24.0a    19
    Euro              43.8       40.6ab 0.1d       52.0     18.6cd       37.4abc  6.1bc       17.3bc 1.1b      0.0d            0.0b    13.9b    21
    Mortlock          44.8       34.3bc 22.5b 52.8          19.7cd       36.8abc  10.1abc     22.9abc 7.6b     25.3bc          5.5b    14.1b    25
    Pallinup          48.8       38.7ab 22.5b 52.7          22.8bcd      35.8abc  3.7c        19.8abc 6.6b     26.5b           0.0b    19.5ab   21
    Yarran            48.6       34.2bc 33.0a 54.0          29.3ab       35.0abc  4.7bc       16.9c    3.1b    2.6d            0.0b    23.5a    18
    Mortlock          46.5       34.3bc 22.6b 54.1          18.6cd       39.5ab   11.7abc     28.6a    5.6b    24.2bc          0.1b    17.1ab   24
    Yarran            47.5       35.0bc 28.8a 56.8          21.0cd       32.4c    14.3ab      19.0bc 6.3b      29.5b           0.0b    16.0ab   23
    Mortlock          45.0       34.2bc 22.2b 55.1          20.0cd       41.4a    5.5bc       23.9abc 6.5b     20.1c           24.8a   14.0b    25
    Yarran            47.0       31.3c     32.6a 52.5       19.6cd       35.5abc  14.3ab      22.0abc 5.1b     36.1a           0.0b    14.9b    20
P-value               0.196      0.000     0.000 0.307 0.000             0.039    0.003       0.000    0.000 0.000             0.000   0.008    0.
*Values followed by the same letter are not significantly different from each other, SSR test at the 5% probability
OAA: Overall aroma; OA: Oat aroma; SA: Starch aroma; OAF: Overall flavour; MF: Metallic flavour; OF: Oat flavour;
BF: Bitter flavour; CF: Creamy flavour; ROF: Raw oat flavour; SF: Starch flavour; SPF: Soapy flavour; AMF: Aftertaste
metallic flavour; AOF Aftertaste oat flavour; ASF: Aftertaste starch flavour.

Table 2 ANOVA of the peak area (F value)

Peak        P1       P2        P3       P4      P5      P6      P7     P8      P9      P10    P11     P12     P13     P14     P15    P16    P17
F value     11.23    12.47     17.58    6.17    1.51    10.8    6.66   7.22    4.65    5.32   5.40    2.80    3.04    6.46    3.23   4.69   4.90

Peak        P24      P25       P26      P27     P28     P29     P30    P31     P32     P33    P34     P35     P36     P37     P38   P39     P40
F value     9.96     17.36     47.09    31.96   18.48   29.66   4.06   44.30   53.22   3.12   15.97   25.63   27.83   24.86   14.81 13.67   3.49

*F0.05(11, 11) = 2.82; F0.01(11, 11) = 4.46)

Table 3 An example of correlations between the various peaks, shown here between the first two peak
areas and other peak areas

Peak         1            2            3         4          5       6         7          8         9          10         11      12          13      14      15
1            1.00         -0.24        -0.35     0.97       0.80    0.86      -0.11      0.93      0.20       0.93       0.79    0.49        0.28    -0.10   0.26
2            -0.24        1.00         0.97      -0.19      0.15    -0.06     0.80       -0.46     0.79       -0.18      -0.18   -0.48       0.31    0.18    -0.41

Peak         24           25           26        27         28      29        30         31        32         33         34      35          36      37      38
1            0.17         -0.23        -0.06     -0.23      -0.11   -0.10     0.08       -0.25     -0.19      -0.18      0.01    -0.12       -0.32   -0.25   -0.15
2            0.42         0.47         0.11      0.49       0.43    0.52      0.54       0.70      0.53       0.47       0.43    0.50        0.21    0.67    0.60

*r (0.05) = 0.58; r (0.01) = 0.71.

Table 4 Peaks significantly contributing flavour components (the larger the number the more important the contribution of
the peak)

Peak      OAA       OA         SA          OAF       MF     OF      BF      CF        ROF     SF        SPF       AMF    AOF     ASF
I         46.5  37.4           16.7                  22.5  36.9             19.3      7.6   15.4        2.6   16.9       22.4    11.1
A               -7.77          42.67                       4.15                       -9.69 21.79       5.82 -7.24       -10.5   17.63
                (27)           (P27)                       (P41)                      (P27) (P27)       (P30) (P27)      (P27)   (P27)
B         1.00 -3.41           14.91                 -2.10 -1.95            3.34            8.05              -3.49      -1.14   7.37
          (P6) (P8)            (P8)                  (P5) (P8)              (P1)            (P1)              (P5)       (P1)    (P8)
C         1.29 5.49            -22.7                       2.50                       6.61 -16.4              4.05       4.71    -11.2
          (P7) (P2)            (P2)                        (P15)                      (P2) (P2)               (P7)       (P2)    (P2)
D               1.96           -9.37                 1.64 -0.95                             -11.0       8.08 -1.88               -8.89
                (P20)          (P20)                 (P20) (P19)                            (P20)       (P12) (P22)              (P20)
E               -2.89          12.14                 4.08 -4.42                             -4.22                        -3.21
                (P14)          (P14)                 (P14) (P45)                            (P33)                        (P14)
F         -2.44 7.87           -44.1                                        -5.18     7.72 -17.9        -7.93 5.62       9.79 -14.6
          (P13) (P21)          (P21)                                        (P39)     (P39) (P21)       (P17) (P21)      (P21) (P21)
R2**      0.67 0.94            0.74                  0.78   0.85            0.43      0.64 0.78         0.68 0.92        0.85 0.80
* I = Intercept of the regression equation.
** R2 = adjusted R-squared statistic.
The predicted value of each sensory measurement could be calculated by a regression equation. For example, OAA = 46.5
+ 1.00 P6 + 1.29 P9 - 2.44 P13. A blank value in the table signifies that the related peak groups had no significant
contribution to the flavour measurement.

Table 5 Kovàts retention indices on SE30 of the 46 peaks detected by flame ionisation detection from the headspace of
cooked meals.

Peak Retention       KI         Possible Compounds                   Peak Retention      KI            Possible Compounds
No.   time                                                           No.   time
1     3.78           822        Hexanal *                            24    15.54         1458
2     5.18           877                                             25    15.76         1477
3     5.66           897                                             26    16.98         1583
4     5.78           902        2-Nonane ??                          27    17.07         1591          Dodecanoic acid ??
5     6.92           951        1-Heptanol ??                        28    17.18         1601          Hexadecane *
6     7.59           981        Octanal ??                           29    17.69         1645
7     7.69           985        3-Octanol ??                         30    17.88         1662
8     7.80           990        2-Octanol ??                         31    18.32         1700          Heptadecane *
9     8.06           1003       Decane *                             32    18.40         1707
10    8.87           1041                                            33    18.56         1721
11    9.11           1053       Octanol *                            34    18.85         1746
12    9.67           1081       Nonanal *                            35    19.18         1775
13    10.67          1132                                            36    19.43         1776
14    11.29          1165       Heptanoic acid *                     37    19.50         1802          Octadecane *
15    11.42          1172       Decanal *                            38    19.63         1814
16    12.37          1225                                            39    19.79         1827
17    12.89          1255       Octanoic acid *                      40    20.08         1844
18    13.26          1277                                            41    20.63         1891          Nonadecane *
19    13.90          1316                                            42    20.96         1917
20    14.16          1336       Nonanoic acid ??                     43    21.33         1953          Palmitic acid *
21    14.37          1355                                            44    22.47         2071
22    14.54          1370                                            45    22.74         2101          Eicosane
23    15.25          1432                                            46    22.84         2112

* Retention time match the standard; ?? deducted from the Kovàts indices.


Shared By: