2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering

EXTRACTION OF GARLIC OIL WITH QUASI-CRITICAL SOLVENTS

A.E. Andreatta, G. Foco, G. Mabe and S.B. Bottini
PLAPIQUI, Universidad Nacional del Sur-CONICET Abstract. Garlic oil can be extracted from fresh garlic bulbs using traditional hydro-distillation or steam stripping processes or alternatively by applying new techniques such as high-pressure extraction with nearcritical fluids. In this respect, carbon dioxide is the preferred near-critical solvent, because it has a low critical temperature, it is non-toxic, non-flammable, easily available and relatively cheap. Garlic oil contains sulfur organic compounds such as allyl sulfides, produced by the action of an enzyme (alliinase) on precursor amino acids. The oil is used not only as a flavor agent, but also as a nutraceutical. In this work garlic oil has been extracted from fresh crushed bulbs, applying the traditional techniques of hydro-distillation and steam stripping at atmospheric pressures, and a high-pressure Soxhlet extractor working with near-critical carbon dioxide at a temperature of 288K and a pressure equal to 51 bar. The extracts were analyzed by gas and liquid chromatography (HPLC and GC-MS). The oil extracted with liquid CO2 has the typical flavor of fresh garlic, while the oils obtained by hydro-distillation and steam stripping have the characteristic smell of cooked garlic. The presence of sulfoxides in the extracts obtained with near-critical CO2 give an indication of the difference in quality between the processes applied.

Keywords: Garlic, Extraction, Quasi-Critical CO2.

1. Introduction
The processing of garlic bulbs has considerable advantages for the food industry, since it reduces the problems derived from seasonal variations, handling and transport of the fresh product. The oil of garlic, rich in sulfur organic compounds, is used not only as flavor agent, but also in the prevention and treatment of several illnesses. Cysteine-sulfoxides and -glutamyl-cysteines are the main sulfur compounds contained in fresh garlic bulbs and they are present at approximately equal quantities. While the first group of compounds are the precursors of the characteristic garlic flavor, the second group are considered as a reserve for the first ones, since -glutamyl-alk(en)cysteines can turn into alk(en)cysteine-sulfoxides during storage and bulb sprouting (Yu et al.,1994). The olfactory and gustatory principles of freshly cut garlic are associated with the presence of alk(en)yl thiosulfinate RS-(O)-SR and related compounds, produced by the action of the enzyme allinase on precursor amino acids such us alk(en)cysteine-sulfoxides, when the cellular tissue is disrupted (Whitaker, 1976). Allicin (diallyl thiosulfinate) is the main thiosulfinate, responsible of many of the bioactive attributes of garlic extracts. Harsh processing conditions such as hydro-distillation or steam stripping convert the thiosulfinates into saturated and unsaturated polysulfurs and sulfur heterocycles (Block et al., 1992). The alkyl thiosulfinates and thiosulfonates have been associated with the flavor of fresh garlic, while the alk(en)yl polysulfurs have the typical flavor of cooked garlic. In this work an alternative extraction process to obtain garlic oil using near-critical carbon dioxide is evaluated and compared to the traditional methods of hydro-distillation and steam stripping.

Address: PLAPIQUI, Camino la Carrindanga Km 7, casilla717, 8000 Baha Blanca - Argentina. E-mail: sbottini@plapiqui.edu.ar

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering
The use of near-critical carbon dioxide in the processing of food, nutraceuticals, flavors and aromas is attractive, due to the low operation temperatures required, the absence of oxygen in the extraction system and the facility of solvent recovery. CO2 is also non-toxic, non-flammable, easily available and relatively cheap. Liquid carbon dioxide is a selective solvent for the extraction of aromatic compounds such as esters, aldehydes, and ketones. Salt, sugars, fruit acids and amino acids are insoluble, whereas water is soluble in liquid carbon dioxide up to 0.1 % by weight (Shultz and Randall, 1970).

2. Experimental
Extractions were carried out using traditional hydro-distillation, steam stripping and liquid CO2 at nearcritical conditions. 200 g of fresh garlic bulbs were chopped into small pieces using a domestic blender at medium speed during 1 minute. For the hydro-distillation experiments, the garlic paste was homogenized in 200 ml distilled water at room temperature and macerated during 1 hour, prior to the extraction. No extra water was added in the extractions with steam stripping process and liquid CO2.

2.1. High-pressure Soxhlet extractor Figures 1 and 2 show a scheme of the high-pressure Soxhlet extractor used in the extractions with nearcritical CO2. The apparatus is based on the design by Jenning (1981) and was built at our workshop. It consists basically of a standard glass Soxhlet placed inside a 5 liters steel cylinder. A metal heating tape that embraces the steel cylinder heats and evaporates the solvent at the bottom. A cold finger hanging from the lid of the cylinder allows solvent condensation and dripping onto the extraction material. The number of extraction cycles is registered by following the sudden change of temperature produced after completion of every cycle. For that purpose a thermistor is located near the discharge siphon of the Soxhlet. After a certain number of cycles the extraction is considered completed and the experiment is stopped. An expansion valve connected to the lid of the cylinder is then opened and CO2 slowly discharges into the atmosphere. The garlic oil is finally removed from the Soxhlet flask. Further details of the apparatus and experimental procedure are given elsewhere (Marchiaro, 2000). The experiments were carried out at a temperature of 15C; the pressure inside the cylinder was equal to the vapor pressure of CO2 at the corresponding experimental temperature. 2.2. Analytical tools Four different techniques were applied to characterize the garlic extracts: GC-MS, HPLC, UV and FTIR. The GC-MS analysis were carried out in a Hewlett Packard HP 6890 gas chromatograph equipped with a selective mass spectrometer detector (Hewlett Packard HP 5972A). A capillarity HP fused silica column of 30 m length, 0.25 mm internal diameter and 0.25m film thickness was used. The operating conditions were as follows: injector temperature 250C, gas carrier (He) 1 ml/min, sample volume 5ml, detector temperature 280C, energy of ionization 70 eV, initial oven temperature 60C (4 minutes) and a heating ramp of 4C/min up to 200C. 2

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering

Fig.1. Experimental set-up: V: valves; G2: CO2 reservoir; MA: pressure gauge; CIE: closing clamps; CCE: extraction chamber; BA REF: thermostatic bath; SU CAL: heating resistance; REG: register; CC: cycle counter; SICAL: thermistor; DR: rupture disk; CT: temperature controller. Fig. 2. High-pressure Soxhlet extractor: CLACL: closing clamps; CCE: extraction chamber; BA REF: thermostatic bath; SU CAL: heating resistance; SICAL: thermistor; TR: condenser; TP: solvent port; SE: extracting material; SIF: siphon; BALSOX: Soxhet flask; T: thermistor; S3VAP: cycle counter.

A SEEN Varian 5500 liquid chromatograph with a photodiode array detector (Waters 996) was used for the HPLC analysis. The column applied was a Varian MicroPack Sy-10m, 300mm length x 4mm internal diameter. A constant flow rate of 1.8 ml/min was used, with an initial composition of 2% 2-propanol / 98% hexane during 6 min, a linear increase of 10% 2-propanol during 10 min, and a hold at 10% for 9 min, before returning to the initial composition over 5 min. Organosulfur compounds were detected by UV absorption at 254nm. A SHIMADZU 160A UV-visible spectrophotometer of simple bundle with a deuterium lamp as source of excitation and a detector of diodes between 200 and 1000 nm was also used to analyze the garlic extracts. Finally, a Nicolet 520 FTIR Spectroscope was used to analyze the possible group configuration of the different garlic extracts.

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering

3. Experimental Results
A series of extractions were carried out to evaluate the effect of process variables on yield and quality of extracts. The yields and physical properties of garlic oils obtained by hydro-distillation, steam stripping and high-pressure extractions with CO2 are given in Table 1. The chemical composition of the extracts was analyzed using the analytical tools described above. The results are compared with previous work from the literature on onion oil (Dron et al., 1997; Simandi et al., 2000). A maceration period of 45 minutes prior to extraction of the garlic paste was sufficient. Longer periods did not give higher yields. Also the addition of extra water to the garlic paste did not affect the yields. In hydrodistillation and steam stripping, extraction periods of 2h 30 were sufficient to exhaust the raw material. For high-pressure extractions with quasi-critical CO2, 8 cycles were adequate.
Table 1. Physical Properties and Yields of Garlic Extracts

Colour Density (g/ml) Yield (weight %)

Hydro-distillation Light yellow 1.08 0.21 0.01

Steam Stripping Light yellow 1.08 0.22 0.01

Liquid CO2 Dark yellow 1.18 5

3.1. Quality of the extracts obtained by the different extraction processes The oil extracted with liquid CO2 has the typical flavor of fresh garlic, while the oils obtained by hydrodistillation and steam stripping have the characteristic smell of cooked garlic. This organoleptic difference is attributed to the presence of saturated and unsaturated thiosulfinates and thiosulfonates in the extracts obtained with near-critical CO2. These compounds are missing in the oils obtained by hydro-distillation and steam stripping due to the high operating temperatures applied in these processes; the main components of these oils are methyl and methyl-allyl polysulfides. Table 2 shows the garlic oil compositions determined by GC-MS. The thermoliability of thiosulfinates and thiosulfonates precludes the identification of these compounds by GC-MS; these compounds decomposes at the high temperatures of the GC injector and detector (Block et al.,1992; Block et al., 1993). HPLC, cryogenic GC and supercritical chromatography have been suggested as appropriate techniques to avoid thermal decomposition. While GC-MS operating conditions will not alter the composition of the garlic oils obtained by hydro-distillation and steam stripping because they have been already subjected to high temperatures, the situation is different for the extracts obtained with liquid CO2. An alternative method should then be explored to determine the presence of thiosulfinates and thiosulfonates in these extracts. Mono-, di- and trisulfides were the main components determined by GC-MS in the garlic oils obtained by hydro-distillation and steam stripping. Semmler et al. (1892) established the importance of diallyl disulfide and diallyl trisulfide in the flavor of these garlic extracts. The remaining aqueous phase in these processes is white and it contains some volatile water-soluble compounds (Yu et al., 1989). Boelens et al. (1971) suggests that the absence of thiosulfinates in the garlic oils obtained by hydro-distillation and steam stripping is due to the

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering
Table2. Results of the GC-MS analysis of garlic extracts Common compound Diallyl sulfide Diallyl trisulfide Allyl methyl sulfide Allyl methyl disulfide Dimethyl trisulfide Diallyl disulfide Diallyl tetrasulfide Dimethyl disulfide Dimethyl tetrasulfide Allyl methyl trisulfide 1-propenyl methyl disulfide 1 oxa-4-6-diazocyclooctane-5thione Toluene 3,5 diethyl 1,2,4 trithiolane Allyl isothiocyanate 1-3 Dithiane 3H-1,2-dithiol-3-one,4,5 dimethyl Di 1-propenyl sulfide 5 methyl 1,2,3thidiazole 1-propenyl methyl sulfide (Z) 1,2,5,trithiepane 2,4 dimethyl thiopheno 4 methyl thiocyclohexanone 1,3 Butadiene Ethyl Thiourea 3,4 dimethyl 1,5 heptadiene Propenyl ethynyl sulfoxide 2,3-dimethyl-, cis Oxirane Acetic acid Acetoacetic acid, 1-thio-,S-allyl ester 2 Furancarboxaldehyde 2- Furanmethanol 2- thiouracil Thiopheno Dihydro-3 methyl 2,5 Furandione Propane, 1 isothiocyanate 2 methyl - 5- (methylthio)- Furan 3 ethyl-2,5-dimethyl- Pyrazine Furyl hydroxymethyl ketone Methyl 2-furoate 5-isopropyltetrahydrothiophen-3-one 2-thiophenecarboxilic acid, 5-(1,1,-dimethylethoxy) 2,4 (1H,3H)-Pyrimidinedithione 1,2,3, thiadiazole, 5 methyl5-acetoxymethyl-2-furaldehyde 4-mercaptophenol 5-methyl-2-thiophenecarboxaldehyde 3-(1,1-dimethylethyl)- thiophene Sulfur, mol. Hydro-distillation 2.207 24.253 0.893 7.231 2.304 34.276 1.019 0.689 0.138 13.105 0.220 8.398 0.117 0.248 0.099 1.570 2.174 0.718 0.341 % peak area Steam Stripping Liquid CO2 2.783 0.051 27.635 0.538 1.127 0.436 7.738 0.420 3.289 0.774 25.838 1.897 1.054 0.176 15.961 1.108 7.458 0.104 0.346 0.183 1.606

0.122 0.164 0.146 0.509 0.166 0.147 0.443 0.165 0.260 0.253 0.988 2.370 2.463 0.326 0.366 2.050 0.932 1.212 0.366 0.236 0.336 3.192 2.887 4.992 0.487 0.661 72.147 0.318 0.285 0.489

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering
solubility of these compounds in water. According to these authors they are present in the aqueous layer and cannot be transferred to the organic phase during the process. The results of GC-MS for the garlic extracts produced by liquid CO2 show many compounds that are not detected in the oils obtained by hydro-distillation and steam stripping (see Table 2). Figure 3 shows the results of HPLC analyses carried out on extracts obtained by high-pressure liquid CO2 to determine the presence of thiosulfinates RS(O)SR. These compounds were characterized by direct comparison with results reported in the literature by Block et al. (1992). The compounds were identified as methyl (Me), allyl (All) and propenyl (Pro) thiosulfinates: AllSS(O)Pro-(E) (1); AllS(O)SPro(Z) (2); AllS(O)SAll (3); AllS(O)SMe (4) and MeS(O)SMe (5). Though the shape of the chromatogram shown in Figure 3 is very similar to that reported by Block et al. (1992), there is a difference of about 3 minutes in the retention times. This discrepancy can be attributed to the difference in length, internal diameter and particle-size of the chromatographic columns used: a 10m 300mm x 4mm column was used in our work, while Block and co-workers used a 5 m 250mm x 4.6mm column. To confirm the presence of allicin (AllS(O)SAll) in this extract, a sample was sent to the Laboratory of Toxicological Analysis of Residues from Facultad de Ciencias Agrarias (Universidad Nacional de Cuyo) for analysis. This laboratory confirmed the present of 3,58 mg of allicine/g in the extract, by comparison with a synthesized standard. Figure 4 shows the results of an HPLC analysis on garlic oil obtained by hydrodistillation. No allicin was detected in this oil. The peak at 2 minutes could correspond to decomposition products of thiosulfinates compounds.

12345-

AllSS(O)Pro(E) AllS(O)SPro-(Z) AllS(O)SAll AllS(O)SMe MeS(O)SMe

minutes

Fig. 3. Si-HPLC separation of thiosulfinates in garlic extract obtained by high-pressure liquid CO2 extraction

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering

minutes

Fig. 4. Si-HPLC of garlic oil obtained by hydro-distillation

Figure 5 shows the UV spectrum of the samples obtained for hydro-distillation and high-pressure liquid CO2 extraction.
2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 200 220 240 260 280 300

Absorbance A

Hydro-dist.

Ext. with CO2

wavelength (nm)

Fig. 5. UV Spectra for extracts obtained by hydro-distillation and quasi-critical CO2 extraction

These spectra confirm the different nature of the garlic extracts. The spectrum of the extract obtained with liquid CO2 coincides with the spectrum of allicin (allyl thiosulfinate), which presents a maximum of absorption at 250 nm (Calvey et al., 1994). Figures 6 and 7 show the results of the FTIR spectroscopic measurements.

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering
100 90 80 70 60 50 40 30 20 10 0

Transmitance

Ext. with liquid CO2

3900 3700 3500 3300 3100 2900 2700 2500 2300 2100 1900 1700 1500 1300 1100 900 700

wavenumbers (cm-1)
Fig. 6. IR Spectrum of garlic extract obtained with near-critical CO2
100 90 80 70 60 50 40 30 20 10 0

Transmitance

Hydro-dist. 3900 3700 3500 3300 3100 2900 2700 2500 2300 2100 1900 1700 1500 1300 1100 900 700

wavenumbers (cm-1)
Fig. 7. IR Spectrum of garlic extract obtained by hydro-distillation The IR spectrum of the sample extracted with near-critical CO2 displayed sulfoxide C-S(O)-C stretch between 1070-1035 cm-1, in addition to strong C=C absorption between 1680-1625 cm-1, thiosulfinate C-S(O)SC absorption between 1095-1080 cm-1 (Block et al, 1986), sulfone -(SO2)- absorption between 1350-1300 cm-1 and 1160-1120 cm-1, CH3 absorption between 3100-2900cm-1 and 1470-1430cm-1, S-C absorption between 800700 cm-1 and S-S absorption between 500-400 cm-1.The IR spectrum of the sample extracted by hydrodistillation displayed CH3 absorption between 3100-2900cm-1 and 1470-1430cm-1, CH absorption between 1400-1200cm-1, C=C absorption between 1680-1625 cm-1, S-C absorption between 800-700 cm-1 and S-S absorption located between 500-400 cm-1.

4. Conclusions
The use of liquid carbon dioxide at near-critical conditions represents an attractive option for the extraction of garlic oil, since it produces higher yields and extracts of better quality than the traditional hydro-distillation and steam stripping processes. The extracts obtained with liquid CO2 have the typical flavor of fresh garlic, while the oils obtained by hydro-distillation and steam stripping have the characteristic smell of cooked garlic. The presence of the biologically active thiosulfinates in the extracts obtained with near-critical CO2 gives an indication of the difference in quality between the processes applied. 8

2nd Mercosur Congress on Chemical Engineering 4th Mercosur Congress on Process Systems Engineering

References
Block, E., Thiruvazhi, M. (1993). Allium chemistry: Synthesis of alk(en)yl 3,4-dimethyl-2-thienyl disulfides, components of distilled oils and extracts of Allium species, J. Agric. Food Chem., 41, 2235. Block, E., Ahmad, S., Catalfamo, J., Jain, M., Apitz-Castro, R. (1986). Antithrombotic organosulfur compounds from garlic: structural, mechanistic, and synthetic studies, J. Am. Chem. Soc,, 108(22), 7045. Block, E., Naganathan, S, Putman, D., Zhao, S. (1992). Allium chemistry: HPLC analysis of thiosulfinates from onion, garlic, wild garlic (ramsoms), leek, scallion, shallot, elephant (great-headed) garlic, chive, and Chinese chive. Uniquely high allyl to methyl ratios in some garlic samples. J. Agric. Food Chem.,40, 2418. Boelens,M., de Valois, P., Wobben, H.J., Van der Gen, A.(1971). Volatile flavor compounds from onion. J. Agric. Food Chem., 19, 984. Calvey, E., Matusik J. E., White K.D., Betz J. M., Block E., Littlejohn M. H., Naganathan S. and Putman D.(1994). Off-Line Supercritical Fluid Extraction of Thiosulfinates from Garlic and Onion. J. Agric. Food Chem., 42(6); 1335. Dron A., Guyer D.E., Gage D.A., Lira C.T., (1997). Yield and quality of onion flavor obtained by supercritical fluid extraction and other methods. J. Food Process. Eng., 20, 107. Jenning (1981). High Presure Soxlet Extractor design. US. Patent 4,265. Marchiaro, A. (2000). Master Thesis in Chemical Engineering, UNS. Extraccin de Aceite de Rosa Mosqueta con gases densos. Semmler, F.W. (1892). Arch Pharm. 230. Shultz, W.G., Randall,J.M. (1970). Liquid carbon dioxide for selective aroma extraction. Food Technol., 24, 1282. Simndi B., Sass-Kiss A., Czukor B., Deak A., Prechl A., Csordas A. And Sawinsky J. (2000). Pilot-scale extraction and fractional separation of onion oleoresin using supercritical carbon dioxide. J. of Food Engineering, 46, 183. Whitaker J.R. (1976).Development of flavor, odor and pungency in onion and garlic. Adv. Food Res., 22, 73. Yu, T., Wu, C., Rosen, R., Hartman, T. Ho, C. (1994). Volatile Compounds Generated from Thermal Degradation of Alliin and Deoxyalliin in Aqueous solution. J. Agric. Food Chem.,42, 146. Yu, T., Wu, C., Rosen, R., Liou, Y. (1989). Volatile Compounds from garlic. J. Agric. Food Chem., 37, 725.

Acknowledgments
The authors are grateful to CONICET and Universidad Nacional del Sur for financial support.