Modeling of the overal kinetic extraction from Maytenus aquifolia using compressed CO2

THERMODYNAMICS

Modeling of the overal kinetic extraction from Maytenus aquifolia using compressed CO₂

M. Minozzo^I; A. Popiolski^I; V. Dal Prá^II; H. Treichel^I; R. L. Cansian^I; J. Vladimir Oliveira^I; A. J. Mossi^III; M. A. Mazutti^IV,*

^IDepartment of Food Engineering, URI - Campus de Erechim, Av. Sete de Setembro 1621, CEP: 99700-000, Erechim - RS, Brazil

^IIDepartment of Pharmacy, URI - Campus de Erechim, Erechim - RS, Brazil

^IIIDepartment of Agronomy, Universidade Federal da Fronteira Sul, Av. Dom João Hoffmann, CEP: 99700-000, Erechim - RS, Brazil

^IVDepartment of Chemical Engineering, Federal University of Santa Maria, Phone: + (55) (55) 3220-9592, Av. Roraima, 1000, CEP: 97105-900, Santa Maria, - RS, Brazil. E-mail: [e-mail]

ABSTRACT

In Brazil, the species Maytenus aquifolia and Maytenus ilicifolia are widely used in popular medicine in the form of teas for stomach and ulcer illness treatment. Despite the great interest in Maytenus aquifolia therapeutic properties and the fact that it is an abundant and native plant growing in Brazil, there is a lack of information in the literature concerning the extraction at high pressures. In this context, this work is focused on the mathematical modelling of the packed-bed extraction of Maytenus aquifolia with compressed CO₂. Three mathematical models were used to represent the experimental data. The experiments were performed in a laboratory-scale unit, evaluating the effects of temperature (293 to 323 K), pressure (100 to 250 bar), and extraction time on the yield of the extracts. Results show that the extraction temperature and solvent density exerted a pronounced effect on yield. The mathematical model of Sovová was the most suitable to represent the experimental extraction data of M. aquifolia.

Keywords: High-pressure extraction; Extraction kinetics; Maytenus aquifolia; Mathematical modeling.

INTRODUCTION

The extraction of products by super/sub-critical fluids (SFE) is based on the contact between a solid raw material and a pressurized solvent, which removes the compounds of interest from the solid phase. In this technique, after removal, the extract is separated from the solvent through pressure reduction (Martínez et al., 2003). The operating temperature, pressure and solvent flow rate generally determine the solubility and the selectivity of the target components, and their diffusion capacity across the fluid (Brunner, 1994).

The use of SFE to obtain extracts from plants with pharmaceutical or food properties has received increased interest nowadays (Carvalho et al., 2012; Paviani et al., 2012; Barroso et al., 2011; Mazutti et al., 2008; Pavaiani et al., 2006; Santos et al., 2007). Among these medicinal plants are the species Maytenus aquifolia and Maytenus ilicifolia, which are widely used in popular medicine, being known as "espinheira-santa". Both species are native growing plants, with natural occurrence in South Brazil where they are used in the form of teas for stomach ulcer illness treatment. For instance, Souza-Formigori et al., (1991) have demonstrated the antiulcerogenic property of Maytenus aquifolium and Maytenus ilicifolia, which seems to be closely related to the presence of two classes of substances, namely, phenols and triterpenes (Pereira et al., 1995; Pereira et al., 2003). The antiulcerogenic activity of aqueous extracts of Maytenus ilicifolia and Maytenus aquifolium has also been confirmed in laboratory tests with rats (Oliveira et al., 1991).

The use of pressurized carbon dioxide for extraction and characterization of extracts of Maytenus ilicifolia has been previously reported (Mossi et al., 2004; Lanças et al., 1997). For example, Mossi et al. (2004) evaluated the influence of particle size, temperature, pressure and carbon dioxide flow rate on the yield and chemical profile of the extracts obtained. In the work of Lanças et al. (1997), the investigation focused only on the characterization of the triterpene fraction and no evaluation of the effects of process extraction parameters was provided.

Although Maytenus ilicifolia and Maytenus aquifolia belong to the Celestraceae family they comprise different species. The different chemotypes of M. aquifolia and M. ilicifolia could influence the yield and chemical profile of the extracts. For medicinal purposes, an interesting possibility is to obtain an extract rich in the desired compounds, like the triterpenes and phenols. As SFE already was employed successfully for extraction of M. ilicifolia, it may also be relevant to evaluate the SFE of M. aquifolia. For scale-up purposes it may be important to carry out kinetic modeling of the extraction curves to determine the proper operating conditions, making feasible SFE processes (Silva et al., 2008; Martínez et al., 2003).

The present report is part of a broader project aiming at the extraction and total characterization of some species of the Celestraceae family. Here, the goal is to investigate the influence of temperature (293 to 323 K), pressure (100 to 250 bar), and extraction time on the yield of the extracts obtained from high-pressure carbon dioxide extraction of Maytenus aquifolia. Three mathematical models available in the literature were used to represent the experimental extraction kinetic data.

EXPERIMENTAL METHODS

Material

Leaves of Maytenus aquifolia were collected in Erechim county (Rio Grande do Sul State, Brazil, 27º 37' 54") from plants growing in full sunlight. After the collection, the samples were immediately dried at room temperature during 3 days, ground and sieved, collecting the particles that passed through a sieve of 100 mesh and were retained in a sieve of 200 mesh. The samples were then stored at room temperature under nitrogen atmosphere prior to the extraction. The carbon dioxide (99.9% purity) was purchased from White Martins S.A. (Brazil).

Characterization of the Raw Material

The real density of the raw material (ρ_p) was determined by helium picnometry (Micrometrics, model Multivolume Pycnometer 1305, Norcross, EUA). The apparent bed density (ρ_ap) was determined by the ratio of mass of raw material and the volume of the extractor, whereas the bed porosity (ε) was determined according to the following equation:

The particle mean diameter was calculated by the equation presented by Gomide (1983) and also used in the work of Mezzomo et al. (2009).

Apparatus and Experimental Procedure

The experiments were performed in a laboratory scale unit, schematically presented in Figure 1, which consists basically of a CO₂ reservoir, two thermostatic baths, a syringe pump (ISCO 260D), a 0.1 dm³ jacketed extraction vessel, an absolute pressure transducer (Smar, LD301) equipped with a portable programmer (Smar, HT 201) with a precision of + 0.12 bar, a collector vessel with a glass tube and a cold trap. Amounts around 25 g of dried Maytenus aquifolia leaves were charged into the extraction vessel. The CO₂ was pumped into the bed, which was held by two 300 mesh wire disks at both ends, and was kept in contact with the herbaceous matrix for at least one hour to allow system stabilization at the same condition of the experiment. Afterwards, the essential oil was collected by opening the micrometering valve and the CO₂ mass flow was accounted for by the pump recordings. The experiments were accomplished isothermally, at constant pressure using a mass CO₂ flow rate of 2 g/min, as suggested by Mossi et al. (2004). The experimental range investigated was 293 - 323 K in temperature and from 100 to 250 bar in pressure. Triplicate extraction runs were accomplished for all conditions.

Mathematical Modelling

Many mathematical models for the extraction of substances from solid material using supercritical fluids were already presented in the literature (Sovová, 1994; França and Meirelles, 2000; Martínez et al., 2003). SFE is usually treated as a cylindrical bed of solid particles, and the pressurized fluid flows axially through it, removing the soluble compounds from the solid phase. Both phases involved in the process, solid and fluid phases, are considered to occupy all the volume of the extraction bed. The mathematical models of Crank (1975), Sovová (1994) and Martínez et al. (2003) were applied here to represent the experimental extraction data of M. aquifolia, considering the extract as a single pseudo-compound.

The data concerning the overall extraction curves were used to estimate the model parameters using software developed in Fortran 90 language, which uses the Particle Swarm Optimization (PSO) heuristic method for the parameter estimation (Schwaab et al., 2008). The objective function used was the sum of least squares. The parameters of the PSO algorithm, namely the number of particles, inertial weight, cognition and social parameters were set as: 40, 0.7, 1.0 and 1.0, respectively. These parameters were fixed based on previous works of the group concerning parameter estimation.

Crank Model

The Crank model supposes that the soluble extracts are uniformly distributed in the spherical solid particles. Based on the Fick law of diffusion, the equation that predicts the mass of extract obtained in the process is given by (Crank, 1975):

where m_ext is the mass of extract (g), m₀ is the initial solute mass (g), t is the extraction time (min), D_eff is the effective diffusivity of solute in the solid particle (m²/min), r is the particle radius and n is the number of the series expansion; in the above equation, the parameter D_eff is estimated using the extraction kinetic data.

Sovová Model

This model describes a situation when solvent flows axially through the bed of milled plant material in a cylindrical extractor. The solvent is solute-free at the entrance of the extractor, and the temperature and pressure are regarded as constants. The solid bed is homogeneous with respect to both the particle size and the initial distribution of solute. The solute is deposited in plant cells and protected by cell walls. However, a part of the walls has been broken, opened by milling, so that a part of the solute is directly exposed to the solvent (Sovová, 1994). The mass of solute contained initially in the solid phase consists of the mass of easily accessible solute. According to Sovová (1994) the extraction curves could be calculated using the extended Lack's plug-flow model:

where t is the extraction time (min), N is the initial mass of inert material (g), x₀ is the mass fraction of extract in the inert material (g/g), Y^* is the solubility of the extract in the CO₂ (g/g), q is the mass flow rate of CO₂ (g/min), t_CER and t_FER are the time related to the steps of constant extraction rate and falling extraction rate, respectively (min), and x_k is the mass fraction of easily accessible solute in the inert material (g/g), which is an estimated parameter in the model.

Parameter Z is directly proportional to the solvent-phase mass transfer coefficient and inversely proportional to the specific solvent flow rate:

where ρ e ρ_p are the density of CO₂ and of the particles (g/cm³), respectively, ε is the bed porosity and kfa is the solvent-phase mass transfer coefficient (min^-1), a fitted parameter.

Similarly, parameter W is directly proportional to the solid-phase mass transfer coefficient and inversely proportional to the specific solvent flow rate:

where k_sa is the solid-phase mass transfer coefficient (min^-1), another adjusted parameter.

Martínez Model

The model proposed by Martínez et al. (2003) neglects the accumulation and the dispersion in the fluid phase as these phenomena are less important for the process compared with the convection effect, which leads to:

where m_ext is the mass of extract (g), m₀ is the initial solute mass (g), t is the extraction time (min), and b_i and t_mi are parameters estimated from extraction kinetic data (min^-1 and min, respectively).

RESULTS AND DISCUSSION

Characteristics of the Raw Material

The dried M. aquifolia particles had a mean diameter of 1.02±0.01 x 10^-3m and a real density of 1390±43 kg/m³. The extraction bed apparent density was 292 kg/m³ and its porosity was 0.79. The values of CO₂ density at each condition of pressure and temperature, presented in Table 1, were estimated from the Angus et al. (1976) correlation. In addition, Table 1 presents other relevant information to be used in the mathematical modeling of the extraction curves, such as the mass-transfer rate of the constant extraction rate period (M_CER), the mass flow rate of CO₂ and the extract solubility (Y^*).

Extraction of Maytenus aquifolia Samples

Table 1 presents the yield along with the experimental conditions investigated in this work. Here, the yield is defined as the weight percentage of the oil extracted with respect to the initial charge of the raw material in the extractor. The general analysis of the results of Table 1 reveals that the extraction obtained in this work ranged from 0.116% (m/m) in run 3 (100 bar and 50 ºC) to 1.104% (m/m) in run 4 (250 bar and 50 ºC), which are in good agreement with the work of Mossi et al. (2004) for the SFE of M. ilicifolia.

The increase in temperature at the lowest extraction pressure (100 bar) led to a sharp decrease in the extraction yield whereas a considerable increase (two time higher) was verified at the highest pressure (250 bar) due to large solvent density changes. The increase of solvent density caused an enhancement of the solvent power, increasing the yield. However, comparing runs 2, 4, and 6 (same pressure, increasing temperature, decreasing solvent density), it is seen that the extraction yield increased with temperature.

This behavior is due to the competition between temperature (vapor pressure of essential oil) and extraction pressure (solvent power) in the range investigated, which is clearly verified in the eighth column of Table 1 (solubility of solute in CO₂). It is seen that the solubility decrease with increasing temperature at low pressure (100 bar), since the density of the solvent decreases sharply. For pressures higher than 175 bar, the change in the density with temperature is less expressive, in a manner that the increase in the vapor-pressure caused by the increase in the temperature is more important than density of the solvent. The thermodynamic solubility is dependent on temperature and pressure (solvent density) and, thus, the appreciable decrease in density observed at 100 bar from 20 to 50ºC is responsible for the slight reduction in solubility, implying a low extraction yield.

Data of Table 1 were used to compute linear (L), quadratic (Q) and interaction effects of temperature and pressure (1Lx2L) on the extraction yields, which are presented in Figure 2 in the form of a Pareto Chart. It is seen that the pressure exerts the most significant effect on the extraction yield due to the enhancement in the solvent power of CO₂ as a consequence of the solvent density increase, whereas the temperature did not present a significant effect on the extraction yield in the variable range evaluated. The positive effect of the interaction between temperature and pressure is clearly verified from Figure 2, which was already expected on the basis of the vapor pressure of volatile compounds and CO₂ solvent power, as discussed in the previous paragraph.

Figure 3 presents the experimental extraction curves for all conditions shown in Table 1. From this figure it is seen that, for the pressure of 100 bar, the constant extraction rate (M_CER) decreases with the increase of temperature (runs 1, 5 and 3 concerning the temperatures of 20, 35 and 50ºC, respectively), whereas for a pressure of 175 bar M_CER was practically constant around 1.75x10^-3 g/min. At 100 bar, a great variation in density was verified, affecting the M_CER, whereas at 175 bar the density was nearly constant among the experiments. At 250 bar and temperatures of 20 and 50 ºC, it was verified that the temperature increase promoted a decrease in M_CER, due to the decrease in the density, but at 35 ºC the greatest value of M_CER (2.340x10^-3 g/min) was verified. Although the density of the solvent is lower than run 2 (250 bar and 20 ºC), at this point the interaction between pressure and temperature was more favorable, since the temperature increases the vapor-pressure of extracts but has little effect on the density decrease. This result is confirmed in Figure 2, which clearly shows the influence of interaction between pressure and temperature on extraction yield.

From Figure 2 it is also verified that the time of constant extraction (tCER) was around 90 minutes for a pressure of 100 bar, independent of temperature, and for another pressure t_CER was 60 minutes. For run 4 (250 bar and 50 ºC) we verified only the constant extraction rate through the extraction time due to the high vapor-pressure of the extract. The falling extraction rate, when present, was short and it is related with the diffusion of the extract from the center of the particle to its surface. Concerning Figure 2, it is important to point out that the extraction yield after 240 minutes was still increasing, but extraction was not extended because the most important stage in supercritical fluid extraction is the constant extraction stage and the parameters associated with this step should be carefully determined in this work. About 80-90% of the extracts are obtained in the constant and falling extraction stages. In addition, the final stage of extraction is characterized by the extraction of wax that presents little interest in the extracts. This information is now provided in the manuscript.

Modeling of Extraction Curves

For the evaluation of experimental data, extraction curves were expressed in terms of mass of extract of M. aquifolia as a function of extraction time. Figure 3 shows the experimental and simulated data for all runs using the three models: Crank, Sovová and Martínez. Figure 3 also gives the determination coefficient (R²) for each model. The models of Sovová and Martínez presented good agreement with the experimental data, with values of R² higher than 0.97 for all conditions, whereas the Crank model showed unsuitable results because of its lower values of R². If R² is calculated considering all experimental points the values obtained are 0.9912, 0.9897 and 0.8504 for the models of Sovová, Martínez and Crank, respectively. A Tukey test (p<0.05) indicated that there are no significant differences between the predictions of the Martínez and Sovová models.

In the Sovová model, the transport phenomena that occur during the process are properly explained, such as the characteristics of the solid substrate (particle size, porosity, and relative amount of broken cells). Although the values of the parameters were not presented here, the fluid-phase mass transfer parameter was always higher than that exhibited by the solid-phase mass transfer. The appreciable difference between the two parameters indicates that, in the M. aquifolia SFE process, the effect of fluid phase convection is much more important than the diffusion in the solid substrate. Of course, at the end of the extraction process, when the solute from broken cells is depleted, diffusion becomes the most important mass-transfer mechanism, hence controlling the extraction process.

Although the model of Martínez provided a satisfactory representation of the experimental data during all the steps of the extraction process, its applicability is affected due to the fact that the values of the parameter t_mi that corresponds to the instant in which the extraction rate reaches its maximum were negative for some experimental conditions, which lacks physical meaning. The model of Crank was the least effective to simulate the SFE of M. aquifolia because of two reasons: first, the extraction process is controlled by convection and not by a diffusion mechanism in the solid substrate as discussed above; second, the model of Crank had only one adjustable parameter, which makes representation of the experimental data more difficult.

CONCLUSIONS

This work presented experimental and predicted data concerning the supercritical extraction of M. aquifolia. From the experimental results, it can be concluded that the extraction yield at low pressure is affected by the pressure or, in other words, by the fluid density. For pressures above 175 bar, the temperature variation of the density of fluid is less accentuated and the extraction yield is influenced by the temperature or, more precisely, by the increase in the vapor-pressure of the extract. The highest extraction yield was 1.10 wt% at 250 bar and 308K. The models of Sovová and Martínez showed satisfactory results concerning the fitting of experimental kinetic data for SFE of M. aquifolia. However, the modeling results obtained here should be evaluated in subsequent studies to verify their capability of predicting extraction curves on larger scales, and to evaluate whether the estimated parameters vary with operational conditions, such as the solvent flow rate, particle diameter, size, and geometry of the extraction bed.

ACKNOWLEDGMENTS

The authors thank CNPq, CAPES and FAPERGS for financial support and scholarships and also FEPAGRO for technical assistance.

(Submitted: January 20, 2011 ; Revised: April 5, 2012 ; Accepted: May 2, 2012)

• Angus, S., Armstrong, B. and Reuck, K. M., (Eds.) International Thermodynamic Tables of the Fluid State: Carbon Dioxide. Pergamon Press, Oxford, UK, Chapter 2, p. 340 (1976).
• Barroso, M. S. T., Villanueva, G., Lucas, A. M., Perez, G. P., Vargas, R. M. F., Brun, G. W., Cassel, E., Supercritical fluid extraction of volatile and non-volatile compounds from Schinus molle L. Brazilian Journal of Chemical Engineering, 28,305-312(2011).
• Brunner, G., Gas Extraction: An Introduction to Fundamentals of Supercritical Fluids and the Application to Separation Processes. 1^st Ed., Steinkopff, Darmstadt, Germany (1994).
• Carvalho, R. H. R., Galvão, E. L., Barros, J. A. C., Conceição, M. M., Souza, E. M. B. D., Extraction, fatty acid profile and antioxidant activity of sesame extract (Sesamum Indicum L.). Brazilian Journal of Chemical Engineering, 29,409-420(2012).
• Cranck, J., The Mathematics of Diffusion. Claredon Press, Oxford (1975).
• França, L. F. and Meirelles, M. A. M., Modeling of extraction of carotene and lipids from pressed palm oil (Elaes guineensis) fibers using supercritical CO₂ Journal of Supercritical Fluids, 18,35-47(2000).
• Gomide, R., Operações com sistemas sólidos granulares. Vol. 1, Catalogation of Brazilian Chamber Publication Book, São Paulo, Brazil (1983).
• Lanças, F. M., Vilegas, J. H. Y., Vasconcelos, E. C. and Celeghini, R. M. S., Novas aplicações de sistemas SFE "home made". I. Plantas medicinais Brasileiras. Ciência e Tecnologia de Alimentos, 4,413-417(1997). (In Portuguese).
• Martínez, J., Monteiro, A. R., Rosa, P. T. V., Marques, M. O. M and Meireles, M. A. A., Multicomponent model to describe extraction of ginger oleoresin with supercritical carbon dioxide. Industrial and Engineering Chemical Research, 42,1057-1063(2003).
• Mazutti, M., Mossi, A. J., Cansian, R. L., Corazza, M. L., Dariva, C., Oliveira, J. V., Chemical profile and antimicrobial activity of Boldo (Peumus boldus MOLINA) extracts obtained by compressed carbon dioxide extraction. Brazilian Journal of Chemical Engineering, 25,427-434(2008).
• Mezzomo, N., Martinez, J., Ferreira, S. R. S., Supercritical fluid extraction of peach (Prunus persica) almond oil: Kinetics, mathematical modeling and scale-up. The Journal of Supercritical Fluids, 51,10-16(2009).
• Mossi, A. J., Cansian, R. L., Carvalho, A. Z., Dariva, C., Oliveira, J. V., Mazutti, M., Nascimento Filho, I. and Echeverrigaray, S., Extraction and characterization of volatile compounds in Maytenus ilicifolia Mart. ex Reiss, using high-pressure CO₂ Fitoterapia, 75,168-178(2004).
• Oliveira, M. G., Monteiro, M. G., Macaúbas, C., Barbosa, V. P. and Carlini, E. A., Pharmacologic and toxicologic effects of two Maytenus species in laboratory animals. Journal of Ethnopharmacology, 34,29-41(1991).
• Paviani, L. C., Saito, E., Dariva, C., Marcucci, M. C., Sánchez-Camargo, A. P., Cabral, F. A., Supercritical CO₂ extraction of raw propolis and its dry ethanolic extract. Brazilian Journal of Chemical Engineering, 29,243-251(2012).
• Paviani, L. C., Pergher, S. B. C., Dariva, C., Application of molecular sieves in the fractionation of lemongrass oil from high-pressure carbon dioxide. Brazilian Journal of Chemical Engineering, 23,219-225(2006).
• Pereira, A. M. S., Menezes, A., Pereira, P. S., Cerdeira, R. M. M., França, S. C., Vilegas, J. H., Cordeiro, P. J. M. and Lanças, F. M., Effect of fertilization on morphologic characteristics and secondary metabolites of Maytenus aquifolium Mart. Journal of Herbs, Spices & Medicinal Plants, 3,43-50(1995).
• Pereira, A. M. S., Pereira, P. S., Cerdeira, R. M. M., Rodrigues, D. C., Moraes, F. R., Moraes, J. R. E. and França, S. C., Pharmacologically active compounds in plant tissue culture of Maytenus ilicifolia (Celastraceae). Acta Horticulturae, 333,205-210(1993).
• Santos, J., Dariva, C., Mossi, A., Stuart, G., Nascimento, I., Preliminary study of the influence of CO₂ extraction conditions on the ester, aldehyde, ketone and hydrocarbon content of grape bagasses from jam production. Brazilian Journal of Chemical Engineering, 24,637-642(2007).
• Schwaab, M., Biscaia, E. C., Monteiro, J. L. and Pinto, J. C. C. S., Nonlinear parameter estimation through particle swarm optimization. Chemical Engineering Science, 63,1542-1552(2008).
• Silva, C. F., Mendes, M. F., Pessoa, F. L. P., Queiroz, E. M., Supercritical carbon dioxide extraction of macadamia (Macadamia integrifolia) nut oil: Experiments and modelling. Brazilian Journal of Chemical Engineering, 25,175-181(2008).
• Souza-Formigoni, M. L., Oliveira, M. G. M., Monteiro, M. G. M., Silveira-Filho, N. G., Braz, S. and Carnelini, E. A., Antiulcerogenic effects of two Maytenus species in laboratory animals. Journal of Ethnopharmacology, 34,21-27(1991).
• Sovová, H., Rate of the vegetable oil extraction with supercritical CO₂ - I. Modeling of extraction curves. Chemical Engineering Science, 49,409-414(1994).

*

To whom correspondence should be addressed

Publication Dates

• Publication in this collection
  03 Jan 2013
• Date of issue
  Dec 2012

History

• Received
  20 Jan 2011
• Accepted
  02 May 2012
• Reviewed
  05 Apr 2012