THE MOST IMPORTANT QUALITY CRITERIA OF SOME HOME-
GROWN BLACK-CURRANT VARIETIES

IV. Gas-chromatographic analysis of aroma

Taina Kuusi, Alpo Siiriä and Terttu Kuusi

The State Institute for Technical Research, Laboratory for Food Research and
Technology, Otaniemi

Received May 30, 1966

In the study of the quality properties of black-currants, the main results which
have been reported earlier in parts I, II and 111 of this series, the last part reported
here is concerned with characterization of the aroma by gas-chromatographic
methods.

Nowadays, gas-chromatography should be the method of choice for the in-
vestigation of volatile components. Application of this method makes it possible
in principle to obtain a qualitative picture of the components which form the basis
of the aroma; moreover, it should also be possible to evaluate the quantities of the
different aroma components.

Here, it was necessary for reasons of methodology to divide the study in two
parts, the first concerned with the lower-boiling part of the aroma, and the second
the higher-boiling part. The former components were studied by means of a column
which tolerated water, but not high temperatures; the higher-boiling compounds
had to be studied with another column which tolerated higher temperatures, but
not water, which was accordingly removed. As a result separate reports are given
on these two groups of aroma compounds.

The material for this investigation was the same as in the rest of the series,
although the analysis had to be limited so that lower-boiling compounds were
analysed from the material obtained in 1962 and 1963,and higher-boiling compounds
only from the material of 1962. Consequently the series obtained here are not so
complete as in the other analyses reported. During the course of the investigation,
it had been possible to improve some points in the methods, but as it was considered
more important that the various series should be mutually comparable, changes

https://www.c-info.fi/en/info/?token=zRsuYLggEepIPaWU.Yx1HdqrWYkuJQFx2PhKw-w.WAbvpa_k-jHkm2xucHnR7eeD8XZDe7-uXfx65BUquwSZ1eGOsK_w23xEM2Fbh7-QdyUQbg_ZBSNklVNFR5hoBrr-17kioCzfm7ZsghNkbTxsHxgeqAEZCU-R9DN5ALni52EBWF3LztKmEj6NWHt6t8niL4dfyQdv7ZQ_U7f0uP6YNZc1X6reOPscaDJsxoKzpeFBu7JpgH_WFt5V


163

in the methods were not adopted once a practicable procedure had been developed.
In the comparison presented here the material was frozen. In addition, studies
were also made with fresh berries; this led to some interesting results, but as in these
studies the standing times of the samples to be analysed necessarily varied consider-
ably in the different cases, the results obtained with the frozen berries seem to be
more reliable.

Lower-boiling aroma compounds

Methods. The aim here was that of applying a method for the isolation
of aroma compounds such that the secondary changes were as small as possible.
Heating and the effect of air were consequently avoided, and the adsorbent employed
in the concentration of aroma was the same mass as that later utilized in the gas-
chromatographic column as filling material. It would probably be most advisable
to use the so-called »head space» technique to avoid changes; here, this was not
possible, as the amount of aroma compounds was not sufficient for the application
of this method in the conditions available. As a result, concentration of the aroma
before gas chromatography was inevitable; this was effected by means of gas stream
and the cooling of filled adsorption tubes.

The sample consisted of 35 g of frozen berries, first thawed, after the addition
of an equal amount of distilled water, in a water bath at 20°C. The thawed mixture
was homogenized in a small closed plastic vessel, with a Bamix homogenizer, for
I—3 min. The aroma was collected from this sample by leading a slow nitrogen
stream through it at room temperature. The stream was released into the sample
through fritted glass to make the bubbles small. The adsorption tubes were cooled
with running water, 10°C (column A), and salt-ice mixture at 18°C (column U);
the diameter of the tubes was 4 mm, and the heights of the mass approximately
12 and 15 cm, respectively. The collecting time was 5 hours; the gas volume was
then 2000 ml, and the flow rate 7 ml/min. If variations occurred, the gas volume
was kept as constant as possible. The apparatus is illustrated in Fig. 1.

In the gas chromatography, a Perkin Elmer Fractometer 116 was used. For
column filling, Mass W (polyethylene glycol/teflon) was chosen. To introduce the
sample into the fractometer, the adsorption tube was heated with a water mantle

Fig. 1. Apparatus for collection of low-boiling
aroma compounds from varieties of black-currant.
A: gas inlet. B; sample container. C: connecting
bulb. D: Column A, cooled with running water. E:
Column U, cooled with NaCl-ice mixture, F: Glass
wool. G; Flow-rate meter. X; Fritted glass.



to 100°C, and a helium gas stream led through for 2 minutes. The retention times
were calculated from the starting point of the introduction. The following conditions
were applied: temperature 80°C, and length of the column 2 m in both years; in
1962, Pire2 0.8, Po 2 0.8, Pir 2 0.5 and flow rate 4.4; in 1963, Ph<- 2 1.0, Po 2 0.8, Pir2
0.45 and flow rate 3.8. A slight difference exists in the column packing between the
years; in 1962 the packing was done by the authors, whereas in 1963 a column
packed by the producer was used. In the latter case, the packing was evidently
tighter, which brought about some slight differences in the retention times. Not-
withstanding this, the curves are well comparable, although exact superposition
is not possible. Two different detectors were used simultaneously, viz. a thermistor
and a flame ionization detector (FID). In general, the former has given only the
macrocomponents (water, ethanol and methanol); the latter is more sensitive, and
has thus given more components; it does not register water. With the macrocompo-
nents as basis, it is possible to make an approximate evaluation of the boiling points
of the aroma compounds present. As a rule, the time of analysis was at least 60
minutes; at this point, the water peak had already passed, and examination was
concerned with whether something else would still come out.

Comparison of the results is founded upon the curves obtained from U-tubes.
In 1962, the aroma in the A-tubes was also analysed to determine the extent
to which the two tubes differed. It was concluded that the pictures are very
similar, although the lowest-boiling compounds are better found in the U-tubes,
whereas the bulk of the water is in the A-tubes. As it was impracticable
to prevent the introduction of water into the fractometer, and the presence of
water is a disturbing factor in analysis, it was considered advisable to employ the
U-tubes for aroma analysis. Although the flame ionization detector does not reg-
ister water, the more of it is present, the more the registration of other compounds
in the same area is disturbed.

The quantity of berries taken here for aroma analysis was constant, and the
fraction collected in the U-tubes has been taken for the run as a whole. Thus quan-
titative comparisons between the different samples are possible in these series,
although standardization of the method was probably not quite sufficient for
exact quantitative determinations.

Results. As a general conclusion it has been found that in the different
analyses a total of 13peaks was present, although they didnot occur everywhere. The
first 10 of these were comparatively sharp, but the last 3, coming after the water

Fig. 2. Aromagram of Brödtorp black-currants
from Piikkiö, low-boiling compounds. (80° C).

Fig. 3. Aromagram of Black of Lcpaa black-
currants from Piikkiö, low-boiling compounds.

(80° C).

164



165

peak, were flattened. It may be assumed that these last peaks represent consider-
able amounts of aroma compounds, although these are not well separated under the
conditions applied.

The results obtained are exemplified by the aromagrams of Brödtorp (Fig. 2)
and Black of Lepaa (Fig. 3) from Piikkiö, 1962, arrived at with the FID; in addition,
the water peak recorded with the thermistor detector is indicated as a striped peak.
At the base, the numbers of the peaks are given, the sensitivity used being given
at the top of peaks; e.g. 8 x means that the real height is obtained by multiplication
by this factor. The dotted lines represent areas in which the sensitivity was changed.

In this study it was not possible exactly to identify the compounds, which
are designated only by numbers in the order of issue from the fractometer. In some
cases it may be possible to guess which is the main component of a peak, although
it might be that one peak represents several components not separated under the
prevailing conditions. Thus, peak 5 probably corresponds to ethanol, and peak 4 to
methanol, the principal components of the low-boiling aroma. In part, the peaks
are well separated, and in part they are discernible only as small elevations on the
side of a higher peak. In some cases the retention times exhibit variations, e.g.
peak 1, ordinarily rather low, may vary somewhat in its location; accordingly it
has been designated as la and lb. No more than a few of the peaks are comparable
in height with the macrocomponents; thus at the level of the 1962 methanol peak
these are only Nos. 2 and 9; in 1963, such heights appear even more seldom, and
then these come into question Nos. 1, 6 and 11.

After these general observations, it is important to consider which factors
may influence the aroma composition, and in which way. Factors which may in-
fluence the aroma include the variety, weather conditions, northemness of the
growing locality, the degree of ripeness and the condition of the berries. The aroma
composition again may be characterized by different features: 1) presence or lack
of some peaks in the aromagrams, 2) relative amounts of the component parts,
evaluated from the height or area of the peaks, and 3) differences in the general
level of the aromagrams. Of these criteria, the first two are mainly qualitative,
and the third quantitative in nature.

As regards the effect of variety, it is concluded that there are to be found
no differences based on the variety factor, evident as the characteristic presence
or lack of any component; in the same variety, the presence of the peaks may vary
in different years and in different localities, and thus the aroma composition of the
varieties in the low-boiling range is not sufficiently constant to be characteristic
of the variety. Neither is the amount of aroma, as evaluated on the basis of the
sum of heights of all peaks, typical of the variety, but varies between wide limits
in the different years and in the various localities. It thus seems that the effect of
variety on the aroma composition cannot be demonstrated by virtue of the strong
influence of the other factors although the existence of such effect is well known;
cf. Larsson (12) concerning varieties in northern Sweden. However, it must be
remembered that the aroma composition considered here represents only the low-
boiling part of the aroma.



166

In contrast, the effect of the weather conditions seems to be very
clear-cut. In 1962, the summer was cool, whereas in 1963 it was warm and sunny.
This difference is discernible in the aroma picture, first in the fact that the frequency
of some peaks is changed. Thus, in 1963, peaks No. 6, 7 and 9 are less frequent
than in 1962, whereas peak No. 8 shows contrary behaviour. This is illustrated by
the following consideration, in which the total number of samples studied for each
case was 17:

Peak G 7 9 8

Frequency in 1962 14 11 17 4
Frequency in 1963 3 4 9 14

Secondly, a great difference between the years is discernible also in the total
aroma, calculated from the sum of the peak heights, as is shown below:

Average of the total aroma Limits of variation

1962 14098 1840-38229
1963 1886 354- 8093

This comparison evidences that the level of the total aroma was much higher
in 1962 than in 1963. This result appears unexpected in that it is well known, that
the berry aroma develops more strong and more typical in a warm and sunny sum-
mer than in a cool and rainy one. Nevertheless, it should be borne in mind that
the aroma compounds studied here are easily volatile, and consequently secondary
losses may occur more easily in a warm than in a cool summer.

As regards the effect of the groving locality, it might be anticipated
that it would correspond to that of the weather conditions, i.e. a southern location
would correspond to a warm summer, which would here mean a diminution in the
total aroma, and changes in the frequencies of some peaks. When the results are
examined with these guiding lines in mind, some similarity is found in that the
peaks No. 6, 7 and 8, which were more frequent in a cool summer, are commoner
towards the north than towards the south whereas peak No. 8 again exhibits con-
trary behaviour. However, this result is only of preliminary nature, and exceptions
exist. Also, the total aroma, measured as the sum of all peak heights, increases
towards the north, as is evident in the following;

Locality Piikkiö Pälkäne Laukaa Ylistaro Maaninka Rovaniemi
1962, average 7002 17822 21ÖÖ3 18684
1963, » 974 815 1429 5391 1718

The increase is irregular, and the years under study differ from each other;
the causes of such irregularities might be not only climatic conditions, but also
the effect of the degree of ripeness and the transport conditions, which may have
raised the values at Maaninka. Although these samples were not definitely fermented,
the longer transport had rendered them riper than samples for which transport
period was shorter.



167

The effect of the degree of maturity and condition seems to have
been evident, although precise evaluation of these factors is difficult. It is clear that
ripening occurs more slowly in a cool summer and in a more northern locality, which
in principle induce delay in the ripening. The clearly unripe berries were eliminated
from the berry samples studied. In such unripe berries the aroma is less developed,
as is wellknown; this was also confirmed here by some gas-chromatographic analyses
of unripe berries. However, in all the samples, slow after-ripening without doubt
occurred during frozen storage before the aroma was studied, which may have to
some degree levelled out the differences. When the notes concerning the condition
of the berries are compared with the aroma picture obtained, it is discernible that
often the number of aroma components is less in samples characterized as unripe;
this result is not without exceptions, possibly because the definitely unripe berries
were eliminated. In contrast, in samples characterized as crushed or over-ripe, the
number of components exceeds the average, in any case as regards 1962, when the
summer was cool. No. 3 was often absent from the peaks when the sample was
designated as unripe; it is possible that this component develops only in ripe berries.
Again, peak No. 10 might be characteristic of over-ripe berries, although the cor-
relation is not completely regular. The degree of ripeness exercises a clear effect
on the sum of peak heights, although sometimes after-ripening changes the situation.
Thus, in samples designated as unripe, the peak heights are lower than usual, and in
over-ripe samples are again higher than the average. The over-ripeness is particularly
clearly evident in peak No. 5, i.e. ethanol, of which the amount is increased ap-
preciably if fermentation takes place, but also increases as a result of over-ripeness
as such, although no definite fermentation has occurred.

An interesting question in consideration of the aroma picture of black-currant,
is the proportion of ethanol and methanol in total aroma, and possible variation
of this relation in the different cases. The proportion of ethanol is of particular
interest. Alternatives might exist here; 1) when the amount of ethanol increases,
the amount of other aroma components does not increase at the same rate, and
thus ethanol reaches a dominant role, or 2) when the amount of ethanol increases,
this simultaneously induces a rise also in the other aroma components, or perhaps
in their »loosening» from the berries, and thus the aroma content is increased in its
entirety. To decide between these alternatives, a comparison was made of the
ethanol content and the total aroma content, evaluated from the peak height in
the aromagrams. This comparison is illustrated in Fig. 4. It can be seen that a rel-
atively good positive correlation exists between the ethanol content and the amount
of total aroma. This result would lead to the conclusion that under the experimental
conditions applied the rise in the ethanol content in general enhanced the typical
black-currant aroma, the ethanol not having taken on any dominant postition.
In these samples, no definite fermentation had taken place; the result may of
course be divergent if this happens. Moreover, in the diagram the two years under
study clearly differ from each other: in 1962 both ethanol content and total aroma
were high, whereas in 1963 both values remained low.

Another interesting point in connection with the aroma of black-currant is
the possible correlation between the aroma number and the height of the ethanol



peak. A comparison of these two factors is shown in Fig. 5. As the aroma number
reflects the volatile reducing substances, the most prominent of these in black-
currant being ethanol, a positive correlation is to be expected. Such a correlation

Fig. 4. Relation between total aroma
and ethanol content, measured as peak
heights, in the black-currant aroma of

the different varieties.

Fig. 5. Relation between the ethanol
content, measured as peak height of
aromagrams, and aroma number of
different varieties of black-currant.

168



169

does seem to exist, although in this case, where the aroma numbers vary relatively
little, the values of the two years investigated differ somewhat, and give two sep-
arate lines; these are nearly parallel. The correlation found is positive, as was
expected: the higher the ethanol peak, the higher the aroma number. Nevertheless,
considerable variation exists in the values.

Again, as concerns the methanol, it might be anticipated that it would show
a correlation with the degree of esterification of the pectin, as a lowering of the
degree of esterification of pectin during its breakdown means that methanol is
liberated from the methoxyl groups. The amount of methanol would thus be higher
the more breakdown had taken place in the pectin, and thus the lower was the
degree of its esterification; ot course this assumes that there had been no loss of

methanol through evaporation. Fig. 6 illustrates the correlation between the height
of the methanol peak and the degree of esterification of pectin. Is it observable
that there is the expected negative correlation, although the variability is great,
and the two years investigated again differ from each other; however, the lines
obtained are nearly parallel. Since methanol is even more easily volatile than ethanol,
it is no matter for surprise that the correlation is not better. Here again it may be
noted that the height of the methanol peak was in general higher in 1962 than in
1963.

Higher-boiling aroma compounds

Methods. It might be assumed from those aroma compounds, which have
a boiling-point higher than that of water, that they are to a great part of terpene
nature. In the buds of black-currants leaves, some terpenes and sesquiterpenes,

Fig. ti. Relation, between the methanol
content, measured as peak height of
aromagrams, and degree of esterification
of pectin in different varieties of black-

currant.



170

such as nopinene, 1-sabinene, d-caryophyllene and d-cadinene were found 1937 by
Glichitch and Igolen (8); it seems evident that similar aroma compounds are
found in both black-currant berries and in the leaves, although other compounds
are probably also present in berries. Such terpene type and other higher-boiling
compounds might be separated from the lower-boiling compounds by fractional
distillation, generally employed in recovery of the aroma. However, it was noted
in preliminary experiments that when an attempt at aroma recovery was made
with steam distillation, marked changes took place in the aroma. The same conclusion
was drawn by Tinner (13). It was accordingly considered advisable to avoid a high
temperature, and the distillation was performed in vacuum. Further, it was borne
in mind that the waxy materials found in the berry skin should be removed already
at this phase; these waxes probably have little to do with the aroma, and in addition
their presence is harmful to the chromatographic column. However, the distillate
obtained cannot be studied as such by gas-chromatographic methods, as the columns
do not tolerate water. The aroma compounds were therefore extracted from the
distillate with an organic solvent. Such a procedure does not fully guarantee that
the aroma obtained corresponds in all respects to the original aroma. It was thus
necessary to check, by organoleptic methods, whether the aroma obtained had
changed in its properties.

The material for this part of the study consisted of frozen berries of the year
1962. For the analysis, 250 g of berries were taken, an equal amount of distilled
water added, and the berries were thawed on a water bath at 20°C. The mixture
was homogenized in 3 lots with Bamix in a closed plastic vessel; the homogenization
time was 1 minute. Isolation of the aroma from the homogenate was effected by
vacuum distillation, with the temperature of the water mantle 50°C. The receiving
vessel was chilled with ice-salt mixture at —lB°C. The vacuum varied between
50 and 80 mm Hg, and the boiling point was correspondingly 38—46°C. The desired
amount of distillate was 100 ml, but this varied in the different cases between 80
and 160 ml; the time required varied between 45 and 120 minutes. At the end of
the distillation, part of the distillate in the receiving vessel was usually in the
frozen state. The distillates were ordinarily kept at -f- 4°C in vessels closed with
fritted-glass cork until the following day. The extraction was performed in a liquid
extraction apparatus holding 100 ml of distillate. The extraction time was 5 hours,
and 100 ml of ethyl ether (p.a.) was used as extractant. The extract was dried with
anhydrous Na2S0 4; the bulk of the ether was evaporated in a Rotavap evaporator,
following which approximately 6 ml of concentrated extract remained. Immediately
before gas-chromatographic analysis, the rest of the ether was removed, so that only
one drop of the aroma sample remained. An endeavour was made to use ethyl
cloride in the extraction (cf. 6); this solvent has the advantage of a boiling point as
low as +l2°C, and thus the aroma losses which occur in the evaporation of the
solvent are minimal. However, in comparison of the two gas-chromatographic anal-
yses it was discovered that ethyl chloride extracts evidenced no more components
than the corresponding ether extracts. Ether was consequently chosen as the
extractant, as it is simpler in use. In gas-chromatography, a Perkin Elmer
Fraktometer 116 was used with a 2 m column, filled with Celite (60—100 mesh),



171

and as the stationary phase 15 per cent butandiol succinate polyester (BBS). The
following conditions were chosen: Phc 0.8, Po 2 0.8, Ph 2 0.5, flow rate 3,1, the tem-
perature in the the main series being 120°C. Before gas-chromatography, the sample
was concentrated to as small a volume as possible; it was drawn in an injection
needle, which then contained partly liquid and partly gaseous phase, with a total
volume of 25 fil. On the average, the time of the run was 2 hrs. Apart from the
main series, a checking run was effected at 150°C, to speed up the separation of the
highest-boiling components. For this series, a new sample, 5/d in size, was taken
for the gas chromatography, by adding some drops of ether, from the sample tubes
kept in the refrigerator. The conditions were as follows: Pne 0.8, Po 2 0.8, Ph 2 0.5,
flow rate 2.6 and temperature 150°C. The column was the same as above.

As in these experiments the amount of distillate taken for the ether extraction
was in all cases 100 ml irrespective of the total volume of the distillate, and since in
addition the transfer of the aroma concentrate in the gas chromatograph was
not quantitative, the results obtained can be compared with each other only
qualitatively; quantitative comparisons cannot be performed. If such comparisons
had been required, the amount of starting material should have been greater,
and aliquots taken for further treatment at the various phases; for this, the
amount of berries available was inadequate. If, again, the ethereal extract were
to have been used quantitatively for gas chromatographic analysis, a precise
injection technique would have become necessary; no reliable method had been
developed for this at the time of the study.

Results. When the curves obtained at 120°C are compared, 18 clear-cut
and well separated peaks are discernible; in the last part, the peaks become flattened,
and their comparability is poor. In all, approximately 25 peaks are present. Some
of these are much higher than others; prominent peaks of this type are Nos. 10,
15, 16, 18, and 19. In the earliest part of the curves some very high peaks are no-
ticeable, but their separation is not good. It was found from some model runs that
easily volatile substances such as ethanol, propanol and butanol would in these
conditions come out before peak No. 3; this area shows a rather high level. There
exist slight differences between the different samples in respect to the occurrence
and relative heights of the peaks; the reasons for such differences are considered
below.

As regards the curves obtained at 150°C, the purpose was that of improving
the picture in the last section; this, of course, led to poorer separation of the peaks
at the start than was the case at 120°C. The identification and comparison of the
peaks at 120 and 150°C is rendered somewhat difficult by the fact that some of the
peaks may be missing at 150°C. However, the general course of the curves is so
similar, that it may be possible to conclude from the highest peaks which of them
correspond to each other in the two instances. Thus, at both temperatures, the high-
est peaks were Nos. 10, 15, and 19. The peaks were thus the same at both tempera-
tures, except as regards the last high peak at 150°C, which would be missing at
120°C. In part, some other peaks may also be missing at 150°C, although they occur
at 120°C; this is not very astonishing, as the samples were left standing for a rather
long period before being analysed at 150°C. Peaks Nos. 16, 17 and 18 in particular



Fig.
7.

Comparison
of

aromagramsof

Brödtorp
samples

obtained
from
5

different
growing

places
in

1962.
Higher-boiling
compounds
run
at

120°
C.

Piikkiö Ylistaro Maaninka
I

_.

.

.

Maaninka
II

Rovaniemi

ITL'



Fig
"

8

"
Com

P"ison
of

aromagramsof
5

A»,

„

20

7~~

Wellington
XXX

Wellington
X

Janslunda Goliath Brödtorp

I7J



174

are poorly presented at 150°C. In contrast, the peaks following No. 19 are more
clearly seen at 150°C than at 120°C.

To illustrate the results arrived at, comparison of the 5 samples of Brödtorp
is presented at both 120 and at 150°C (Fig. 7, Fig. 9). For the comparison, the curves
of these samples, obtained from different localities, have been drawn on a semilogarith-
mic paper, so that even the highest peaks are clearly visible, and the various curves
are easily mutually comparable. In the same way, a comparison of 5 different varie-
ties obtained from Piikkiö is presented at the two temperatures applied (Fig. 8,
Fig. 10).

On examination of these illustrations, itmust be borne in mind that the retention
times are not always exactly the same, by reason of small variations in temperature
regulation and other conditions during the run. This is seen in the aromagrams of
the different varieties at 120°C, where peak No. 10 has still almost the same location
for the varieties, although the peaks of Janslunda and Brödtorp are slightly more
advanced than the others; this difference becomes more conspicuous towards the
end. Nevertheless, the curves are well comparable, as in all cases the general features
are relatively similar, and the differences between the varieties are easily detectable.

The results obtained here rendered it possible, in the same way as with the
low-boiling compounds, to consider the effect exerted on the aroma picture by the
variety, locality of growing, degree of ripeness and condition. In contrast, the effect
of weather conditions is more difficult to assess, as the material is from the year

Fig. 9. Comparison of aromagrams of Hrödtorp samples obtained from 5 different growing places
in 1962. Higher-boiling compounds run at 150° C.

Piikkiö
Ylistaro
Maaninka I
Maaninka II
Rovaniemi



175

1962 only; the analyses of the 1963 material are lacking. Neither is quantitative
comparison of the aroma practicable, as the amounts analysed do not correspond
exactly to each other. Consequently, the qualitative aroma picture is employed
for comparative purposes.

As concerns the effect of variety on the aroma picture, conclusions may
be drawn from the comparison shown in Figs. 7,8, 9, and 10. When a study is made
of the aroma of the same variety at different localities as an example of this,
Brödtorp has been chosen, since the material was greatest here it is noticeable
that the aroma picture is very much alike in the different cases. For instance, peaks
Nos. 14 and 24 are missing in all of them. Small differences discernible may be
owing to the effect of locality, degree of ripeness and condition. By contrast, a
comparison between various varieties from the same locality shows more differences.
In particular, the relative amounts of components in the area of peaks Nos. 14—16
differ, although the peaks in the last part, best visible at 150°C, also evidence dis-
similarities.

Conclusions on the effect of locality are rather difficult to draw, as it is
not easy to be certain which is in fact the effect of locality, and which that of degree
of ripeness and condition. However, it might be assumed that in more northern
localities, more numerous components would be present; an indication in this di-
rection is given by the comparison between the variety Brödtorp at Piikkiö and
Rovaniemi, of which the former is southern, the latter northern, and where the

Fig. 10. Comparison of aromagrams of 5 different varieties from Piikkiö in 19H2. Higher-boiling com

pounds run at lf>o° C. The curve of Brödtorp has been drawn at a lower level to facilitate comparison

Wellington XXX
Wellington X

_._._._._._._. Janslunda
Goliath
Brödtorp



176

berries were to some extent unripe in both cases. A similar general trend is observ-
able in the comparison of the samples from Piikkiö with corresponding samples
from Maaninka in the north, although here the possible effect of the prolonged
transport in the latter case cannot be ignored.

Very few conclusions may be drawn in respect of the effect of degree of
ripeness and condition. If an additional sample of Wellington XXX,
originating in Great Britain and having by virtue of the long transport period
definitely been fermented, is employed for comparative purposes, it is seen that in
this case peaks Nos. 3, 9 and 10 were particularly high. However no similar features
were noted in the other samples studied where long transport might have caused
corresponding changes; thus in the Brödtorp samples, peaks Nos. 9 and 10 were
higher than usual in the sample of Rovaniemi, which was more or less unripe. It
thus appears that in the higher-boiling compounds the effect of degree of ripeness
and the condition do not exercise such a marked effect upon the aroma picture
as they evidenced in the lower-boiling compounds.

Organoleptic checking experiments. When the aroma is
isolated from the starting material at room temperature by means of a nitrogen
stream, it is probable that the aroma so obtained corresponds to the aroma com-
position, as it influences the sense of smell. This would accordingly also be the case
here in respect of the lower-boiling compounds. By contrast, when isolation of the
aroma is effected by vacuum distillation and subsequent ether extraction, it is less
clear that the aroma obtained corresponds to that originally present in the berries.
Some organoleptic tests were accordingly made to check the extent to which isolation
of the aroma had been successful in retaining the original properties.

In smell-testing, it was found that there was still present a strong and rather
typical smell ofblack-currant in the residue of the vacuum distillation; in the opinion
of the judges it would have been difficult to make a mistake in this respect. Evi-
dently, separation of the aroma by vacuum distillation had not been quantitative.
However, it may be mentioned that isolation of the lower boiling compounds by
the aid of a nitrogen stream had not been superior in this respect. In the distillate,
the smell was not so characteristic as that of the residue: some of the judges referred
to the smell as only »ester-like», others mentioned the smell of some other berries,
and no more than a small proportion of the judges evaluated the smell as typical
of black-currant. Thus the aroma of the distillate seems to correspond only to
one part of the original aroma both qualitatively and quantitatively. The aroma
of a distillate sample, which had been left standing, was evaluated as weakened.
The residue of the ether or ethyl chloride extract was found even more divergent;
here the descriptions of the judges hinted to the presence of terpenes: »bitter»,
»sticking smell», »orange peel», »terpene-like», »something pine-like». Again, the
water phase of the distillate after ether extraction was characterized as »berry-like»,
although one of the judges even recognized the black-currant aroma in it.

It thus seems that the method applied in isolation of the higher-boiling com-
pounds separated into the organic solvent mainly compounds of the terpene type,
whereas at least a proportion of those substances which form the basis of the general
berry aroma remained in the aqueous phases. These conclusions are. of course, of
a preliminary nature, as identification of the compounds has not been effected.



177

Discussion

Although a vast quantity of information is available on gas-chromatographic
aroma studies in general, and also a considerable amount of work has been done on
fruit and berry aroma, very few studies have been concerned with the aroma of
black-currant. In fact, the only significant contribution in this field is constituted
by the investigations made in Sweden by von Sydow and his group (cf. 7; 1; 2; 3,
4, and 5), which may be useful for comparison. However, in this case as well the
methods applied differ so greatly that comparison seems to be rather difficult. In
the Swedish studies, the main aim was that of making precise identification of the
aroma components, and measurement of their relative amounts, and a long list of
various compounds in the higher-boiling range, mostly terpenes, has been reported.
Nevertheless, not so much attention was attached to such factors as variety, locality,
weather conditions and so on, as in the present study. Notwithstanding the con-
siderable differences in methods, in both cases similar results have been found to
the extent that definite varietal differences in the higher-boiling compounds have
been noted (7; 3 and 5).

Summary

A gas chromatographic study has been made of the volatile aroma components
of black-currant. The material consisted of 12 home-grown varieties from 6 localities
taken over a period of 2 years, although the series are not complete. Separate in-
vestigation was made of the low-boiling and high-boiling parts of the aroma; in the
former case, material of two years was studied, and in the latter case, that of one
year only.

The lower-boiling components were isolated in chilled adsorption tubes from a
nitrogen stream led through a berry homogenate at room temperature. Analysis
was effected at 80°C with a polyetylene glycol/teflon column (the same mass served
as filling for the adsorption tubes). In all, 13 components were registered, although
their occurrence, and the amounts in the different samples, were somewhat variable.
As regards the different factors, the variety exercised no more than a slight effect
on the aroma picture, whereas weather conditions were important, influencing the
frequency of the peaks and the general aroma level; a cool summer increased the
level. The northernness of the growing locality had a similar effect, although this
result its of preliminary nature. The degree of ripeness and the condition had a con-
siderable effect; the level of aromagrams was higher with more ripe samples, and in
particular the ethanol content, which forms the bulk of the low-boiling aroma, was
increased. A positive correlation was detected between the height of the ethanol
peak and the total aroma or aroma number. By contrast, the height of the methanol
peak exhibited a slight negative correlation with the degree of esterification of the
pectin.

The higher-boiling compounds were isolated by vacuum distillation and sub-
sequent extraction with ether. In the gas chromatography, use was made of a bu-
tandiol succinate polyester/celite column, and the temperatures were 120 and 150°C.



178

In all, 25 components were observed. Of the various factors, the variety showed a
clear-cut influence on the aroma picture, as differences in the occurrence of the peaks
and the relative heights were noted between the varieties; in this series, quantitative
comparisons between the samples were impossible. The effect of locality and degree
of ripeness is less well defined.

Organoleptic checking showed that the changes were small in isolation of the
lower-boiling components. In contrast, isolation of the higher-boiling compounds
induced more changes, so that an aroma fraction obtained in the organic solvent
was more terpene-like than the original aroma.

Acknowledgment. Our thanks are due to the Department of Horticulture and
to the experimental stations which provided the material, as well as to the Finnish
Techno-Scientific Research Council, for financial support.

REFERENCES:

(1) Andersson, J., Bosvik, R. & von Sydow, E. 1963. The composition of the essential oil of black
currant leaves (Ribes nigrum L.) J. Sei. F'ood Agr. 14: 834 840.

(2) Andersson, J. & von Sydow, E. 1964. The aroma of black currants. I. Higher boiling compounds.
Acta Chem. Scand. 18: 1105 1114.

(3) — »— & —*— 1966. Gaskromatografisk karakterisering av aromen hos svarta vinbär av
olika sort och mognadsgrad. Lecture at Nordisk symposium om aromaforskning, Fredens-
borg, Denmark, 27 29. 4. 1966.

(4) — »— & —1966. The aroma of black currants. 11. Lower boiling compounds. Acta Chem.
Scand. 20: 522 528.

(5) —*— & —»— 1966. The aroma of black currants, ill. Chemical characterization of different
varieties and stages of ripeness by gas chromatography. Ibid. 20: 529 535.

(6) Bidmead, D. S. & Welti, D. 1960. Isolation and identification of volatile fruit-flavours. Research
in Appi. Ind. 13: 295-299.

(7) Bosvik, R. & von Sydow, E. 1962. En gaskromatografisk jämförelse av aromen hos nägra bär-
safter och bärsorter. Svenska Institutet för Konserveringsforskning, Göteborg. SIK-
Rapport Nr. 111: 1 13.

(8) Gx-Ichitch, L. S. & Igolen, M. G. 1937. Contribution ä l’etude de l’huile essentielle de Bourgeons
de Cassis. Les Parfyms de France 15:241—244.

(9) Kuusi, T. 1965. The most important quality criteria of some home-grown black-currant varieties.
I. Ascorbic acid. J. Sei. Agr. Soc. Finland 37: 254 281.

(10) —» 1965. The most important quality criteria of some home-grown black-currant varieties.
11. Dry matter, pectin, acid content, colour and formol value. Ibid. 37: 282 304.

(11) »— 1966. The most important quality criteria of some home-grown black-currant varieties.
111. Aroma number, organoleptic evaluation, and degree of ripeness. Ibid. 38: 150—161.

(12) Larsson, G. 1959. Norrländska sortförsök mcd svarta vinbär 1944—1958. Stat. trädg. förs. Mcdd.
122: 1-30.

(13) Tinner, H. 1962. Ueber mögliche Veränderungen der Fruchtaromen bei höheren Temperaturen.
Symposium FTuchtarome Bern. Berichte der wissenschaftlich-technischen Kommission
der Internat. Fruchtsaftunion, pp. 377 387.



179

SELOSTUS:

MUSTAHERUKAN TÄRKEIMMISTÄ LAATUOMINAISUUKSISTA ERÄILLÄ KOTIMAASSA

VILJELLYILLÄ LAJIKKEILLA

IV. Aromin kaasukromatografinen analyysi

Taina Kuusi, Alpo Siiriä ja Terttu Kuusi

Valtion teknillinen tutkimuslaitos, Elinlarvikeleollisuuslaboralorin, Otaniemi

Mustaherukan haihtuvia aromiaineita on tutkittu kaasukromatografian avulla. Tutkimusaineis-
tona oli 12 kotimaassa viljeltyä lajiketta, joista saatiin näytteitä 6 eri paikkakunnalta 2 vuoden aikana,
vaikkakaan sarjat eivät ole aukottomat. Alhaalla kiehuvat ja korkealla kiehuvat aromiaineet tutkittiin
erikseen; edellisistä saatiin kahden vuoden sarja, jälkimmäisistä yhden vuoden.

Alhaalla kiehuvat aromiaineet eristettiin jäähdytettyihin adsorptioputkiin johtamalla huoneen-
lämmössä typpivista marjahomogenaatin läpi. Kaasukromatografia-ajossa käytettiin polyetylenglyko-
li/teflon pylvästä; ajolämpötila oli 80°C (adsorptioputkien täytteenä käytettiin samaa massaa). Kaik-
kiaan 13 komponenttia saatiin esille, vaikkakin niiden esiintyminen ja määrät eri näytteissä vaihtelivat
jonkin verran. Eri tekijöistä lajikkeella oli vain vähäinen vaikutus alhaalla kiehuvien aromiaineiden
suhteen, kun taas sääolosuhteilla oli voimakas vaikutus sekä huippujen esiintymisfrekvenssiin että
yleiseen aromimäärään; viileänä kesänä aromimäärä oli suurempi. Kasvatuspaikan pohjoisuudella oli
samansuuntainen vaikutus, mutta tätä tulosta on pidettävä alustavanluontoisena. Kypsyysasteella ja
kunnolla oli melkoinen vaikutus; aromagrammien taso oli korkeampi kypsemmillä näytteillä, erityisesti
etanoli, joka muodosti pääkomponentin alhaalla kiehuvassa fraktiossa, oli määrältään suurentunut.
Positiivinen korrelaatio todettiin etanolihuipun korkeuden ja kokonaisaromimäärän sekä aromiluvun
välillä. Päinvastoin taas metanolihuipun korkeus osoitti lievää negatiivista korrelaatiota pektiinin
esteröitymisasteen kanssa.

Korkealla kiehuvat aromiaineet eristettiin vakuumitislauksella ja senjälkeen suoritetulla eetteri-
uutolla. Kaasukromatografiassa käytettiin butandiolsukkinaattipolyesteri/celiitti pylvästä, ajolämpö-
tilat olivat 120 ja 150°C. Kaikkiaan todettiin 25 aromikomponenttia. Lajikkeella oli selvä vaikutus
aromikuvaan, mikä näkyi huippujen esiintymisessä ja niiden suhteellisissa korkeuksissa; tässä koesar-
jassa ei ole mahdollista suorittaa varsinaisia kvantitatiivisia vertailuja eri näytteiden aromin kesken.
Kasvatuspaikkakunnan ja kypsyysasteen vaikutus on vähäisempi.

Organoleptiset tarkistuskokeet osoittivat, että huoneenlämmössä suoritetussa aromieristyksessä
aromin muutokset jäivät vähäisiksi. Sensijaan korkealla kiehuvien aromiaineiden eristys aiheutti enem-
män muutoksia siten, että orgaaniseen liuottimeen saatu aromifraktio oli enemmän terpeenityyppinen
kuin alkuperäinen aromi.

4