Yanxiang Gao:
Extraction of oleoresins from fennel seeds and dried onion with supercritical CO₂

Fennel (Foeniculum vulgare M.) and onion (Allium Cepa L.) are commonly used spices in food industry, their flavors are incorporated in processing foods, mainly by adding the hydrodistilled oils, to some extent, by using the solvent-extracted oleoresins.

Both the hydrodistilled oil and the oleoresin have a different aroma from that of the natural spices. This is due to the heat and water caused changes taking place during distillation in the oil and the solvent residue, removal of more volatiles, etc. in the oleoresin. To obtain the essential oil or oleoresin, devoid of the above limitations, supercritical carbon dioxide ( SC-CO₂ ) as a solvent has many advantages: colorless, odourless, tasteless, non corrosive, chemically inert gas which retains no solvent residue in the extract. The solvation property of the carbon dioxide can also be modified by adjusting pressure and temperature through density and dielectric constant changes. Supercritical CO₂ extraction is more selective than the conventional extraction with commonly used organic solvents. Recently there are a few full-scale plants applying this technology for the extraction of essential oil and oleoresin from several spices.

Only a few studies exist in which the subcritical or supercritical CO₂ extraction of oleoresin or some component from fennel and fresh onion was performed in small-scale apparatus. The yield of oleoresin was 2.0% when unripe fennel umbels were extracted by SC-CO₂ at 200 bar and 40 ^oC (Ravid et al., 1983). Liquid CO₂ gave a 5.8% yield in 2 hours at 58 bar and 20 ^oC when fennel seeds were used (Naik et al., 1989). Extraction of onion flavor from fresh onion was investigated using SC-CO₂ at 200 bar and 35 ^oC, combined gas chromatography - mass spectrometry analysis of SC-CO₂ onion extract showed the presence of 28 sulfur - containing compounds (Nirmal et al., 199?). Thiosulfinates were extracted from onion juice with SC-CO₂ at 240 bar and 35 ^oC, the SC-CO₂ extraction of thiosulfinates was ca. 60% as efficient as the extraction with diethyl ether, as determined by GC-MS (Calvey et al., 1994).

The lack of information on the effect of extraction parameters on the yield and composition of fennel and onion oleoresins, optimisation of the extraction process for them, and fractionation of oleoresin into essential oil and fatty oil, makes the development of industrial products difficulty. At the same time, publications on the sensory evaluation of SC-CO₂ extracts are scarcely available.

Aims of the research work

1. Investigation of the possibility and feasibility of supercritical CO₂ extraction of oleoresins from fennel seeds and dried onion in different scales apparatus.
2. Fractionation of oleoresins into the essential oil rich and fatty oil rich products in pilot-scale apparatus.
3. Optimisation of the extraction process for SC-CO₂ extraction of fennel oleoresin in order to minimise the CO₂ consumption and the unwanted components content in both products at the same level of the yield
4. Optimisation of the extraction parameters for SC-CO₂ extraction of onion oleoresin to obtain high yield and quality of onion product.
5. Comparison of the similarity and difference of the yield, volatile compound composition and sensory characteristics of SC-CO₂ extracts, organic solvent extracts and steam distilled oils of fennel and onion.
6. Modelling the extraction process in order to predict the important parameters such as the maximum yield and overall transfer coefficient.

Steam distillation
The essential oil of fennel seeds was isolated with a Karlsruber apparatus (Stahl. 1953).
The essential oil from onion was extracted in a special steam distillation apparatus (Sass-Kiss, 1989).

Solvent extraction
Laboratory-scale type of Soxhlet extraction apparatus was used for the preparation of oleoresin, and the solvents were hexane and alcohol ( 96 % v/v ).

SC-CO₂ extraction

1. Small-scale SC-CO₂ extraction

Analytical SFE apparatus (Model 7680, Avondale, PA) was used for the extraction of fennel oleoresin, and a laboratory-scale type SFE apparatus (SFX2-10, Isco, Inc., USA) for the extraction of onion oleoresin.
2. Middle-scale SC-CO₂ extraction

A middle-scale SC-CO₂ extraction has been performed with the home-made apparatus (extractor, 1.14L).
3. Pilot-scale SC-CO₂ extraction

For the extraction of fennel and onion oleoresins, the home-made SFE equipment with one extractor ( 5L ) and two separators in series was adopted.

1. GC analysis of volatile compounds in fennel and onion oleoresins For GC/MS analysis of volatile compounds in fennel oleoresin from the small-scale experiment, a Hewlett Packard 5890 Series II gas chromatography equipped with a 7673 autosampler and a 5889 A MS Engine mass spectrometer was used.

For the qualitative and quantitative analysis of the composition of volatile components in fennel oleoresin from the large-scale experiment, gas chromatography was conducted on a Shimadzu GC-A 14 equipped with FID detector.
The composition of volatile components in onion oleoresin was analysed by Varian 3700 gas chromatograph (Varian Associates, Walnut Creek, CA) equipped with FID detector.
2. GC analysis of fatty acid composition of fennel oleoresin The composition of fatty acids of fennel oleoresin was analysed by Hewlett Packard 5720 A gas chromatograph equipped with FID detector.
3. HPLC analysis of pigments in fennel oleoresin

The pigments of fennel oleoresin were fractionated isocratically by reversed-phase HPLC method. A Beckman series of liquid chromatograph equipped with a Model 165 variable wavelength detector was used.
4. Determination of sulfur content of onion essential oil and oleoresin Schöniger's titration method modified by the Department of Analytical Chemistry in the Technical University was adopted.

Experimental design
A central composite rotatable design of second order was used for the SC-CO₂ extraction of fennel and onion oleoresins.

Statistical analysis
All related experimental data were submitted to analysis of variance and response surface, using the STATGRAPHICS computer program (1985).
 

III. SUMMARY OF NEW SCIENTIFIC RESULTS

The new scientific results presented in this dissertation are as follows:

1. The result from small-scale extraction of fennel oleoresin showed that the first-extracted oleoresin contained much more volatile compounds. With the increase of extraction time, the composition of volatile compounds altered and the last-extracted oleoresin only contained trans-anethole, fenchone, methyl chavicol and limonene as principle volatile compounds. Sensory evaluation indicated that the first-extracted oleoresin exhibited much more pleasant flavor than the others.
2. The fennel oleoresin was successfully fractionated into the essential oil rich and fatty oil rich products in the preliminary pilot-scale extraction. With an increase of the pressure from 70 to 90 bar in the first separator, the yield of the fatty oil rich product decreased by 41 %, while the yield of the essential oil rich product increased as much as 5 times.

As compared with other methods, the SC-CO₂ extraction resulted in higher yield of fennel oleoresin than steam distillation, almost the same as hexane extraction.
GC analysis of volatile compounds revealed that the compositions of volatile compounds in fennel essential oil and oleoresins prepared by different solvents were very different. With the increase of extraction time, the composition of volatile compounds in the essential oil rich and fatty oil rich products also changed.
3. Response surface methodology (RSM) was first used to design the large-scale extraction and fractionation experiments of fennel oleoresin. The results from 27 experimental runs indicated that the CO₂ amount consumed for the 80 % relative overall yield with respect to hexane extraction was at the lowest level when the extraction pressure and temperature were above 285 bar and 63 ^oC, respectively. The minimum level of total unwanted components in both separators appeared at the pressure 80-84 bar and temperature 31-35 ^oC in the first separator. The composition of volatile compounds of fennel extracts prepared was not obviously influenced by extraction and separation parameters.

The result from HPLC analysis implied that in most cases the pigments in fennel oleoresin were pheophytins and lutein, and their contents in the fatty oil- rich product increased along with the rise of extraction pressure and temperature and the decrease of separation temperature in the first separator. Extension of extraction time also resulted in an increase of the pigment content. As compared with hexane and alcohol, SC-CO₂ exhibited higher solubility for pigments at high pressure and temperature than hexane, whereas the solubility of the pigments was the highest in alcohol.
The result from sensory evaluation of fennel essential oil and oleoresins showed that the SC-CO₂ product and disitilled essential oil were more intensive in odour and taste than alcohol and hexane extracts.
4. The data from 15 experimental runs for the small-scale extraction of onion oleoresin were used to develop the polynomial model and fitted the experimental data very well (R₂>0.903). The contribution of each parameter to the yield was ranked in the following order: extraction pressure > time >temperature, all of them exerted the significant effect on the yield (p<0.05). According to the contour plot, the maximum yield of onion oleoresin could be reached when the extraction pressure and temperature were above 350 bar and 57 ^oC, respectively. Sulfur content of onion oleoresin was also influenced by extraction parameters, the low pressure and high temperature of extraction led to an increase of sulfur content of onion oleoresin.
5. The yield and sulfur content of onion oleoresin from middle-scale extraction increased obviously with the rise of extraction temperature, however, the extraction pressure generated no significant effect on them.
6. The experimental results from the pilot-scale extraction further confirmed the conclusion from the small-scale extraction: high extraction pressure and temperature resulted in significantly positive influence on the yield. The sulfur content reached maximum level at high temperature and low pressure.

The comparison of the yield and sulfur content, of essential oil and oleoresins prepared by different methods revealed that alcohol extraction possessed the highest yield and extraction efficiency of sulfur. The yield of onion oleoresin by hexane and SC-CO₂ extraction was similar although the extraction efficiency of sulfur by SC-CO₂ extraction was higher than hexane extraction. The yield of onion essential oil obtained by steam distillation was at the lowest level, nevertheless, its extraction efficiency of sulfur was higher than hexane extraction, similar to SC-CO₂ extraction.
The composition- of volatile compounds of onion oleoresins produced by different scales SC-CO₂ extractions was characterised by low concentration of sulfides, di-, tri, tetrasulfides, which are the main compounds of distilled onion oil. The ratio of area % of I/II; (compounds appearing in distilled oil / compounds occuring in extracts only) fractions, except some cases, did not show significant changes by the effect of pressure and temperature.
Comparising the composition of volatile compounds of distilled oil, and the oleoresins prepared by SC-CO₂, hexan, alcohol, it could be suggested that the extracts contained much less sulfides, di-, tri-, tetrasulfides than distilled oil. The similarity of the composition of II. fraction on the chromatograms showed the analogy of extraction methods, but the differences characterise the extraction solvents.
The result from sensory evaluation of onion essential oil and oleoresins indicated that the SC-CO₂ extract was of the best quality, distilled essential oil took the second position and the worst one was alcohol extract.
7. Based on the experimental data gathered from the SC-CO₂ extraction of fennel and onion oleoresins, a mathematical model was developed using differential mass balances, and an analytical solution with two parameters (y_¥ , k') was obtained. The model was checked by the experimental data and a good fit was found. y_¥ and k' in the model were estimated by the nonlinear regression program and statistical analysis of their values implied that y_¥ was controlled by the extraction temperature and k' by the extraction pressure for the SC-CO₂ extraction of fennel oleoresin. For the SC-CO₂ extraction of onion oleoresin, k' was associated with particle size distribution and y_¥ was influenced by the extraction pressure and temperature. The reasonable explanation for this conclusion was also presented.

IV. APPLICABILITY OF THE RESEARCH RESULTS

The experimental data, in particular, the optimum extraction parameters from the research work can be directly used to guide the practical design of SC-CO₂ extraction technology of fennel and onion oleoresins. The selection and design of SC-CO₂ extractor can be also performed on the basis of optimum parameters and developed mathematical model. In order to decrease production cost and improve the product quality, the desired yield of oleoresin can be predicted by the model.

Sensory analysis program can be successfully applied to evaluate and compare the quality of SC-CO₂ extracts, solvent extracts and distilled oil. This is much more economic method than instrumental analysis and it can be extended to other spice products.