Supercritical fluid extraction in plant

() Journal of Chromatography A, 1163 (2007) 2–24 Review Supercritical fluid extraction in plant essential and volatile oil analysis Seied Mahdi Pourmortazavi ∗, Seiedeh Somayyeh Hajimirsadeghi Faculty[.]

Journal of Chromatography A, 1163 (2007) 2–24 Review Supercritical fluid extraction in plant essential and volatile oil analysis Seied Mahdi Pourmortazavi ∗ , Seiedeh Somayyeh Hajimirsadeghi Faculty of Material and Manufacturing Technologies, Malek Ashtar University of Technology, P.O Box 16765-3454, Tehran, Iran Received 21 February 2007; received in revised form June 2007; accepted June 2007 Available online 17 June 2007 Abstract The use of supercritical fluids, especially carbon dioxide, in the extraction of plant volatile components has increased during two last decades due to the expected advantages of the supercritical extraction process Supercritical fluid extraction (SFE) is a rapid, selective and convenient method for sample preparation prior to the analysis of compounds in the volatile product of plant matrices Also, SFE is a simple, inexpensive, fast, effective and virtually solvent-free sample pretreatment technique This review provides a detailed and updated discussion of the developments, modes and applications of SFE in the isolation of essential oils from plant matrices SFE is usually performed with pure or modified carbon dioxide, which facilitates off-line collection of extracts and on-line coupling with other analytical methods such as gas, liquid and supercritical fluid chromatography In this review, we showed that a number of factors influence extraction yields, these being solubility of the solute in the fluid, diffusion through the matrix and collection process Finally, SFE has been compared with conventional extraction methods in terms of selectivity, rapidity, cleanliness and possibility of manipulating the composition of the extract © 2007 Elsevier B.V All rights reserved Keywords: Supercritical fluid extraction; Essential oils; Antioxidant; Cartenoids; Terpenes; Volatile components Contents Introduction Introduction to supercritical fluid extraction Solubility and mass-transfer rate of plant oils in supercritical fluid Effect of matrix on supercritical fluid extraction Effect of extraction parameters 5.1 Effect of pressure and temperature 5.2 Effect of modifiers on supercritical fluid extraction 5.3 Effect of extraction time 5.4 Effect of flow rate 5.5 Sample particle size and packing density 5.6 Effect of water in supercritical fluid extraction 5.7 Drying effect 3 7 10 12 12 13 14 14 Abbreviations: AHF, adhyperforin; AIDS, acquired immunodeficiency syndrome; AZA-A, azadirachtin A; BHT, butylated hydroxytoluene; CER, constant extraction rate; CWO, cedarwood oil; YCER , concentration of the oil in the supercritical phase; CC-SFE, countercurrent supercritical fluid extraction; DAD, diode-array detection; DPPH, 1,1-diphenyl-2-picrylhydrazyl; kf , external mass-transfer coefficients; FID, flame ionization detection; GC-O, gas chromatography-olfactometery; HD, hydrodistillation; HS-SPME, headspace solid-phase microextraction; HF, hyperforin; HM, hydrocarbon monoterpene; HS, hydrocarbon sesquiterpene; LC-CO2 , liquid carbon dioxide; MCER , mass of extract at constant extraction rate; MS, mass spectrometry; MEKC, micellar electrokinetic chromatography; MWHD, microwave assisted hydrodistillation; ODS, octadecylsilica; OM, oxygenated monoterpene; OS, oxygenated sesquiterpene; Q, solvent flow rate; RPLC, reversed-phase liquid chromatography; Re, Reynolds number; SDE, simultaneous distillation–extraction; SFC, supercritical fluid chromatography; SFE, supercritical fluid extraction; SJW, St John’s Wort (Hypericum perforatum L.); S/F, solvent-to-feed ratio; SPE, solid-phase extraction; S-HS, static headspace; tCER , time of the CER period; Y* , solubility; YCER , oil concentration at the tCER ; YieldTOTAL , total amount of solute collected for the entire measuring time ∗ Corresponding author Tel.: +98 2122952285; fax: +98 2122936578 E-mail address: pourmortazavi@yahoo.com (S.M Pourmortazavi) 0021-9673/$ – see front matter © 2007 Elsevier B.V All rights reserved doi:10.1016/j.chroma.2007.06.021 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Collection of extracted analyte 6.1 Solvent collection 6.2 Solid-phase collection 6.3 Collection in empty vessels 6.4 Novel collection methods 6.5 On-line coupling of supercritical fluid extraction with chromatographic techniques Extraction of oxygenated compounds from plant materials by supercritical fluid extraction Extraction of terpenes and sesquiterpenes by supercritical fluid extraction Comparison of supercritical fluid extraction to conventional methods 10 Conclusion References 15 15 15 16 16 17 17 18 20 22 22 Introduction Introduction to supercritical fluid extraction Essential oils represent a small fraction of a plant’s composition but confer the characteristic for which aromatic plants are used in the pharmaceutical, food and fragrance industries Essential oils have a complex composition, containing from a few dozen to several hundred constituents, especially hydrocarbons (terpenes and sesquiterpenes) and oxygenated compounds (alcohols, aldehydes, ketones, acids, phenols, oxides, lactones, acetalse, ethers and esters) Both hydrocarbons and oxygenated compounds are responsible for the characteristic odors and flavors The proportion of individual compounds in the oil composition is different from trace levels to over 90% (δ-limonene in orange oil) The aroma’s oil is the result of the combination of the aromas of all components Trace components are important, since they give the oil a characteristic and natural odor Thus, it is important that the natural proportion of the components is maintained during extraction of the essential oils from plants by any procedure [1] Steam distillation has traditionally been applied for essential oils recovery from plant materials One of the disadvantages of the hydrodistillation methods is that essential oils undergo chemical alteration and the heat-sensitive compounds can easily be destroyed Therefore, the quality of the essential oil extracts is extremely impaired [2] The extraction of essential oil components using solvents at high pressure, or supercritical fluids, has received much attention in the past several years, especially in food, pharmaceutical and cosmetic industries, because it presents an alternative to conventional processes such as organic solvent extraction and steam distillation [3] The increasing use of vegetable extracts by the food, cosmetic, and pharmaceutical industries can make the extraction of essential oils using supercritical carbon dioxide an attractive technology compared to conventional processes with respect to the product quality [4–6] The knowledge of the mass-transfer mechanism, the kinetics parameters and the thermodynamics restrictions of the extraction conducted in a bed of vegetable material can be used to economically evaluate the extraction process This requires information on the thermodynamic restrictions of the system vegetable material/CO2 On the other hand, the understanding of the various process variables and how they can be connected to a theoretical model to describe the extraction kinetics are also desirable Supercritical fluids have been used as solvents for a wide variety of applications such as essential oil extraction [3], metal cation extraction [7,8], polymer synthesis [9] and particle nucleation [10,11] In practice, more than 90% of all analytical supercritical fluid extraction (SFE) is performed with carbon dioxide (CO2 ) for several practical reasons Apart from having relatively low critical pressure (74 bar) and temperature (32 ◦ C), CO2 is relatively non-toxic, non-flammable, available in high purity at relatively low cost, and is easily removed from the extract In the supercritical state, CO2 has a polarity comparable to liquid pentane and is, therefore, best suited for lipophilic compounds The main drawback of CO2 is its lack of polarity for the extraction of polar analytes [12] In the 1990s, some reports were published about the choice of N2 O as extraction fluid for analytical SFE [13,14] This fluid was considered better suited for polar compounds because of its permanent dipole moment One of the applications where N2 O has shown significant improvements when compared to CO2 is for example the extraction of polychlorinated dibenzodioxins from fly ash [13] Unfortunately, this fluid has been shown to cause violent explosions when used for samples having high organic content and should, therefore, be used only when absolutely necessary [13,14] Other more exotic supercritical fluids which have been used for environmental SFE are SF6 and freons SF6 is a non-polar molecule (although easy polarizable) and as a supercritical fluid, it has been shown to selectively extract aliphatic hydrocarbons up to around C-24 from a mixture containing both aliphatic and aromatic hydrocarbons Freons, especially CHClF2 (Freon-22), has on several occasions been shown to increase the extraction efficiency compared to conducting extractions with CO2 [15] Although supercritical H2 O has often been used for the destruction of hazardous organics, the high temperature and pressure needed (T > 374 ◦ C and P > 221 bar) together with the corrosive nature of H2 O at these conditions, has limited the possible practical applications in plant oil analysis [16] H2 O at subcritical conditions has, however, been shown to be an effective fluid for the extraction of several classes of essential oil In 2000, Gamiz-Garcia et al [17] tested a static–dynamic subcritical water extraction for the isolation of fennel oil Their results showed that, subcritical water extraction is an efficient method in terms of quantitative composition of the extract 4 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Ethane, propane, ethylene, dimethyl ether, etc have also been recommended as solvents under sub- and supercritical conditions for extraction [18,19] Catchpole et al [20] extracted ginger, black pepper, and chilli powder using near-critical carbon dioxide, propane, and dimethyl ether on a laboratory scale to determine the overall yield and extraction efficiency for selected pungent components They also determined the volatile contents of ginger and black pepper extracts Extraction of all spice types was carried out with acetone to compare overall yields Their results showed that subcritical dimethyl ether was as effective at extracting the pungent principles from the spices as supercritical carbon dioxide, although a substantial amount of water was also extracted Subcritical propane was the least effective solvent All solvents quantitatively extracted the gingerols from ginger The yields of capsaicins obtained by supercritical CO2 and dimethyl ether were similar and approximately double that extracted by propane The yield of piperines obtained by propane extraction of black pepper was low at ∼10% of that achieved with dimethyl ether and CO2 , but improved with increasing extraction temperature Mohamed et al [21] explored supercritical extraction using ethane and CO2 for the recovery of the methylxanthines caffeine and theobromine and cocoa butter from cocoa beans using a high-pressure apparatus Their finding showed that the extraction yields of cocoa butter obtained with ethane are much higher than those obtained with CO2 because of the higher solubility of this fat in ethane A pronounced effect of pressure on the extraction of methylxanthines and cocoa butter was observed for both solvents The methylxanthines in cocoa beans were slightly more soluble in ethane than in CO2 probably because of co-solvency effects of cocoa butter, which was extracted more easily using supercritical ethane Despite the higher cost of ethane, its critical pressure is lower than that of CO2 , and the higher butter solubility could render ethane a viable solvent through lower energy costs Jaubert et al [27] chose the ternary system CO2 – limonene–citral as a model system in order to study the extraction of terpenes from lemon oil using supercritical carbon dioxide Extractions were performed at several pressures and temperatures to evaluate the influence of these parameters on the separation efficiency They used a theoretical model, based on a modified Peng–Robinson equation of state to understand the thermodynamic and mass-transfer aspects of the extraction column The critical parameters and the acentric factors of limonene and citral were estimated by group contribution methods They applied the method developed by Abdoul et al [28], which allows the calculation of binary interaction parameters (kij ) to terpenic compounds The extraction experiments were simulated using this model, and the extraction profiles were accurately reproduced On the other hand, Gaspar et al [29] measured the solubility of borage, echium, and lunaria oils in compressed CO2 using the dynamic method Pressure and temperature were varied from 60 to 300 bar and 10 to 55 ◦ C, respectively Their measured solubilities of echium, borage, and lunaria oils in compressed CO2 are presented in Table As shown in this table, at a given temperature, the solubility increases with the increase of pressure, as a direct result of the increased solvent density They showed that the effect of extraction temperature is also similar for all oils At low pressures (60 and 100 bar), an increase of temperature leads to a decrease of solubility, and the opposite is observed at the highest tested pressure (300 bar) At 200 bar, there is an improvement in solubility when increasing the temperature from 10 to 25 ◦ C Also, they compared the solubilities to those of other vegetable oils and were correlated using the density-based model proposed by Chrastil [29] They predicted by the model, a linear relationship between the logarithm of solubility and the logarithm of solvent density was obtained Average deviation between their measured and calculated solubility did not exceed 14% Solubility and mass-transfer rate of plant oils in supercritical fluid Table ∗ ) in compressed Results of experimental study of solubility for seed oils (Cexp CO2 [29] There are many variables to be considered in SFE and method development can seem a daunting task The choice of extraction conditions has largely been determined empirically, which is time consuming One initial area that must be assessed is the solubility of the analyte to be extracted in the supercritical extracting fluid This can be investigated by spiking an inert medium, usually celite or sand, with the analyte of interest In addition to providing an indication of the solubility of the analyte in the supercritical fluid, additional information is also obtained relating to the efficiency of collection of the analyte after depressurization The time consuming nature of even these simple experiments has led several groups of workers to propose techniques for modeling analyte solubility [22] Fundamentals, theory and equations of states for estimating the solubility of various compounds can be found in the literature [23–26] In this review we will focus on the investigations involving the solubility of plant volatile components P (bar) T (◦ C) ρ (kg/m3 ) ∗ (kg/m3 ) Cexp Echium Borage Lunaria 60 10 25 40 55 883.8 190.5 149.2 129.6 3.17 1.90 0.11 0.00 0.00 0.49 100 10 25 40 55 921.9 819.5 629.3 327.1 4.78 3.01 0.81 0.02 2.80 1.75 0.26 0.13 0.64 0.52 0.09 0.00 200 10 25 40 55 980.8 915.2 840.8 755.5 7.72 8.61 7.31 5.27 5.07 5.77 4.39 2.05 1.10 1.70 1.21 0.65 300 10 25 40 55 1020.2 966.8 910.3 850.6 10.20 13.44 14.56 16.35 6.93 9.87 9.90 9.96 1.40 2.88 2.84 3.23 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Fig Comparison of solubility values for the linalool in supercritical CO2 obtained by different studies (values obtained by Berna et al.: (䊉) at 318.15 K, () at 328.15 K Values obtained from Iwai work: (△) at 313.15 K, () at 323.15 K, () at 333.15 K) (From ref [30] with permission.) Berna et al [30] measured the solubilities of essential oil components of orange in supercritical carbon dioxide In their study, the solubilities of pure limonene and linalool in compressed carbon dioxide were measured using a flow apparatus at 318.2 and 328.2 K and pressures ranging from 69 to 111 bar The obtained values of these solubilities are shown in Figs and From the values that were obtained, Berna et al showed that the solubility of limonene in supercritical carbon dioxide is higher than the solubility of linalool at the same conditions but that at higher pressures they approach each other Also, both systems show a sudden increase in the solubility at pressures up to approximately 80 bar For the linalool–CO2 system, it can be noticed that when the temperature rises at pressures under 80 bar, the solubility increases but at pressures over 80 bar, the solubility decreases The comparison between their results and previous studies for the system limonene + CO2 showed that the solubility increases when the temperature rises at pressures under 80 bar but at pressures over 80 bar, it decreases Also, it can be observed that this Fig Comparison of solubility values for the limonene in supercritical carbon dioxide obtained at different studies (values from Berna et al.: (䊉) at 318.15 K Values from Iwai et al.: (△) at 313.15 K, () at 323.15 K, () at 333.15 K Values from Matos et al.: () at 323.15 K Values from Di Giacomo et al.: () at 323.15 K) (From ref [30] with permission.) behavior is similar to that of the system linalool + CO2 This behavior shows that the best conditions for supercritical extractions in both systems will be at pressures >90–100 bar and a temperature near the critical temperature of carbon dioxide in order to obtain the maximum amount of product Their results were in agreement with those of kinetic studies which fixed the best pressure for the extraction or deterpenation of citrus oil peel at about 200 bar [31] Also, they successfully modeled the solubilities using equations of state (Peng–Robinson, Soave–Redlich–Kwong, 3P1T, Dohrn and Prausnitz non-polar) and a semi-empirical equation (Chrastil model) They obtained generalized parameters for the Peng–Robinson equation of state for each system These parameters were independent of temperature, and they reproduce successfully all data available in the literature The results showed that the solubility of limonene in supercritical carbon dioxide was higher than the solubility of linalool On the other hand, Sovova et al [32] investigated fatty oil influence on the solubility of limonene in CO2 under pressures 8–12 MPa at 313.2 K They measured solubility in CO2 using the dynamic method both for limonene and for the mixture of limonene and blackcurrant seed oil The concentration of fatty oil in the vapor phase was found to be negligible in comparison with the concentration of limonene Limonene was distributed between the liquid phase, rich in fatty oil, and the vapor phase, rich in CO2 , and its equilibrium concentration in the latter decreases with the diminishing limonene-to-oil ratio in the saturator Also, there was a steep increase of the limonene partition coefficient with pressure between and 10 MPa, near the critical pressure of the binary mixture of limonene and CO2 Their applied thermodynamic model was the Soave–Redlich–Kwong cubic equation of state with either the one fluid linear van der Waals mixing rule or with the MHV2 mixing rule Extraction pressures should be approximately 20% larger than the critical pressure of the essential oil + CO2 binary mixture and rather tight packing of the ground seed in the bed should be applied Catchpole and Proells [33] measured the solubilities of lipids typically found in marine oils and seed oil refining byproducts in subcritical R134a to determine whether R134a could be a viable, low-pressure alternative to supercritical CO2 The measured solubilities of squalene, oleic acid, soya oil, and deep sea shark liver oil in subcritical R134a in a countercurrent packed column apparatus over the temperature range 303–353 K at 60 bar Solubility measurements were also made over the pressure range 40–200 bar at 343 K for shark liver oil and oleic acid Their results indicated that, the solubilities of all solutes in R134a are low, ranging from 0.8 to 10 g solute/kg solvent The solubilities increased almost linearly with increasing temperature at fixed pressure and increased logarithmically with increasing pressure at fixed temperature The recorded strong temperature dependence of the solubility allows for two-stage fractionation of extracts They used a linear solvation energy relationship approach to correlate the enhancement factors of the solutes as a function of the solvent polarity/dipolarizability factor and obtained linear relationships However, the dependence of the enhancement factor on other solute–solvent parameters could not be determined 6 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Table Spline parameters for the assays performed at 66.7 bar and 288.15 K, used to choose the adequate flow rate [34] Q (×105 kg/s) tCER per 60 s MCER (×108 kg/s) YCER (× 103 kg/kg) YieldCER (%) YieldTOTAL (%) 0.62 1.08 1.50 2.15 3.16 159 93 106 75 60 6.67 2.05 2.92 3.05 3.77 10.8 18.9 19.5 18.3 11.9 0.71 1.03 1.64 1.45 1.44 1.47 1.73 3.11 2.24 1.96 Fig The effect of solvent flow rate on the mass ratio of solute in the fluid phase at the bed outlet at T = 288.15 K and P = 66.7 bar (From ref [34] with permission.) Sousa et al [34] measured the solubility of the essential oil present in L sidoides Cham in liquid CO2 by the dynamic method, in which the solvent is saturated by the solute as it flows through the bed of solids at a predetermined constant flow rate Table reports their results and Fig shows the effect of solute in the supercritical phase at the measuring cell outlet Their finding showed that for experiments accomplished at extremely low solvent flow rates, the effects of axial dispersion were important, resulting in a smaller value of YCER (the concentration of the oil in the supercritical phase at the outlet of the column) At high solvent flow rates, smaller values of YCER were obtained due to shorter residence times As shown in Fig 3, solvent flow rates in the vicinity of 1.5 × 10−5 kg/s is appropriate to be used for measuring solubility, using the dynamic method for the L sidoides + CO2 system The variation in the thermophysical properties of the solute and the solvent was relatively small due to the narrow interval of both temperature and pressure in this study Based on this, the solubility for the L sidoides + CO2 system was measured at solvent flow rates in the vicinity of 1.5 × 10−5 kg/s Table shows the measured solubility In this table, the amount of solute collected up to the end of the CER period (YieldCER ) along with the total amount of solute collected for the entire measuring time (YieldTOTAL ) are shown The effect of temperature on solubility is complex, due to the combination of two variables, density and vapor pressure The vapor pressure of the solute increases with temperature, causing an elevation in solubility However, decreasing of solvent density may cause decreases of the solute solubility The dominant effect will depend on the magnitude of each effect on the others for each system Higher solubility values were obtained at 66.7 bar in the range of 288.15–293.15 K Therefore, the increase in the solubility for this range of temperatures should mainly be due to the increase in the vapor pressure of the solute For essential oils the vapor pressure is low, however, small changes in temperature can cause significant changes in solubility For example, a 5◦ increase in the temperature (from 288.015 to 293.015 K) at 66.7 bar caused an increase in solubility of 14% However, the same increase in temperature, but from 293.15 to 298.15 K, resulted in a reduction in solute solubility of 42% Therefore, in the first case, the dominant effect was solute vapor pressure, while in the second it was density The effects of both temperature and pressure on solute solubility can be shown in Fig The overall saturation curves at 66.7 bar and various Fig The influence of temperature on the overall saturation curves at 66.7 bar (From ref [34] with permission.) Table Solubility measured by the dynamic method for the pseudo-ternary system [34] T (K) P (bar) Q (×10−5 kg/s) Y* (×10−3 kg/kg) YieldCER (%) YieldTOTAL (%) 283.15 288.15 288.15 293.15 293.15 295.65 298.15 66.7 66.7 78.5 66.7 78.5 66.7 66.7 1.60 1.50 1.60 1.53 1.60 1.53 1.57 13.4 19.5 17.8 22.7 20.1 19.0 13.2 1.34 1.64 1.63 1.76 1.45 1.55 1.16 2.19 3.11 3.03 3.21 2.75 3.29 2.82 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 temperatures (Fig 4) reflect, as expected, the behavior of solubility, since the slope of each overall saturation curve is directly proportional to solubility Also, investigations on the effect of pressure on solubility for temperatures of 288.15 and 293.15 K showed that, solubility decreases by enhancing pressure from 66.7 to 78.5 bar The behavior is consistent with the literature on solid–fluid equilibria [35] From the industrial point of view, optimization of the processes that employ supercritical fluids not only requires the knowledge of thermodynamic parameters (solubility and selectivity) but also requires consideration of mass-transfer rate [36,37] Ferreira and Meireles [38] modeled the mass-transfer mechanism for extraction of black pepper essential oil using supercritical carbon dioxide by the extended Lack’s plug flow model developed by Sovova (Sovova’s model) [39] Their procedure was used to evaluate the parameters of the Sovova’s model from experimental data—it quantitatively described the experimental data for the majority of conditions analyzed Their finding indicated that temperature and pressure levels used in their work did not change mass-transfer coefficient values and is a function of the bed characteristic and initial amounts of solute Germain et al [40] showed that, the external masstransfer coefficients (kf ) during supercritical fluid extraction of high-solubility solutes, under solvent up-flow conditions and low superficial velocities, can be small because of the negative influence of natural convection phenomena They used a shrinking-core model for mass transfer to estimate best-fit values of kf for data on SFE of lipids from pre-pressed rape seeds Values of kf at a high Reynolds number (Re = 14.1) were similar when using solvent up-flow or down-flow, but kf at lower Re (1.57) was 3.6 times smaller when using solvent up-flow than that predicted from a literature correlation for down-flow conditions These kf values are consistent with values estimated by fitting literature data, or gathered from various sources under similar, non-adequate conditions (solvent up-flow under low Re) for the extraction of both fatty and essential oils Effect of matrix on supercritical fluid extraction Different factors such as the particle size, shape, surface area, porosity, moisture, level of extractable solutes and the nature of the matrix will affect the supercritical fluid extraction results Similarly, the interactions between solutes and active sites of the matrix can necessitate strict extraction conditions The success of a SFE method not only depends on the extraction step itself (i.e nature of the supercritical fluid and choice of extraction parameters) but also on the matrix considered (a pretreatment may be recommended) as well as on the analyte trapping system [41–43] Consequently, SFE must be regarded as a four-stage process: (1) desorption of the compound from the matrix with (2) subsequent diffusion into the matrix, (3) solubilization of the analyte by the supercritical fluid, and (4) sweeping out of the extraction cell by the fluid Each part of the process has to be carefully optimized in order to obtain quantitative and reproducible recoveries Most of the time, the first step remains most difficult to control, as solute–matrix interactions are difficult to predict The physical structure of the matrix is of critical importance, as the extraction efficiency is related to the ability of the supercritical fluid to diffuse within the matrix For that reason, the extraction conditions of the same group of oils may differ from one matrix to another As a general rule, decreasing the particle size of solid matrices leads to a higher surface area, making extraction more efficient Yet, excessive grinding may hinder the extraction due to readsorption of the analytes onto matrix surfaces (this could be avoided by increasing the flow rate) and pressure drop inside the extraction chamber On the other hand, environmental agents may affect the composition and essential oil contents of matrix Esmelindro et al [44] assess the influence of light intensity (plants exposed to direct sun and in controlled lighting conditions), and the age of leaves (6–24 months) on the characteristics of the extracts of mate tea leaves obtained from carbon dioxide at high pressures Samples of mate were collected in an experiment conducted under agronomic control at Industria e Comercio de Erva-Mate Bar˜ao, Brazil Quantitative analysis of caffeine, theobromine, phytol, vitamin E, squalene, and stigmasterol was performed, and the results showed that field variables exert a strong influence on the liquid yield and on the chemical distribution of the extracts Effect of extraction parameters 5.1 Effect of pressure and temperature Four parameters are extremely helpful in the understanding of solute behavior in supercritical media, and thus in executing successful analytical supercritical fluid extractions [45,46]: (i) the miscibility or threshold pressure [47,48], which corresponds to the pressure at which the solute partitions into the supercritical fluid, (ii) the pressure at which the solute reaches its maximum solubility, (iii) the fractionation pressure range, which is the pressure region between the miscibility and solubility maximum pressures (in this interval it is possible to extract selectivity one solute by choosing the correct pressure) and (iv) a knowledge of the physical properties of the solute, particularly its melting point (in fact most solutes dissolve better when they are in their liquid state, i.e above their melting point) To illustrate the difference between the threshold pressure and the solubility maximum pressure, the solubility–pressure curve of naphthalene is given in Fig [46] This solute is slightly soluble in CO2 at 75 bar (threshold pressure) as the pressure increases the solubility rises, especially around 90 bar, up to its maximum value The fluid pressure is the main parameter that influences the extraction efficiency An elevation of this pressure at a given temperature results in an increase in the fluid density (Fig 6), which means an enhanced solubility of the solutes Consequently, the higher the extraction pressure, the smaller is the volume of fluid necessary for a given extraction For example, one needs to double the volume of CO2 in order to extract 70% of diuron herbicide from a contaminated soil when working at 110 bar instead of 338 bar (Fig 7) [49] However, high pressure is not always recommended for complex matrices owing to the higher solubility of solutes when the pressure is elevated, resulting in complex S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Fig Variation of the extraction yield of diuron from a polluted soil vs the volume of CO2 percolated at different pressures (( Fig Variation of the solubility of naphthalene with the pressure of supercritical CO2 at 45 ◦ C (From ref [46] with permission.) extracts and difficult analysis On the other hand, it must be borne in mind that the presence of co-extracted solutes can dramatically change the solubility level of the solute of interest At a constant pressure, the density of CO2 decreases when the temperature is increased This effect becomes more pronounced as the fluid compressibility increases, as shown in Fig Temperature also affects the volatility of the solute Hence, the effect of a temperature elevation is difficult to predict because of its dependence on the nature of the sample For a non-volatile solute, a higher temperature would result in lower extraction recovery owing to a decrease in solubility thus the distribution coefficient of phenol between water and supercritical CO2 decreases when the fluid temperature rises from 25 to 30 ◦ C [50] On the other hand, for a volatile solute, there is a competition between its solubility in CO2 (which decreases as the temperature increases) Fig Pressure–density diagram for carbon dioxide The shaded area corresponds to the experimental domain of supercritical phase extraction and chromatography (From ref [46] with permission.) ) 110, () 235, (䊉) 338 bar) Extraction conditions: extractant, CO2 –CH3 CN (90:10, v/v), extraction cell, 25 cm × 4.6 mm I.D.; temperature, 100 ◦ C; flow rate of liquid CO2 , 16.5 ml/min (From ref [49] with permission.) and its volatility (which rises with increasing temperature) For example, when the temperature increases from 80 to 120 ◦ C, the extraction recovery of diuron from soil with methanol-modified CO2 is enhanced from 75 to 99% [51] Baysal and Starmens [52] studies supercritical carbon dioxide of carvone and limonene from caraway seed Their results showed that pressure and temperature have main effects on the extraction efficiency They showed that at moderate temperatures just above the critical temperature of CO2 (31.1 ◦ C), extraction yield for limonene is considerable at pressures just above the critical pressure of CO2 (73.8 bar) Below this value, hardly any limonene is obtained from the seed matrix, regardless of the temperature applied during the extraction procedure At elevated temperatures, a pressure of up to 125 bar is required to extract limonene in only small quantities Further increasing of the pressure yield resembles those found for limonene Careri et al [53] used a chemometric approach to investigate the effects of different parameters on the supercritical extraction of carotenoids from spirulina Pacifica algae Their results showed that the temperature of the supercritical fluid did not influence extraction efficiency However, the pressure of the supercritical fluid plays an important role in the SFE of carotenoids from Spirulina pacifica algae Reverchon et al [54] extracted volatile oil from rose concrete, using different pressures and temperatures of supercritical fluids The results showed at the highest extraction densities (for example 100 bar and 40 ◦ C) that large quantities of paraffins and steroptens were co-extracted with the rose volatile oil Therefore, lower pressure and temperature were used for SFE of the rose concrete Hamburger et al [55] studied the effect of supercritical pressure on yield of extracted substances from three medicinal plants (marigold, hawthorn and chamomile) They reported that, at pressures above 300 bar, the yields of total extract increased Increased yield of non-volatile lipophilic compounds, such as faradiol esters, are achieved at pressure above 300 bar Brachet et al [56] used central composite designs in the study of three tropane alkaloids: hyoscyamine and scopolamine from Datura S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Fig All-trans-lycopene and cis-lycopene contents (expressed as percentage of the total lycopene) in the tomato supercritical fluid extracts performed in the 0.55–0.90 g/ml density range Error bars are indicated at each density of CO2 used (From ref [57] with permission.) arborea hairy roots and cocaine from coca leaves They showed that for cocaine extraction, pressure had a weak influence on cocaine recovery The influence of pressure and temperature on extraction yields is fairly identical for hyoscyamine and scopolamine At higher pressure, increasing the extraction temperature from 40 to 100 ◦ C yielded significant increases in hyoscyamine and scopolamine recoveries, despite the CO2 density decrease At low pressure, the effect of temperature is less pronounced These results confirm that SFE is governed by the volatility of analytes as well as by fluid density Gmez-Prieto et al [57] proposed a procedure for the SFE of all-trans-lycopene from tomato using carbon dioxide at 40 ◦ C without modifier Their method minimizes the risk of degradation via isomerization and oxidation of health-promoting ingredients, such as lycopene The effects of different experimental variables on the solvating power of the supercritical fluid were evaluated in terms of both the selectivity for and the yield of the extraction of all-trans-lycopene Fig gives the relative amounts of the cis and all-trans forms of lycopene obtained at the different densities investigated in the range over which lycopene was extracted (i.e 0.55–0.90 g/ml) The amount of the translycopene extracted increases (and the corresponding cis form content decreases) if the extraction pressure increases From these results, it seems clear that the enhancement of the fractionation of trans-lycopene requires a proper choice of CO2 density An extract 88% all-trans-lycopene and 12% cis-lycopene could be produced Ambrosino et al [58] applied a new supercritical extraction methodology to extract azadirachtin A (AZA-A) from neem seed kernels They used supercritical and liquid carbon dioxide (CO2 ) as extractive agents in a three-separation-stage supercritical pilot plant They carried out comparisons by calculating the efficiency of the pilot plant with respect to the milligrams per kilogram of seeds (ms /m0 ) of extracted AZA-A Conventional pressure extraction on raw seeds (320 MPa) led to a low yield in oil (8%) together with the lowest concentration of AZA-A per kilogram of oil and also to the lowest concentration of AZA-A per kilogram of seeds (44 mg/kg of seeds) Compared to conventional pressure extraction, both supercritical and subcritical pressures gave rise to a greater enrichment of AZA-A Significant differences were evident by comparison of supercritical Fig Overall extraction curve for marigold extraction with near-critical CO2 at various conditions (From ref [59] with permission.) and subcritical extractions at ms /m0 = 64 and 119 mg/kg of seeds The most convenient extraction was gained using an ms /m0 ratio of 119 rather than 64 For supercritical extraction, a separation of cuticular waxes from oil was performed in the pilot plant In 2005, Campos et al [59] investigated extraction of marigold (Calendula officinalis) oleoresin with supercritical carbon dioxide They showed that, the temperature effect in the process yield is complex due to the combined effect of solvent density and solute vapor pressure These opposite effects must be evaluated to observe crossing of the yield isotherms In Fig 9, a constant yield at 15 MPa with increasing temperature (from 303 to 313 K) is observed, while at 20 MPa the extract yield increases with temperature going from 299 to 313 K This behavior is an indication that the solute vapor pressure is the dominant effect at 20 MPa, while at 15 MPa the difficulty to observe the dominant effect is a suggestion that the crossover region (yield isotherms) is close to 15 MPa, for the temperatures studied Gaspar [60] studied the effect of the extraction pressure and temperature on the extraction of essential oils and other co-extracted components (cuticular waxes) from oregano (Origanum virens L.) bracts by compressed CO2 from 50 to 300 bar and 300 to 320 K, respectively Moderate conditions, using solvent densities between 0.3 and 0.5 kg m−3 , were found to be sufficient for efficient extraction of essential oils The use of higher pressures and temperatures, despite slight advantages for the rate of extraction and yields of essential oils extraction, led to significant co-extraction of waxes and, consequently, to extracts with lower essential oil content For CO2 densities below 0.25 kg m−3 , selective extraction of individual essential oils was attained At these low-density conditions, the lighter and more volatile hydrocarbons were preferentially extracted Canela et al [61] studied the supercritical fluid extraction of fatty acids and carotenoids from the Microalgae spirulina Maxima with carbon dioxide, assessing the effect of pressure and temperature on the yield and chemical composition of the extracts Their experiments were conducted at temperatures of 20–70 ◦ C and pressures of 15–180 bar Statistical analysis showed that neither the temperature nor the pressure significantly affected the total yield, but both the temperature and the pressure affected the extraction rate, and temperature was more significant than pressure upon the SFE The extracts were rich in essential fatty acids and carotenes, and at 100 bar and 45 ◦ C the 10 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 extract contained no carotenes Also, temperatures larger than 50 ◦ C degraded the carotenes, as expected 5.2 Effect of modifiers on supercritical fluid extraction The nature of the modifier depends on the nature of the solute to be extracted [62] For example, the extraction of diuron is considerably enhanced with methanol instead of acetonitrile as a modifier [63] A reasonable starting point consists of selecting a modifier that is a good solvent in its liquid state for the target analyte It should be noted that the addition of large amounts of modifier will change the critical parameters of the mixture [64], as shown in Fig 10 for methanol–carbon dioxide mixtures [64] As a result, binary mixtures of carbon dioxide and an organic solvent are often used in a subcritical state, where the diffusion coefficients are smaller than in a supercritical state Modifiers can be introduced as mixed fluids in the pumping system with a second pump and mixing chamber [65], or by simply injecting the modifier as a liquid into sample before extraction [66] (the later way being less successful because it leads to concentration gradients within the matrix) Alternatively, one may use directly Fig 10 Variations of the critical pressure and temperature of CO2 –CH3 OH mixtures with the molar fraction of methanol (From ref [64] with permission.) a cylinder tank of modified CO2 , but this is much more expensive and, as the tank is emptied, the content of modifier tends to increase Kohler et al [67] investigate the effects of modifier concentration and its nature on the supercritical carbon dioxide extraction of artemisinin and artemisinic acid from Artemisia annua L They selected methanol, ethanol, methanol–water (50/50, v/v) and toluene as modifiers based on preliminary tests and of their solvating power for artemisinin Methanol, ethanol and toluene gave similar results, however, the use of toluene presents a major drawback due to its high boiling point and concomitant longer evaporation times Methanol–water gave unsuccessful results based on the slopes of extraction kinetic curves This result can be explained by the low solvating power of water for artemisinin In order to improving supercritical CO2 extractability of hyoscyamine and scopolamine hydrochloride salts, Choi et al [68] investigate the effect of methanol and water on the extraction yields It was found that addition of methanol drastically increased the extraction yield of hyoscyamine and scopolamine However, water did not show any significant influence on the extractabilities of hyoscyamin, although it slightly increased in the case of scopolamine The poorer result for water relative to methanol may be due to the fact that water could not sufficiently improve the polarity of CO2 as much as methanol, since only 0.3% (v/v) of water can be completely miscible with CO2 They also studied the effects of basified modifiers (diethylamine) added to methanol and water on analyte extractability The addition of 10% diethylamine to methanol or water dramatically enhanced the extraction of hyoscyamine and scopolamine hydrochloride compared to using pure methanol or water This result may be due to the fact that the salts are changed to free bases by minor addition of methanol or with the diethylamine, allowing supercritical CO2 to more easily dissolve the free bases Ethanol, water and an equimolar mixture of these two solvents were chosen as co-solvents for supercritical carbon dioxide extraction of stevia glycosides by Pasquel et al [69] The results indicated that due to the high polarity of water in comparison with ethanol, an increase in glycosides solubility resulted Comparison of the yields for the experiments using water as co-solvent showed that regardless of whether it was used with or without ethanol, water increased the solubility of the glycosides Monteiro et al [70] investigated the CO2 extraction of bacuri fruit They showed that the co-solvent influenced the extraction yield of soluble material from shells of the fruit in two different ways First, owing to its polarity, the co-solvent favored dissolution of the polar substances present in the bacuri shells Second, the co-solvent diluted the extract, diminishing its viscosity, thereby enhancing the flow of the extract through the extractor Cocero and Garcia [71] studied supercritical fluid extraction of sunflower oil with carbon dioxide in a pilot plant at 30.0 MPa and 40 ◦ C, using different amount of methanol, ethanol, butanol and hexanol as co-solvent Comparing the co-solvent extraction experimental data with SFE using neat CO2 , it was found that the use of 10% (w/w) of a co-solvent increases oil solubility 10fold They used a mathematical model to investigate the effects of co-solvents on two adjustable parameters, i.e equilibrium S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 coefficient and mass-transfer parameter Their results showed that equilibrium parameter increases with co-solvent concentration Butanol gives the higher values which are exactly the same result obtained comparing oil solubility Solute recovery values using methanol values are higher than when using hexanol and ethanol This is explained by comparing the mass-transfer values, methanol values being the lowest Also, they showed that mass transfer increases with co-solvent concentration depending on the nature of co-solvent This variation is probably caused by the change of physical properties in the supercritical mixture Kerrola and Kallio [72] investigated the effect of adding various quantities of water and a mixture of ethanol and water (1:1, v/v) to the carbon dioxide extraction system on the relative amounts of the volatile components of angelica roots Their results showed that the yields of recovered components increased when an aliquot of water was added to the system After the extraction had completed and carbon dioxide released, the water co-solvent formed a separate phase The increase in volume was considered as a positive factor The co-solvent was suggested to act as the primary solvent, thus enhancing the diffusivity within the matrix The mixture of ethanol and water appeared to increase the proportion of monoterpene hydrocarbons when compared with water alone used as modifier and pure liquid CO2 without a modifier The effect of the co-solvent added on extractability of a compound varied considerably ␤- 11 Phellandrene was the most prominent compound in all extracts obtained with the mixture of ethanol and water as modifier, but the relative abundance varied from 9.2 to 30.7% depending on extraction conditions The relative proportion of both sesquiterpene hydrocarbons and oxygenated sesquiterpenes decreased when the polarity of the solvent was increased by the mixture of ethanol and water in comparison to liquid CO2 extracts obtained without any modifier Palma et al [73] extracted white grape seeds by sequential supercritical fluid extraction Their results showed that, modifier has a main effect on the SFE process and by increasing the polarity of the supercritical fluid using methanol as a modifier of CO2 , it was possible to fractionate the extracted compounds They obtained two fractions: the first, which was obtained with pure CO2 , contained mainly fatty acids, aliphatic aldehydes, and sterols The second fraction, obtained with methanol-modified CO2 , had phenolic compounds, mainly catechin, epicatechin, and gallic acid Pourmortazavi et al [74] studied the supercritical fluid extraction of aerial parts of Perovskia atriplicifolia Benth In this research, the effect of different modifiers at a constant pressure (100 atm) and temperature (35 ◦ C) on the extraction efficiency was also evaluated Fig 11 shows the effect of different volumes of modifiers on the main contents of P atriplicifolia essential oil Fig 11a shows the effect of methanol (1 and 5%, v/v) on the Fig 11 Effects of different volume of modifiers [(a) methanol, (b) ethanol, (c) dichloromethane and (d) n-hexane] on the main contents of Perovskia atriplicifolia Benth essential oil these main compounds are: (1) ␣-pinene; (2) ␦-3-carene; (3) 1,8-cineole + limonene; (4) camphor; (5) ␤-caryophyllene; (6) ␣-humulene; (7) unknown compounds with large retention number (the extraction pressure was 100 atm and extraction temperature was 35 ◦ C) (From ref [74] with permission.) 12 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 composition of P atriplicifolia essential oil Methanol decreased the number of extracted compounds in comparison with the extraction by pure supercritical CO2 from 22 compounds to 11 compounds Also, using 1% methanol as modifier increase content of ␣-pinene, ␤-pinene, ␤-caryophyllene, ␦-3-carene, ␣humulene and decrease the percent of 1,8-cineole + limonene and camphor in the extracted essential oil Methanol increased the percent of co-extracted compounds with large retention number (unknown compounds) in the oil composition Ethanol was also tested for extraction of essential oil from P atriplicifolia Results showed that addition of ethanol (Fig 11b) enhanced the concentration of ␣-pinene, camphene, ␤-pinene, myrcene, ␦-3-carene, and decrease the content of 1,8-cineole + limonene, camphor and ␤-caryophyllene in the plant oil However, these results showed that addition of %1 ethanol, as modifier was more effective than 5% ethanol Dichloromethane as modifier (Fig 11c) decreased the number of extracted compounds from 22 with pure supercritical CO2 to and compounds in the presence of and 5% (v/v) dichloromethane, respectively Dichloromethane also increased the concentration of ␣-pinene, ␤-pinene, ␦-3-carene, 1,8-cineole + limonene, ␣-humulene and reduced the concentration of camphor in the P atriplicifplia essential oil composition However, dichloromethane increased the content of co-extracted compounds with large retention indices (unknown compounds) in the essential oil composition Furthermore, it was found that dichloromethane was a selective modifier for the extraction of 1,8-cineole + limonene and ␦-3-carene from P atriplicifolia By using hexane as modifier (Fig 11d), the number of identified compounds decreased from 22 to 12 and compounds (hexane volumes were and 5%, v/v, respectively) In the presence of hexane as modifier the percent of ␣-pinene, ␤-pinene, ␦-3-carene, ␣-terpenyl acetate increased and the percent of 1,8-cineole + limonene and camphor decreased in comparison with extraction by pure carbon dioxide The result of these studies showed that changing modifier type and identity could significantly affect the selectivity of the extraction process 5.3 Effect of extraction time It is important to maximize the contact of the supercritical fluid solvent with the sample material in order to enhance the efficiency of SFE Several variables that influence the solvent contact with sample material include flow rate, SFE time, and SFE mode (static with no follow-through or dynamic with follow-through) It was reported [75] that, a 10–20 static extraction prior to dynamic extraction improved the extract recoveries in SFE extraction of aflatoxins Cui and Ang [76] developed a small-scale supercritical fluid extraction system for the selective extraction of phloroglucinols from St John’s Wort (SJW) leaf/flower mixtures using carbon dioxide The extraction efficiency was investigated as influenced by pressure, temperature, time, and modifier They optimized condition of SFE at 367 bar and 50 ◦ C Samples were held in static extraction for 10 min, followed by a dynamic extraction for 90 at the flow rate of ml/min In this study, they showed static extraction longer than 10 did not increase extraction efficiency At the Table Effect of dynamic extraction time on the composition of fennel oils obtained by SFE (at static time of 25 and pressure 200 and 350 bar) [78] Pressure (bar) Temperature (◦ C) Dynamic time (min) 200 45 30 200 45 45 350 45 30 350 45 45 Compound Run1 Run2 Run3 Run4 ␣-Pinene Sabinene Limonene (z)-␤-Ocimene ␥-Terpinene Fenchone Estragole (E)-Anethol Germacrene D 1.32 0.89 9.34 1.39 0.77 9.25 2.94 72.30 1.78 0.89 – 7.94 0.79 0.83 8.45 3.07 76.64 1.38 1.05 – 7.26 1.01 – 9.93 3.09 77.67 – 1.00 – 7.17 – – 8.36 2.81 72.70 1.61 flow rate ml/min (60 ◦ C and at 367 bar), about 83.2% of the total extractable hyperforin (HF) and 88.3% extractable adhyperforin (AHF) were extracted in the first hour, and about 95.3% extractable HF and 97.2% extractable AHF were extracted in the first 1.5 h No more than 4% HF and AHF were extracted within another hour after the 1.5 h dynamic extraction Thus, they used 10 static extraction followed by 1.5-h dynamic extraction for all samples Pourmortazavi et al [77] showed the influence of the dynamic extraction time on the composition of the essential oil of Juniperus communis L leaves that was studied by performing extraction with supercritical carbon dioxide this consists of a static extraction of 25 min, followed by 20 and 30 of dynamic extraction time Results showed that increasing dynamic extraction time at constant pressure 350 atm, enhances the content of heavy compounds with large retention indices in the plant oil They also tested a static–dynamic SFE approach for the isolation of fennel oil [78] A static extraction period (for enhancing sample–extractant contact thus favoring the attainment of the portion equilibrium) was carried out, followed by a dynamic extraction period in which extractant passed continuously through the extraction chamber The influence of the dynamic extraction time on the composition of the essential oil was studied by performing extraction with supercritical carbon dioxide consisted of a static extraction step of 25 min, followed by 30 and 45 of dynamic extraction The results are shown in Table It was found that increased dynamic extraction time enhance the extraction of most of the compounds 5.4 Effect of flow rate The speed of the supercritical fluid flowing through the cell has a strong influence on the extraction efficiencies The slower the fluid velocity, the deeper it penetrates the matrix The fluid speed can be expressed by the linear velocity, which is strongly dependent on the flow rate and the cell geometry For a given extraction cell, the flow rate can be easily changed by using a new restrictor with a different inside diameter Decreasing the flow rate resulted in a lower linear velocity S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 13 Fig 13 Effect of solvent flow rate on the extraction yield vs the specific amount of solvent (Q) at 150 bar, 45 ◦ C (From ref [80] with permission.) Fig 12 Effect of solvent flow rate on the extraction yield vs extraction time at 150 bar, 45 ◦ C (From ref [80] with permission.) and usually in increased extraction For example, 14 C-labeled linear alkylbenzenesulphonates were better extracted by supercritical CO2 modified with methanol (40 mol%) (at 380 bar and 125 ◦ C) from a sludge-amended soil with a liquid CO2 flow rate of 0.45 ml min−1 (mean recovery 90.8 ± 1.3%) instead of 1.2 ml min−1 (mean recovery 75.6 ± 1.1%), the same volume of fluid being used in each instance [79] Higher flow rates result in a decrease in the recovery either by using an elevated pressure drop though the extraction cell this phenomenon probably occurred during the extraction of diuron from a contaminated soil with a CO2 –methanol (90:10, v/v) mixture, or by increasing analyte loss during decompression of the fluid Papamichail et al [80] studied the SFE of oil from milled celery seeds using CO2 as a solvent They investigated the effect of flow rate of CO2 on the extraction rate of celery seeds They showed that the increase of the solvent flow rate leads to the increase of the amount of oil extracted versus extraction time (Fig 12) The amount of the extracted oil per kilogram of CO2 used is higher for the lower flow rate due to the intra-particle diffusion resistance This, actually, has as a result the smaller slope of the extraction curve in Fig 13 for the higher flow rate matrix effects Sample particle size and vessel packing density uniformity in day-to-day operations may be a factor in some SFE applications that achieve variable results [81] Louli et al [82] extracted parsley seed oil with supercritical carbon dioxide at different conditions Fig 14 presents their results about the effect of particle size on the extraction rate As it was expected, the extraction rate increases by decreasing the size of the seeds This is due to the higher amount of oil released as the seed cells are destroyed by milling Moreover, after milling the diffusion paths in the solid matrix become shorter resulting in a smaller intra-particle resistance to solute diffusion Sabio et al [83] extracted tomato skins and their mixtures with seeds by supercritical CO2 extraction The results of their experimental study at a pressure of 300 bar and temperature of 60 ◦ C (and a flow rate of 0.792 kg/h) for two different particle sizes are shown in Fig 15 From this figure, it is evident that the smaller particle size reduces the yield obtained in the extractions Because of their small specific surface area, large 5.5 Sample particle size and packing density In general, decreasing particle size in SFE creates more surface area and benefits extraction, but it also may hinder extraction if the analytes re-adsorb on matrix surfaces Hawthorne and coworker [81] discussed elution of the analytes from the matrix in the vessel, to determine the rate-limiting step in SFE A higher flow rate can help reduce partitioning back onto matrix sites if this is the limiting factor (otherwise, solubility factors are limiting) Larger particles, decreased packing density, smaller sample size, and a wider extraction vessel reduce potential Fig 14 Effect of particle size of Parsley seed samples on the extraction rate (data are presented as the extraction yield vs the specific amount of solvent (Q) at 10 MPa, 318 K, and a solvent flow rate of 1.1 kg CO2 /h) (From ref [82] with permission.) 14 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Fig 15 Effect of tomato sample particle size on the yields of supercritical fluid extraction (yields of lipids, lycopene, and carotenes obtained at 300 bar, 60 ◦ C, and the CO2 flow rate of 0.792 kg/h from mixtures of tomato skins and seeds with two different particle sizes) (From ref [83] with permission.) particles lead to a distinct, diffusion-dominated extraction and long processing times For example, decreasing the particle size of ground peanuts from a range of 3.35–4.75 mm to a range of 0.86–1.19 mm was found to increase the total oil recovery from 36 to 82% However, particle sizes that are too small can result in inhomogeneous extractions due to fluid channeling effects in the fixed bed In this study Sabio et al verified an inhomogeneous color distribution in the 80 ␮m particle size solid after the extraction via post-extraction examination of the substrate, indicating an uneven extraction due to the small particle size 5.6 Effect of water in supercritical fluid extraction Water in the sample often affects SFE There have been applications of direct SFE of aqueous samples [84], but precautions must be taken to avoid damaging or destructive interaction of sample contents with water; the water must be removed or controlled before performing SFE Water can aid in the extraction process, or be detrimental, depending on water can open pores, swell the matrix, thereby allowing the fluid better access to analytes, and aid in flow through the matrix Also, even though water is only ≈0.3% soluble in supercritical CO2 [84], it serves to increase the polarity of the fluid and enable higher recoveries of relatively polar species However, if excess water remains in the vessel, a highly water soluble analyte will prefer to partition into the aqueous phase and its SFE recovery will be low Semipolar analytes will dissolve in the aqueous phase, but readily partition into the supercritical CO2 , and yield high recoveries For analytes that are insoluble in water, the analytes precipitate onto matrix surfaces, and even though the analyte may be highly soluble in the extraction fluid, the excess water in the sample acts as a barrier in transfer of the analyte to the fluid [84] The solubility of water in CO2 (0.3%) can cause restrictor plugging upon the fluid depressurization, including the pressure of water in the collection system Removal of water is usually done by freeze-drying the sample matrix, as oven drying may result in solute volatilization Alternatively, addition of drying agents to the sample may be used This sample treatment is attractive as it favors the dispersion of the analytes in the matrix and the sample homogenization [85] Leeke et al [86] extracted Origanum vulgare L using supercritical carbon dioxide at 100 bar and 313 K in the presence of water They observed that the addition of water as a modifier resulted in an increase in the extraction of essential oil For extraction where water was added discontinuously, a large increase in the extraction of the essential oils was recorded The extraction degree increased with increasing w/w% water added to the bed, reaching an optimum at 80% (w/w) At these water concentrations, a low but finite yield of waxy material also resulted Such a result could be beneficial to attain an essential oil rich product In this study, the continuous addition of water also resulted in an increase in the degree of extraction of essential oils, but not necessary an increased yield of waxy material For tests where a higher w/w% was added, the removal of water would have had little effect on the extraction efficiency of essential oils, even at prolonged extraction times, because of its abundance Finally, they showed that their discontinuous method is sufficient to influence the extraction degree of the essential oils and on a commercial scale, further equipment for continuously humidifying supercritical carbon dioxide would be avoided, providing that the extraction process was completed within a suitable timescale 5.7 Drying effect Drying of plant samples, a major preservation process for spices, can be carried out conventionally by air-drying (with or without heat) or by freeze-drying It is obvious that drying and the drying process may have an influence on the content of aroma compounds Literature data indicate that the changes of aroma compounds during drying depend on the drying method as well as the type of the spice Ibanez and co-workers [87] proposed a two-step supercritical fluid extraction of rosemary leaves at selected conditions of pressure and temperature to divide the oleoresin into two fractions with different antioxidant activities and essential oil compositions They showed that, there are two main factors that have to be considered: the drying process, which influences the essential oil composition and, therefore extract quality, and the effect of the drying process on the plant cells Damage to plant cell walls can result in compounds being more easily extracted at the quoted SFE conditions Rosemary leaves obtained by different methods of drying have been extracted and evaluated in terms of antioxidant activity and essential oil yield and composition By analyzing the results for three samples of dried rosemary by using different procedures, it was shown that the treatment that provided the highest quantity of rosemary essential oil was freeze-drying followed by drying in oven at 45 ◦ C, and then vacuum rotary evaporation Freeze-drying is the mildest temperature treatment Therefore, less aroma loss is expected to be obtained Drying in oven at 45 ◦ C implied a higher temperature treatment than rotary evaporation (35 ◦ C), but the first method is faster than the second and is performed in the absence of light The vacuum treatment was a slow process conducted in daylight, and artifacts could easily be formed In fact, olfactory tests showed S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 15 development of a non-characteristic rosemary aroma in this study Collection of extracted analyte The development of appropriate collection system for target analytes after SFE is often ignored by new users, despite the obvious fact quantitative extraction conditions cannot be developed and evaluated unless the collection step is efficient Thus, the first job of the analyst is to optimize the collection system and determine its efficiency for the target analytes Apart from the extraction process, the single most important process in SFE is the efficient trapping of extracted material Two different approaches are commonly used for off-line SFE, liquid solvent collection and solid-phase trapping Both systems have their advantages and disadvantages with respect to ease of handling, choice of restrictor type, maximum gas flow, and compatibility with the various types of supercritical fluids, modifiers and analytes [88–90] 6.1 Solvent collection Liquid solvent collection is mechanically simple and has been the most widely used approach for natural samples Two common approaches have been used In the first approach, the end of the flow restrictor is placed directly into the collection solvent, and CO2 –analyte mixture is depressurized directly in contact with the solvent In the second approach, the CO2 –analyte effluent is first depressurized to the gas phase in a glass transfer tube before contacting the solvent For this system, efficient collection appears to depend on efficient transfer of the analytes from the gas phase to the collection solvent Somewhat surprisingly, the first approach has been shown to yield better collection efficiencies of more volatile components, and efficient collection using the second approach has often required the addition of a second solid-phase trap [91,92] or a glass wool insert in the glass liner [93] Various articles in the literature report the use of “liquid trapping.” One version of liquid trapping involves immersion of the restrictor into a liquid, as illustrated in Fig 16, while a second version concerns an inert solid surface in tandem with a liquid trap In the Dionex (Sunnyvale, CA, USA) 703 instrumentation for example, non-volatile analytes are thought to deposit on a hang-down tube (solid surface) while the more volatile analytes partition into the collection liquid after decompression A schematic of this type of trapping device is shown in Fig 17 Fig 16 Schematic representation of the liquid trapping process involving immersion of the restrictor into a liquid solvent of a modifier to the extraction fluid caused a decrease in trapping efficiencies Eckard and Taylor [96] found that solid-phase trapping has an additional drawback in that the sorbent trap has a finite capacity They found a 50/50 mixture of Porapak Q and glass beads exhibited the highest sample capacity and was the most effective trap for a wide range of analyte types In another study, Moore and Taylor [97] found that stainless steel ball trap efficiency is greatly affected by the addition of a modifier to the supercritical fluids They found that it was necessary to raise the trap temperature in order to achieve quantitative recoveries of digitalis glycosides Solid-phase trapping is normally performed by depressurizing the CO2 and the analytes prior to the trap and collecting the analytes from the gas (or aerosol) phase directly onto sorbents such as silica gel, Florisil, or bonded phase packing or onto insert surfaces such as glass or stainless steel beads After SFE, the trap is eluted with liquid solvents for subsequent analysis Both cryogenic and adsorption mechanism are active in solid-phase trapping, however, cryogenic trapping on inert materials (glass or stainless steel beads) is largely unsuccessful for analytes with even a small vapor pressure The use of sorbent phases allows adsorption to be used to increase the collection 6.2 Solid-phase collection Several different solid-phase trapping methods are used in SFE Mulcahey and Taylor [94] conducted a study with nonmodified CO2 to determine the best solid-phase trap composition and trapping conditions for a test mixture of analytes representing a wide range of polarities Mulcahey et al [95] continued their studies with a wider assortment of solid-phase sorbent traps They found that a single trap composition may not effectively trap a wide range of analytes Furthermore, the addition Fig 17 Schematic representation of the liquid collection systems used in the Dionex extractors 16 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 efficiencies, and the selectivity of the adsorption mechanism can be used to gain a degree of compound-class fractionation during the SFE collection step In addition, the choice of rinsing solvent(s) can be used to selectively elute different compound classes from the trapping system This degree of selectivity based on the elution of the trap is a particular advantage of sorbent collection over solvent collection systems However, solid-phase trapping can be more trapping material, the trap temperature, and the rinsing solvent Some articles may be found in the literature about investigation on the effect of solid-phase collection on the composition of supercritical fluid extracted plant oil Dugo et al [98] reported a method for the deterpenation of the citrus essential oils with supercritical CO2 using adsorption on silica gel to enhance the selectivity of the separation between the hydrocarbons and the oxygenated compounds They showed that the silica gel retains the oxygenated, more polar compounds, thus allowing at low temperature and pressure, the extraction of the non-polar terpene hydrocarbons By increasing the pressure and the temperature after a defined time, it is possible to elute the oxygenated components the volatile fraction Araujo et al [99] extracted Pupunha (Guilielna speciosa) oil in a fixed bed using carbon dioxide In their study, the supercritical carbon dioxide containing the solubilized oil was expanded in the small stainless steel tube the precipitated extract was collected in a glass tube placed inside the separator Blanch et al [100] in 1994 proposed a simple procedure for off-line SFE and capillary gas chromatography analysis of the essential oil obtained from Rosmarinus officinalis L During the extraction step, they deposited the obtained analytes on an internal trap where the supercritical fluid evaporates Subsequently, a suitable solvent is pumped through the trap so that the analytes are rinsed off into a vial Fig 18 shows scheme of their assembly They used different sorbent for trapping extracted essential oils, such as Hypersil octadecylsilica (ODS), glass beads, GasChrom 220, Tenax TA, Thermotrap TA, and Volaspher A-2 silanized sorbents 6.3 Collection in empty vessels Collection in an empty vial or vessel has been successfully practiced by a number of investigators and is particularly appropriate for bulk extraction of fat and similar exhaustive extractions It is also applicable however, for the extraction of trace levels of analytes, such as pesticides [101], but larger collection vessels are required for capturing such trace analytes in order to minimize their loss Avoidance of entrainment of analytes in the escaping fluid stream can be minimized by adding a glass, steel wool, or ball packing to the empty container The chosen material should be chemically inert, provide a high surface for condensing the analyte from the rapidly expanding fluid, but allow ready desorption of the analyte after completion of the extraction 6.4 Novel collection methods In 2004, Sarmento et al [102] studied the performance of three commercial reverse osmosis membranes: SG, CG and AG regarding the permeability to supercritical CO2 and the reten- Fig 18 Scheme of the trap assembly (From ref [100] with permission.) tion of lemongrass, orange and nutmeg essential oils at 12 MPa and 40 ◦ C The results of their work demonstrate the potential for use of a commercial reverse osmosis membrane in the separation of lemongrass, orange and nutmeg essential oils from supercritical mixtures with CO2 All the membranes exhibited good resistance to the severe pressure conditions employed The oil retention index was reduced with the increase in transmembrane pressure from to MPa The best retention results were obtained with the SG membrane, which retained up to 90% of all the essential oils tested However, at the same time, this membrane allowed the lowest CO2 permeate fluxes with values up to 8.75 kg h−1 m−2 at a pressure difference of MPa The experimental results indicated the occurrence of fouling for all the membrane models after permeation of lemongrass essential oil Spricigo et al [103] showed that the association of membrane separation processes to the supercritical fluid extraction of essential oils from vegetable matrices can be an alternative to the reduction of recompression costs derived from the depressurization step necessary for the recovering of the extracts In their work, a cellulose acetate reverse osmosis membrane was applied to perform the separation of nutmeg essential oil and dense carbon dioxide The effects of feed stream essential oil concentration, temperature and trans-membrane pressure on essential oil retention and CO2 permeability were investigated The average retention of essential oil by the membrane was of 96.4% and it was not affected significantly by any of the process variables The CO2 flux was linearly proportional to the trans-membrane pressure applied and decreased as the essential oil concentration S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 in the feed stream increased The membrane presented good CO2 permeability and resisted well to the severe pressure conditions applied 6.5 On-line coupling of supercritical fluid extraction with chromatographic techniques Sometimes SFE has been directly coupled to chromatographic detection systems as this offers some advantages over off-line SFE In on-line SFE the entire extracted sample is introduced into the system and therefore greater detection sensitivity can be obtained However, smaller sample sizes are often used in on-line SFE to prevent overloading of the chromatographic column There is also less chance of sample contamination because there is no intermediate handling of the extracted analytes Initial studies have focused on coupling SFE to supercritical chromatography (SFC), as carbon dioxide is used as both the extraction fluid and the chromatographic mobile phase, although much research into coupling SFE to GC has been carried out Andersen et al [104] discussed some theoretical considerations involved in using SFE as a method for sample introduction in chromatography An SFE–SFC protocol was used to extract and determine the herbicides linuron and diuron from a sandy loam soil and wheat After initial SFE optimization, carried out off-line, the two techniques were coupled with flame ionization being used for detection A modifier was required to achieve quantitative extraction, which eluted as a solvent front on the chromatogram Both capillary and microbore columns were investigated, with better sensitivity being obtained with a microbore column and larger sample loops The most common technique used in on-line SFE of natural samples is coupled SFE–GC This coupling technique is of particular important for non-polar analytes, which not require derivation either in situ or prior to GC separation A variety of approaches have been used to couple SFE extraction with capillary GC, which can be roughly categorized into three main groups: (a) An external loop, through which the extract passes, is used to introduce the analytes into the GC column The analytes are transferred into the GC column by sweeping the heated loop with carrier gas This has the advantage of allowing offline collection to be carried out simultaneously, although it also suffers from a decrease in sensitivity as only a fraction of the sample is passed to the chromatograph (b) The extract is transferred into a sorbent trap external to the gas chromatograph that may be cryogenically cooled Subsequent heating of the trap and purging with carrier gas allow the extracts to enter the chromatograph The addition of the Tenax-GC trap gave better peak shapes than direct SFE–GC because of the refocusing effect It also allowed the methanol modifier, used in the extraction, to be removed before flushing the analytes on to the column, which previously caused a “hump” in the baseline in conventional SFE–GC (c) The simplest approach utilizes a conventional GC injection port to couple SFE with GC with both on-column and split–splitless ports being used All three techniques gener- 17 ally require cryogenic refocusing of the analytes on the front of the capillary column prior to analysis to obtain good peak shapes The first time SFE was coupled to another separation technique (i.e SFC) was in extracting caffeine from roaste coffee beans [105] SFE–GC with mass spectrometry (MS) or flame ionization detection (FID) has been demonstrated for flavor compounds in spices, chewing gum, orange peel, spruce needles and cedar wood [106], from lime, lime peel, eucalyptus [107,108], basil [107,109], grapefruit oil [110,111], thyme [109,112], orange juice [108] and chamomile [109] Many of the earlier applications were of qualitative nature In comparison between SFE–GC and SFE–SFC, the latter was given an advantage due to higher yields of oxygenated terpenes and no need for derivation [109,113] Sato et al [114] used direct connection of supercritical fluid extraction and supercritical fluid chromatography as a rapid quantitative method for capsaicinoids in placentas of Capsicum annuum L They compared this method with usual extraction-HPLC method Their finding showed that the SFE/SFC method has the advantages of no need for pretreatment and no (or minimal) need for organic solvents Also, this method is useful as a rapid (20 min) and safe screening test for the pungency of various Capsicum fruits Extraction of oxygenated compounds from plant materials by supercritical fluid extraction Today, antioxidants from natural resources are associated with health benefits since oxygenated compounds are related to a positive action against heart diseases, malaria, neurodegenerative diseases, AIDS, cancer and longevity [115] On the other hand, some of the artificial antioxidants are related to health damage, such as kidney edema [116] For these reasons, the market for natural antioxidant should rapidly increase Several methods have been used to extract antioxidants from aromatic plants, such as solid–liquid extraction, aqueous alkaline extraction, extraction with aqueous solutions [117,118], and supercritical fluid extraction [119,120] Products obtained by SFE from different plants, in general, have a higher antioxidant activity than extracts obtained by using solvent extraction with organic solvents [121,122], probably due to a difference in composition deriving from the extraction conditions applied In 2004, Hu et al [123] investigated the influences of extracting pressure, temperature, and flow rate on the yield of sesame seed extract and the antioxidant activity of extracts by supercritical carbon dioxide and solvent from black sesame seed as compared to ␣-tocopherol, Trolox, and butylated hydroxytoluene (BHT) They found the highest extracted yield was achieved at 35 ◦ C, 40 MPa, and a CO2 flow rate of 205 ml min−1 Results for the linoleic acid system showed that the antioxidant activity follows the following order: extract at 35 ◦ C, 20 MPa > BHT > extract at 55 ◦ C, 40 MPa > extract at 55 ◦ C, 30 MPa > Trolox > solvent extraction > ␣-tocopherol The supercritical carbon dioxide extracts exhibited significantly higher antioxidant activities comparable to those obtained by n-hexane extraction 18 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 The volatile components of Terminalia catappa (green, yellow and red fallen) leaves were extracted using supercritical carbon dioxide at various pressures [124] The extracts from yellow and red fallen leaves exhibited higher inhibition of preoxidation that those from green leaves On the other hand, supercritical carbon dioxide extraction at 133 atm and 40 ◦ C resulted in the extracts with better antioxidant activity whereas lower inhibition of pre-oxidation was observed with extracts prepared from the extraction at 267 atm and 40 ◦ C Also, Zancan et al [125] studied the influence of the use of co-solvent in the kinetics of SFE of ginger oleoresin, in the chemical composition of the extracts and in their antioxidant action They showed that the major substances present in the ginger extracts were ␣-zingiberene, gingerols and shogaols the amounts of these compounds were significantly affected by temperature, pressure and co-solvent Nonetheless, the antioxidant activity of the ginger extracts remained constant at 80% and decreased to 60% in the absence of gingerols and shogaols In 2003, the selectivity of subcritical water extraction several temperatures to extract antioxidant compounds from rosemary leaves was investigated [126] Results indicate high selectivity of subcritical water toward the most active compounds of rosemary The antioxidant activity of fractions obtained by extraction at different water temperatures was high comparable to those achieved by SFE of rosemary leaves In this study, Ibanez et al demonstrated the possibility of tuning the selectivity for antioxidant extraction by a small change in extraction temperature Senorans et al [127] isolated antioxidants from orange juice by the use of countercurrent supercritical fluid extraction (CC-SFE) and characterized by reversed-phase liquid chromatography (RPLC) coupled to mass spectrometry (MS) and diode-array detection (DAD) They employed a pilot-scale SFE plant equipped with a packed column for countercurrent extraction and fractionation of raw orange juice with carbon dioxide Several experiments have been performed in order to study the effect of the countercurrent conditions on the content of antioxidative compounds In their study, the main variable that has been considered was the solvent-to-feed ratio (S/F) because it plays an essential role in the extraction efficiency The values tested covered a wide range of sample and solvent (CO2 ) flow rates They obtained three different products after extraction and fractionation of the orange juice: those in separators (F1) and (F2) and the raffinate (R) which is the byproduct of the extracted samples collected at the bottom of the column Fig 19 shows the percent of total area of the identified compounds versus solvent-to-feed ratio When low S/F ratios are used, a maximum extraction of flavonoids is obtained, with a low percentage recovered in the raffinate The opposite is found at S/F equal to 11 where almost 96% of the compounds identified are found in the raffinate In each experimental run, two different extracted fractions and the residual non-extracted juice were obtained and characterized Different flavonoids have been identified in the fractions obtained after CC-SFE Also, they discussed possibility of using this process for antioxidant compounds enrichment Enrichment results are shown in Fig 20 as a function of the S/F ratio data corresponding to the enrichment achieved in separators and individually and the total (consid- Fig 19 Graph representing the percent of total area of the identified compounds vs solvent-to-feed ratio (From ref [127] with permission.) Fig 20 Graph representing the log (enrichment) as a function of the S/F ratio (a selective enrichment can be observed toward the extracts or the raffinate as a function of the S/F ratios selected) (From ref [127] with permission.) ering both separators together) are also presented They obtained a high correlation (96%) for log (total enrichment) versus S/F using a linear regression y = 0.3365x + 2.4304r2 = 0.963 In 2002, Simoa et al [128] determined antioxidants from orange juice by the combined use of CC-SFE prior to RPLC or micellar electrokinetic chromatography (MEKC) They achieved separation of antioxidants found in the SFE fractions by using a new MEKC method and a published LC procedure, both using diode-array detection In the same year, Yepez et al [129] obtained fractions from seeds of coriander (Coriandrum sativum) by extraction with supercritical carbon dioxide in a semi-continuous lab-scale equipment, and were tested for their antioxidant activity Fractions from coriander seeds obtained by SFE exhibited a significant antioxidant activity, as determined by removal of DPPH free radicals larger than 50% after 300 In addition, high extraction yields, close to 2% (w/w), were achieved using SFE conditions in the range of 183–326 K, corresponding to a CO2 density in the range of 0.74 g/ml Extraction of terpenes and sesquiterpenes by supercritical fluid extraction Goto and co-workers [130] used supercritical CO2 to separate oxygenated compounds from essential oils This technique still cannot replace vacuum distillation as an industrial process S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 because of low recoveries and inconsistent results Vacuum distillation and supercritical CO2 are complementary processes for producing high quality oxygenated compounds with high recovery rates The former is more suitable for removing monoterpenes at low fraction temperatures (40%) in sesquiterpenes, while those from leaves and flowers were abundant in nitrogenated and oxygenated compounds However, SDE extracts from stems, leaves, and flowers of the plant contained sesquiterpene levels of 32, 28 and 20%, and a generally higher proportion of oxygenated compounds and monoterpenes than the SFE counterparts Also, some heavy hydrocarbons originated from flower pigments and waxes were isolated by SFE but not by SDE It was found that SFE is both selective and highly efficient in the isolation of sesquiterpenes, heavy hydrocarbons and nitrogenated compounds in this particular case Vilegas and Lancas [142] used supercritical CO2 extraction to compare chemical composition of the essential oils obtained by steam distillation procedure from two lauraceae (Ocotea caesia Mez.) Their study showed that the supercritical carbon dioxide extractions had lower yields than those from steam distillation, but the extracts obtained were similar in both species studied In 1999, Eikani et al [143] compared the extraction obtained from Cuminum cyminum L by supercritical carbon dioxide with cumin essential oil obtained by conventional steam distillation They showed that the physicochemical properties (such as refractive index, specific gravity and optical rotation) of the oils extracted by supercritical CO2 and steam distillation were different The results of the GC–MS analysis showed that the most noticeable difference between the two methods is in the pmentha-1,3-dien-7-al and p-mentha-1,4-dien-7-al composition Hydrodistillation and solvent extraction using pentane, ethanol and supercritical carbon dioxide were used to isolate essential oils from Grapefruit flavedo [144] The compositions of different extractions were compared Monoterpene hydrocarbons decrease in supercritical carbon dioxide extracts at 87–90% with respect to their quantity in pentane extracts (95%) and in hydrodistillate (97%) these levels in monoterpene hydrocarbons were related to the limonene content, the most abundant compound in grapefruit essence Sesquiterpenes, aldehydes, alcohols 21 S.M Pourmortazavi, S.S Hajimirsadeghi / J Chromatogr A 1163 (2007) 2–24 Table Comparison of the overall chromatographic area percentages for the four main classes of hydrocarbon monoterpenes (HMs), oxygenated monoterpenes (OMs), hydrocarbon sesquiterpenes (HSs), and oxygenated sesquiterpenes (OSs) (at different stages of supercritical fluid extraction) in the constituents of the Laurel oil [146] Class 60 120 180 240 SFE-T HD HMs OMs HSs OSs 12.67 79.43 7.45 0.5 2.74 77.62 14.91 4.17 1.75 76.94 14.11 6.92 0.31 66.68 15.52 17.07 13.65 70.86 11.53 3.68 15.51 70.28 7.19 7.03 Table Percent yield and cumulative percent yield of the Laurel oil at different stages of the supercritical extraction (60–240 min), the overall run (SFE-T), and the hydrodistillation [146] Quantity 60 120 180 240 SFE-T HD Yield (%) Cumulative yield (%) ms /m0 0.27 0.27 0.28 0.56 0.22 0.77 0.05 0.82 0.82 0.82 0.90 0.90 5.36 10.72 16.08 21.44 21.44 – Columns SFE-1 to SFE-4 refer to the oil fractions collected after each hour of the supercritical extraction SFE-T is the overall essential oil obtained by SFE Columns SFE-1 to SFE-4 refer to the oil fractions collected after each hour of the supercritical extraction SFE-T is the overall essential oil obtained by SFE and esters increased their GC area percentage in supercritical carbon dioxide extracts obtained at a high fluid density relative to hydrodistillate and the pentane extracts In 2000, Cassel et al [145] compared results of hydrodistillation and supercritical carbon dioxide methods for the extraction of Baccharis leave oil They observed the non-oxygenated monoterpenes, present in the hydrodistilled oil in high contents (␤-pinene 28.2% and limonene 10.6%), were not detected in the supercritical carbon dioxide extracts The selectivity of supercritical carbon dioxide allowed one to maximize the concentration of oxygenated compounds Also, Caredda et al [146] studied supercritical carbon dioxide extraction of essential oil from Laurus nobilis Extraction conditions were as follows: pressure 90 bar, temperature 50 ◦ C and carbon dioxide flow 1.0 kg/h Extracted waxes were trapped in the first separator set at 90 bar and −10 ◦ C The oil was recovered in the second separator held at 15 bar and 10 ◦ C The main components were 1,8-cineole (22.8%), linalool (12.5%), R-terpinyl acetate (11.4%), and methyleugenol (8.1%) Four classes, hydrocarbon monoterpenes (HMs), oxygenated monoterpenes (OMs), hydrocarbon sesquiterpenes (HSs), and oxygenated sesquiterpenes (OSs), on the basis of their chemical structure or retention time were reported The area percentages relative to each class are shown in Table In general, volatile compounds (HMs) are extracted almost completely during the first hour of extraction (12.67, against 0.31% in the fourth hour) The OMs decreased to a minor extent from 79.43 to 66.68% HSs and OSs are present at, respectively, 15.52 and 17.07% in the fraction obtained after 180–240 and at 7.45 and 0.50 in the first hour sample These results confirm that a long time run is necessary to obtain oil with a stable composition Comparison with the hydrodistilled oil did not reveal any significant difference The yields of each fraction of the supercritical extraction and hydrodistillation as w/w%, with respect to the charged material, are reported in Table In the same table the amount of CO2 consumed in the process, expressed as the specific mass of solvent, ms /m0 (m0 is the mass of leaves charged in the extractor) is specified The overall yield of the supercritical extraction was 0.82%; 1,8-cineole (22.84%) was the major component In 2004, Pourmortazavi et al [77] showed that different extraction compositions could be obtained by different extraction methods applied to natural products They studied supercritical fluid extraction of volatile oil from J communis L leaves using carbon dioxide was carried out under different conditions of pressure, temperature, modifier content and dynamic extraction time Then, they compared proposed extraction method with hydrodistillation A total 22 compounds have been determined in SFE extracts while in the hydrodistilled oil only 11 components were identified and quantified SFE products were found to be markedly different from the corresponding hydrodistilled oil A large amount of ␤-phellanderene was present in the hydrodistilled essential oil, also the ratio of ␣pinene and 3-carene in distilled oil were high in comparison with the supercritical carbon dioxide extracts Their results showed that under pressure 200 atm, temperature 45 ◦ C and dynamic extraction time of 30 min, SFE of limonene was more selective and under pressure 350 atm, temperature 45 ◦ C and dynamic extraction time of 20 min, extraction was more selective for the ␣-thoujone, which was not found in the hydrodistilled oil Pourmortazavi et al [147] also showed that the composition of the SFE products and the hydrodistilled black cumin essential oils is significantly different They compared the composition of the SFE product with hydrodistilled oil and found a higher level of the ␥-terpinene and cuminaldehyde in the hydrodistilled oil Table shows the p-cymene content of the distilled oils is considerable, however, this compound was not found in the SFE extracts The method contributes to the automation of pharmaceutical industry Ebrahimzadeh et al [148] isolated essential oil of Zataria multiflora Boiss, cultivated in Iran, by steam distillation and compared with supercritical fluid CO2 extracts The extracts obtained by SFE at different conditions were compositions similar to that of the oil obtained by steam distillation Table Comparison of the main components of Iranian black cumin oils obtained by SFE (under 200 atm pressure and 45 ◦ C temperature for 15 static followed by 20 dynamic) and hydrodistillation [147] Compound SFE HD ␣-Pinene ␤-Pinene Myrcene p-Cymene o-Cymene Limonene ␥-Terpinene Cuminaldehyde Cuminyl alcohol ␣-Methyl-benzenemethanol 0.8 1.5 0.6 – 7.8 6.8 38.0 11.5 – 25.6 2.8 3.7 1.0 5.6 0.1 10.6 45.7 12.7 6.4 3.5 ... extracted in the supercritical extracting fluid This can be investigated by spiking an inert medium, usually celite or sand, with the analyte of interest In addition to providing an indication... of extraction kinetic curves This result can be explained by the low solvating power of water for artemisinin In order to improving supercritical CO2 extractability of hyoscyamine and scopolamine... than in a supercritical state Modifiers can be introduced as mixed fluids in the pumping system with a second pump and mixing chamber [65], or by simply injecting the modifier as a liquid into