CA1072395A - Isolation of an aromatic fraction of coffee - Google Patents

Abstract

INVENTOR: PAUL CAZENAVE

TITLE: A PROCESS FOR ISOLATING AN AROMATIC
COFFEE FRACTION

ABSTRACT OF THE DISCLOSURE :

A process for isolating an aromatic coffee fraction from an aqueous medium containing said fraction, which compri-ses contacting the aqueous medium with an organic solvent which is an azeotropic mixture capable of boiling at a temperature below about 50°C of(a) at least one non-aro-matic hydrocarbon and(b) at least one non-aromatic halo-genated hydrocarbon or an ether and recovering a solvent phase containing the aromatic fraction.

The azeotropic mixture may be one of the following mixtures:

pentane/methylene chloride 68 - 32 %
pentane/isopropyl chloride 58 - 42 %
pentane freon 113 (trichloro-trifluoroethane) 94 - 6 %
pentane/diethylether 20 - 80 %
pentane/methylene chloride/diethylether 34.5 - 3.5 62 %

Application to flavouring or strengthening.

Description

: ~7~:395 This invention relates to a proces~ for isolating an aromatic coffee fraction, to the aromatic fraction obtained by this process and to it~ use.
After Shey have been extracted from their natural S medium, coffee aromas are fragile entities which are difficult to preserve for prolonged periods in the presence sf water.
: It is for this reason that~ in the majority of processes for isolating aromatic coffee fractions where these fractions are converted into aqueous solutions or come into contact with ` 10 high levels of moisture, the fractions are immediately ; extracted with a suitable liquid, such as an organic ~olvent, a~ oil, etc. However, lt has been found that, even if the liquid cont~ning said frac~ions is carefully dried, the fractions deteriorate rapidly, even at low temperatures and aLso irre~pective of the concentration of the~Q fractions in the liquid. Thus, a coffee aroma in the form of a 20%
solution in methylene chloride will not keep for more than a few days, even at 200C and in the absence of light. Attempt~
have been made to eliminate thi~ acute instabillty of volat~le coffee fract~on~ by combining the extracts obtained with solvents of different polarity or by using mixtures of solvents for extraction. The stability of the volatile fr~ctions would not ~ppe~r to be greatly lmproved~
By contrast, the present invention enables volat$1e coffee fractions of outsta~ding stability to be obt~ined.
: .

. . .
; ~2-.,,'' `
~

' '.' ;

:``
~`
: `

`:

~.~17Z3'95 The present invention relates to a process for isolating an aromatic fraction of coffee from an aqueous medium containing said fraction, which comprises contacting the aqueous medium with an organic solvent which is an azeotropic mixture capable of boillng at a temperature below about 50 C of(a) at least one non-aromatic hydrocarbon andlb) at least one non-aromatic halogenated hydxocarbon or an ether and recovering a solvent ; phase containing the aromatic fraction.
; The present invention also relates to a process such as defined above, in which the aromatic fraction collected in the azeotropic mixture is concentrated by elimination of this azeo-tropic mixture, more especially by distillation.
`~ Finally, the invention relates to the aromatic fraction thus obtained and to its use as an aromatising agent.
It is known that an a~eotropic mixture, or a3eotrope, is a mixture which, under a given pressure, behaves in the same way as a pure substance and, in particular, has a fixed boillng point PE. According to the invention, the aæeotropic mixture has ~; a boiling point below about 50 C. Thus, in cases where ~he pro cess is carried out under atmospheric pressure, the constituents of this azeotropic mixture are compounds of relatively low mole-cular weight. According to the invention, suitable non-aromatic . hydrocarbons are those containing from 4 to 7 carbon atoms; sui-`` ~ table non-aromatic halogenated hydrocarbons are methyl~ne chlo~-~ ride, ethyl chloride, isopro~yl chllo~ide, chloroorm, ~YY==r~-`~ tdichlorodifluoromethane), ~re~-~ (trichlorofluoromethane), freon 113 (trichlorotrifluoroethane); suitable ethers are dlethyl ether, ethyl isopropyl ether, diisopropyl ether, furan (a cYclic ether), providing these compounds are capable of formlng, with a ~' tro de rf~k . ' .

`:
~ ` .

. ~
.. ..
. ; .

':.
non-aromatic hydrocarbon, an azeotrope with a boiling point below about 500C and insoluble in water. In addition, the : azeotrope may be ~ more complex mixture than a simple binary ;; mixture~ Some suitable azeotropic mixtures are shown by way of example below:
Constituents Propor~ions PE(760 mm Hg) : bv volume_ pentane/methylene chloride 68-32 300C
pentane/isopro~yl~ hloride 58-42 30.5C
. ~ 10 pentane/Fre~-~ 3(trichloro-. trifluoroethane) 94-6 360c pentane/diethyl ether 20-80 3~oC
pentane/methylene chloride/
. diethyl ether 34.5-3.5-62 330C
:` 15 It is obvious that the process accordi~g to the invention may also be canied out with azeotropic mixtures which have a ~ boiling point below 500C under a pressure different from atmospheric pressure, the constituents of this aæeotropic mixture having to be selected accordingly. For ~xample, if `` 20 tke process is carried out under reduced pres~ure, the con~tituents of the azeotrope may be selected from compounds : of relatively higher molecul~r weight. Conversely, if the :: process ~ carried out under pressure, the constituents of . the azeotrope may be selected from compound~ of very low : 25 molecular weight, for example from compounds that are gaseou-q under ordinary conditions.
The ~tarting material is an aqueou~ medium : containing aromatic ingredients of coffee. The aqueous medium : may contain only aromatic ingredients o coffee, irrespective of ~he means by which the~e i~gredient3 h~ve been dissolved, although it may al~o contain other elements, .....
,e :

. . .

;.
.
.. . .

~72395 e3pecially solids, such as the solids found in coffee for example9 arom~ ~trengtheners, agents intended to correct any imbalances, e~c. For example, the aromatic coffee ingredients may emanate from the condensation of aromas S released during the roastlng of green coffee or during the grinding of roasted coffee, or even froo the condenQation of aromas obtained by the stripping with gases or with steam either of a roasted coffee, in dry fonm or in suqpension in water, or of an aqueou~ extract of roasted coffee.
As will be seen hereinafter, the aromatic fraction obtained by the process according to the invention is characterised by i~ remarkable stability. It has been found that this stability is not governed by the type of coffee tre~ted (arabica, robucta~ excelsa, santos, etc), or by its geographic origin (America, Africa or Asia), or by its degree of roas~ing or by the concentration of the ~tarting aqueous medium which may vary within wide limits.
One convenient measure of the concentration of a condensate ` obtained by stripping dry roasted cofee with steam is, for ~0 example, the stripping level, i.e. the ratio by weight of condensatewater to dry starting coffee. This level generally smount~ to between 1 and 300%.
The process according to the invention may becarried out in different w~ys. In a first embodiment, the aqueous medium i8 extracted in batches. In a second embodiment, the aqueous medium is continuously extracted. Broadly spe~king, the extraction techniques are well knownO Some of them ~re mentioned ~y way of illustration in the followingO For example, extraction m~y be carried out in a stirrer equipped tank. By virtue of the azeotropic characteri3tic~ of the ~; -5-. .

~07Z3~5 s~ent, extraction m~y also be carried out in an ~pparatus of the Soxhlet or similar type at the maximum temperature penmitted by the solvent. Several variants are pos~ible, - including for example recycling of the a~eotropic mixture, or -l 5 cascade extraction in bat~eries of t~nks.

; The aromatic ingredient~ may also be continuously extracted~
for example by countercurrent clrculation of the two liquids.
In cases where extraction is carried out in a tank9 ; 10 relatively stable emulsions can be formed, especially if the aqueous medium containing the aromatic coffee ingredients also contains solids. In this case, simple freezing of the emuls`ion9 followed by return to the starting temperature9 is generally sufficlent to break the emulsion.
In addition, it has been found that the guantity of azeotropic mixture to be u~ed, measured by the ratio of ; mixture t`o aqueous medium, is not a detenminin~ factor.Accordingly, it is desirable to use quantities as reduced as possible whilst, at the s~me time, ensuring correct extraction of the aromatic coffee ingredients.
These quantities (by volume) ~re for ex~mple of the order of 0.3 to 3 time~ the volume of the aqueous medium to be treated for each extraction operation.
So fsr as concentr~tion of the aromatlc fractions i9 ~5 concerned, it may be accomplished by any known me~hod which does not denature them, more especially by distillation of the solvent, the ~olvent being an azeotropic mixture of low boiling point. In addition,slnce distillation is carried out at a constant temperature and with a constant composition of ` 30 the solvent, the degree of concentration of the aromatic .

.

1~972395 :. ;
fraction is no longer a critical factor. It ha~ been found in practice that the aromatic fraction may be concentrated without damage to a residual solvent content of the order of 5~/0.
", 5 In one preferred embodiment of the process according to the invention, an aqueous medlum obtained by the l to 300 stripping of ground roasted Colombian arabica coffee or of a mixture of Colombi~n, Cameroons and Mexican arabica coffee or even of a mixture o Nicaraguan arabica and Angolan robusta coffee, i9 extracted with the azeotropic mixture. To this end, extraction is carried out in batches, in tanks and in three or four stages, using a total volume of azeotrop~e mixture equal to approximately 1.5 times the volume of the ` aqueous medium. Extraction is carried out under atmospheric pressure and at a temperature of approximately 200G. The aromatic fraction contained in the azeotropic mixture is then concentrated by distilling the azeotropic mixture at atmospheric pressure up to a reduction in the initial volu~e of the order of 40 to Sot~s.The concentrated aromatic fraction collected then contains approximately 10~ of residual azeotropic mixture.
It has already been mentioned that, apart from its representative character of the ~rom~ of coffee, the chief characteristic of the ~romatic fraetion obtained by the proces~ i~ its exception~l stability. For example, whereas a methylene chloride extract of ~ stripped coff~e aroma, concentrated to a residual solvent content of approximately 20%, only keeps for a few days at -200C and only for a few more days at -800C, an otherwise comparable extract obtained with the azeotropic mixture of pentane and methylene chloride " .

_7 ~;

' . `

keeps for at l~ast one year at a temperature of -20~C. This stability may be demonstrated with precision by chromatographic analyses in the gas phase.
The degree of concentration of the aromatic fraction wouLd not appezr to have any effect upon its stabillty. It is for this reason that most of the az~otropic mixture is generally eliminatad by distillation.
'! ~ The organoleptic quality of the aromatic fractions obt~ined by the process according to the invention is remarkable although it is influenced by the concentration of thc aqueous medium extracted with the azeotropic mixtur~.
Thus, in the case of a medium consi~ting of a condensate obtained by stripping ground roasted coffee with steam, an ; optimum quality has been observed for media with stripping levels of from 100 to 150%, i.e. in cases where the ground coffee haR been treated with 1 to 1.5 times its weight of ste2m.
The aromatic fraction thus prepared, concentrated or otherwise, may be used as an aromatising agent either for imparting a coffee aroma to a material from which such an aroma is absent or for strengthening an already existing coffee aroma.
In the first application, known as ab initio aromati~ation, the quanti~ies of aromatic fraction to be added to the material to be aromatised are governed by the nature of that m~terial.
If this materi~l is neu~ral from the aromatic polnt of view, the additions may amount to between 0.5 and 12/Oo(per thou~and) calculated on~weight of aromatic fraction(without residual solvent), based on the weight of the solids in the material to be treated. Thus, it is possible to aromatise, ~or example, -- .
~ -8-'.' `

. ~
... .

. .
, ~ .
,. , ~ . .

1~7Z395 ., ~-- milk, coffee powders poor in aroma, creams, ice creams, chewing gums, etc. If this material already has an inherent aroma, the additions have to be defined case by case. In general, it is necessary to use proportionally greater quantities to supplant the inherent aroma, although in some cases relatively smallar quantities may prove sufficient as a result of a synergistic effect.
In the second application, namaly strengthening an already existing coffee aroma, the quantities of aromatic ` 10 fraction to be added are appreciably smaller and nonmally amount to between 1/5th and 1/20th of the quantities required for ab initio aromatisation. For ex~mple, in order to strengthen or enhance the aroma of a medium-quality instant coffee, it is sufficient to add from 0.1 to 0.6ho of aromatic - 15 fraction.
In one variant of these two applications, the aromatic fraction is fixed to an "aroma support", for example a salt or a material based on polysaccharides, by any known process.
It is this fixed arcmatic fraction which is used for ' 20 aromatisation.
; Finally, in order to aromatise coffee solids or to `- strengthen their coffee aroma, it i~ possible for ex~mpleto add the aromatic fraction either directly to a decoction of coffee, to an instant coffee powder or to the reconstitutad 2S instant coffee, or indirectly, i.e. after having fixed the aromatic fraction to coffee oil.
It i9 preferred to use a concentrated aromatic fraction ~` with a very low residual solvent content, so that there is ` no need to provide specific means for eliminating the residual solvent in the end product. In this co~ection, `. .
.
_g_ ' ' .

.':
` ' . ,. . , . ~ , .
:~ . . . ..

~7Z3~5 it is pointed out that, even at the higher aromatisation levels such as defined earlier on, the aromatised material does not h~ve any solvent taste.
In one preferred application9 the aromatic fraction prepared by the process according to the invention is directly added, after ooncentr~tion to a residual solvent content of the order of 10/~, to coffee solids in a quantity of from 0.35 to 1/~o(as strengthener) and in a quantity of from 3.5 to 10~/~o in the case of low-aroma coffee solids.
` 10 The following Examples illustra~e the process acccrding to the inventio~ and the use of the aromatic fraction prepared by this process. In these Examples, percentage~
and ratio~ are expressed in tenms of weight.

.. . .
lS A series of aromatic fractions of Colombian ~rabica coffee is prepared by separately suspending in one litre of water 100 g s~mples of this coffee, ground and previou~ly roasted at approximately 165C tD a weight loss of th~ order of 18~/o~ and by subjeoting these suspensions to stripping for variable periods. The quantities of condensate collected are measured, being expressed as stripping lev~ls by relating these quantities to the quantities of starting coffee used.
Each of these condensates is th~n treated as follows:
extraction with 4 x 25 ml of an azeotropic mixture of pentane (68~/.) and methylene chloride (32Vh)~ combining the 4 vol~es of organic phase, drying with anhydrous sodium sulphate and concentration by distillation of the azeotrope at 30~C to a residual solvent oontent of the order of 10%, evaluated . .
by gas-phase chromatography. The preparation parameters and the characteristics of the concentrated aromatic fra~tions :: , '' .
..

~7'~395 ....
. .
; thus obtained are set out in the following Table, in which : the columns represent:
.; 1. stripping times in mm

2. weight of condensates collected in g - stripping level in %
: 3. volumes of concentrated aromatic fractions in ml ~` 4. minimal stability 8 50 1.5 6 months :; 11 100 1.9 1 year 14 150 2.3 1 year 17 200 2.5 6 months ., _ . . 250 2.B 6 months ., .

The corresponding non-concentrated aromatic frac~lons ~ have a similar stability and, after concentration, give ;- 10 concentrated fractions whlch have the same characteristics.
~` In addition, comparable results are obtained ~rom a , mlxture of 407. of Colombian arabica, 30% of Cameroon ~rabica and 30~/. o Mexican arsbic~ coffee, or from a mixture of 60% of Nicaraguan arabica and 40% of Angolan robu~ta coffee.
Finally, parallel results are obtained from the3e `.: coffees or co~fee mixtures with different degrees of roasting.
~ EXAMPLES 2 to 5 : The tests described in Example 1 are repea~ed using Colombian arabica coffee and other azeotropic mixtures.
', .

.~

:

~0~7~395 , . .

The results obt~ined are set out in the following Table, in which the figures of columns 2 ~nd 3 h~ve the same meaning as in Example 1.

. . Examples ~ 2 3 azeotropic m~xture , , ~
. 2 pentane 58%, 100 1.9 isopropyl chloride 42% 30.50C 150 2.3 .,,~ , _ .... _

3 pent~ne 947~, 360C 100 1.6 . ~ ~ ~c~mr~ 6% 150 2.0 "L~ F~ \ 113~
_ __ .: 4 pentane 20~, 3~oc 100 1.2 diethyl ether 80% 150 1.6 ,................. ,,, ~ , ,, , _,, ,, . . . , . .~
. 5 pentane 34.5/0, 100 2~0 : methylene chloride 3.5%, 330C 150 2.4 ether 627o ': ~ ~ _ ._, .
~, .
: In every casa, the shelf lives amount to more than 6 months.
.. E~AMPLE 6 ~ Aromatlc fractions of Colombian arabica coffee are ,. ~
'~ prepared rom decoctions obtained by separately treat-L~g 100 g s~mples of this coffee with 1 litre of boiling wa~er.
After separation of the grounds, these decpctions are th~
. treated in the same way as in Example l(stripping wlth ~team, extraction with the azeotropic mixture of pentane (6~) . and methylene chloride (32~ etc). The co~centrated ~ .
~ aromatic fractions thus prepared have volumes of 116 ml .": 15 (stripping level 100~3 and 1.9 ml (stripplng level 150~).
They have a minimum shelf life of 1 year.
'~ ~ r~e ~

~ . .
. ' ,: ' ~ -';' - ' . :. :

7~39S

.
The operations described in Example 6 are repeated using an azeotropic mlxture of pentane (20 %) and diethyl ether (80 ~) as solvent. The aromatic fractions prepared have volumes of 1.2 ml (stripping level 100 %) and 1~6 ml (stripping level 150 ~). They keep for a minimum period of 6 months.

= ~
60 kg of Colombian arabica coffee, ground and prevlously roasted at 165 C to a weight loss of the order of 18 %, are treated for 18 minutes with 300 kg of steam superheated to 185 C. Af~er elimina~ion of 100 kg of condensate, 196 kg of an extract with a concentration of dry matter of the order of 11 ~ are obtained. This extract is then stripped with steam for approximately 10 minutes until g kg of çondensate have been recovered, corresponding to a stripping level of 15 %. The ope-rations are then continued in the same way as described in Exam-ple 1, using as solvent the azeotropic mixture of pentane ~68 ~) and methylene chloride ~32 %).
The aromatic fraction (90 ml) obtained after concentration to a residual solvent content of approximately 10 g is extremely stable and keeps for at least 6 months without damage.
The results are similar when the aromatic fraction has been prepared using the azeotrope pentane (58 %) / isopropyl chloride (42 %).

., 60 kg of Colombian arabica coffee, ground and previously ~ roasted at 165 C to a weight loss of the order of 18 ~, are `~ humidified with steam superheated to 185 C, after which this `~ humidified coffee is stripped with steam to a stripping level ; 1 3 ` ' .' : ` , : ' ' ' ' : ': . , ` ~
: :
. .

` ~72395 .
. of approximately 4 %, the vapours being condensed by passage through two ~uccessive condensers cooled with tepid water (300C) for the first and with cold water (10C) for the second~ The ~Icold~ condensate with a S volume of 1.2 litres is then removed. The operatio~s are ` then continued in the same way as des~lbed in Example 1~
; using the azeotropic mixture of pentane (68%) and methylene chloride (32%) as solvent.
The aromatic fraction obtained after concentrat~on to a residual solve~t content of 1~/~ is extremely stable. It keeps for at least 6 months without changing.

~` A neutral base powder is prepared as follows:
a mixture of 3 arabica coffees (Colombian 40%, Cameroon 30%
and Mexican 3~/0), ground and previously roasted at 165C
to a weight loss of the order of 18%, is extracted ln tne same way as described in Example 8, and the ` extract thus obtained is subsequently subjected to stripping `~ with steam ln order to remove ~s aroma~ Finally, the de-aro~atised extract is spraydried to give the required ~- neutral base powdPr.
Three 15 g s~mples of this powder are then removed and have added to them 50 ~1 (i.e. in round figures 3.5%o), 100 ~1 (7~00) and 150 ~1 (10/.o), respectively, of the aromatic fr~ction with a 3tripping level of 100/~ obt~ined in accordance with Example 1 (pentane/methylene chloride). Three instant coffees are then prepared by dissolving each of these aromatised powders in 1 litre of hot water. The coffee~
are then submitted to a jury of eight trained tasters who are asked to indicate their pr~ference. The preferred ,."`
~' ; .`,`

` .

'. - , ~

~ ~7 ~ 39 5 sample i.s the sample aromatised to 7ZOO~ qualified as b~lanced9 followed by ~he sample Aromatised ~o 10ho and then by the sample aromatised to 3.5~/Oo. These samples are generally described as having a highly representative aroma and as being free from any after taste of the solvent.
Other instant coffees are then prepared in the same way with the aid of the neutral ba~e powder aroma~ised to 7~ho with the aromatic fractions of Example 1 with various . stripping levels (50, 150, 200 and 250%). The verdict : 10 of the tasters is as follows:
50% fresh aroma, slightly pyrazinic 150% and 100%, balanced 200~/. "burnt" aroma, too heavy 250% like 200%, "burnt",too heavy.
lS The results obtained are similar when the neutral base powder used is a freeze-dried powder.
; EXAMPLES 11 to 13 The procedure is as described in Example 10, using the aromatic fractions o~ Examples 6 to 8. The results obtained are set out in the following Table:

, ,... ~ ~
Examples Aromatic fraction used Degree of Preferences 11 r~ ~ti-~tlu~ - 3 .
. (pentaneJmethylene 7 _ , chloride) 10 2 ; 12 . . . - . 3.5 ~ .
(pentane/diethyl ether) 7 1 _ ~ ~ 3.5 _ 13 (pentane/methylene chloride 7 andpentane/isopropyl 10 3 chloride) .. _ __ _ . .

~15-.

, :
~1~72395 .

An approximately 50 g/l decoction is prepared from a mixture of 40~ of Brazillan santos coffee, 30% of Colombian arabica coffee and 3Gqo of Salvador ~rabic~
coffee. This decoction is strengthened by the addition of 0.7~/Oo (based on the solids) of the aromatic fraction of Example 1 with a stripping level of 100b. Accor~ing to the tasters, the decoction thus strengthened has much better organoleptic characteristics than the untreated starting decoction, and posYesses an extremely intense aroma.
: EXAMPLE 15 ~,~,. ~
300 ~l (i.e. 2.5/~o) of the aromatic fraction of Ex~mple 1 with a stripping level of lO~o are added to one litre of standard commercial-grade pasteurlsed milk with a dry matter content of ~he order of 12%. The ~; beverage thus obtained has an acid taste which, nevertheles~, is very similar to that of a coffee cont~ining milk witho~t ``~ having its appearance.
~ 20 EXAMPLE 16 ;`; A cre~m is prepared from 3.5% of flour, 137. of castor sugar, 2.5% of caramel, 0.02~ of vanillin ~nd approximately 80% of pasteurised milk. After cooking, O.l~t~o by weight of the aromatic fraction of Example 1 (stripping level 100%) is added. This cream con~tîtutes a very pleasant-t~sting coffee-f~voured cream.

An ice cre~m is prepared from 12Z of fresh cream containing 35~ of fat, 7~h of egg yolk, 12~ of sucrose, ~`` 30 3.5g of ~070 glucose 9yrup, 3.5~b of invert ~ugar (glucoRe .

.

.

1(~7Z395 and fructose), 14~b of caramel and 48% of whole milk. Before freezing, 0.1~/oO by weight of the aromatic fraction of Example 1 . (stripping lPvel 100%) is added. This coffeP-flavoured ice cream is highly rated by the tasters.

' .~

~' ..

:' ., '``'. '- ' ` '.' ~ . ..
- . , ~ -

Claims (15)

I CLAIM :

1. A process for isolating an aromatic coffee fraction from an aqueous medium containing said fraction, which comprises contacting the aqueous medium with an organic solvent which is a water-insoluble azeotropic mixture capable of boiling at a temperature below about 50 °C of(a) at least one non-aromatic hydrocarbon and(b) at least one non-aromatic halogenated hydrocarbon or an ether and recovering a solvent phase containing the aro-matic fraction.

2. A process as claimed in Claim 1, wherein the recovered sol-vent phase is concentrated by distillation.

3. A process as claimed in Claim 2, wherein the recovered sol-vent phase is concentrated to a residual solvent content of the order of 10 %.

4. A process as claimed in Claim 1, wherein the azeotropic mix-ture is one of the following azeotropic mixtures:

pentane/methylene chloride 68 - 32 %
pentane/isopropyl chloride 58 - 42 %
pentane-trichloro-trifluoroethane 94 - 6 %
pentane/diethylether 20 - 80 pentane/methylene chloride/
diethylether 34.5 - 3.5 - 62 %

5. A process as claimed in Claim 1, wherein the aqueous medium is an extract of roasted coffee.

6. A process as claimed in Claim 1, wherein the aqueous medium is a condensate of the aroma given off during the roasting of green coffee or during the grinding of roasted coffee.

7. A process as claimed in Claim 1, wherein the aqueous medium is a stripped aroma condensate of roasted coffee.

8. A process as claimed in Claim 7, wherein the aqueous medium is a condensate obtained by stripping of roasted coffee at a stripping level of from 1% to 300 %.

9. A process as claimed in Claim 1, wherein extraction is car-ried out at a temperature in the range from 20 to 30 °C.

10. A process as claimed in Claim 1, wherein extraction is car-ried out at the boiling temperature of the azeotropic mix-ture used.

11. An aromatic coffee fraction obtained by the process claimed in Claim 1, characterised by a stability at -20 °C of at least 6 months.

12. A process for flavouring a material without a coffee aroma so as to provide it with a coffee aroma wherein from 0.5 to 12 %0 (per thousand) of the aromatic fraction of Claim 11 is added to said material.

13. A process for flavouring a material with a coffee aroma so as to strenghten its aroma wherein from 0.1 to 0.6 %0 of the aromatic fraction of Claim 11 is added to said material.

14. A process as claimed in Claim 12, wherein the material is an instant coffee powder.

15. A process as claimed in Claim 13, wherein the material is an instant coffee powder.