Supercritical carbon dioxide extraction of glycyrrhizic acid from licorice plant root using binary entrainer: Experimental optimization via response surface methodology

Accepted Manuscript Title: Supercritical carbon dioxide extraction of glycyrrhizic acid from licorice plant root using binary entrainer: Experimental optimization via response surface methodology Author: Ali Hedayati S.M. Ghoreishi PII: DOI: Reference:

S0896-8446(15)00109-6 http://dx.doi.org/doi:10.1016/j.supflu.2015.03.005 SUPFLU 3268

To appear in:

J. of Supercritical Fluids

Received date: Revised date: Accepted date:

20-1-2015 9-3-2015 9-3-2015

Please cite this article as: A. Hedayati, S.M. Ghoreishi, Supercritical carbon dioxide extraction of glycyrrhizic acid from licorice plant root using binary entrainer: Experimental optimization via response surface methodology, The Journal of Supercritical Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.03.005 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Supercritical carbon dioxide extraction of glycyrrhizic acid from licorice plant root using binary entrainer: Experimental optimization via response surface methodology

ip t

Ali Hedayati, S.M. Ghoreishi* Department of Chemical Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran

cr

*Corresponding Author: e-mail: [email protected]; phone: +98-31-33915604

us

Abstract

In this study, the extraction of Glycyrrhizic acid (GA) from Glycyrrhiza glabra (licorice) root

an

was investigated by Soxhlet extraction and modified supercritical CO2 with methanol and water as co-solvents and 30 min of static time. Design of experiment was carried out with

M

response surface methodology (RSM) using Mini Tab software 17. The operating temperature (45-85°C), pressure (10-34 MPa), dynamic extraction time (40-120 min), CO2 flow rate (0.8-2

d

ml/min) and methanol concentration in water (0-100% as the binary co-solvent) were

te

considered as the range of operating variables. The high performance liquid chromatography

Ac ce p

(HPLC) was used to identify and quantitatively determine the amount of extracted GA. Response surface analysis verified that R2 and modified R2 of the model were 98.05% and 94.51%, respectively. The RSM modeling predicted the optimal operating conditions to be the pressure of 29.6 MPa, temperature of 68◦C, CO2 flow rate of 2 ml/min, dynamic extraction time of 108 min, methanol concentration of 46.5% in water (v/v) in which the maximum recovery of 54.4% was obtained. The accuracy of the modeling optimal GA recovery was validated with triplicate experiments giving the average extraction recovery of 54±1%. Keywords: Supercritical extraction; Glycyrrhizic acid; Response surface design; optimization; Methanol/Water Binary entrainer

1 Page 1 of 35

1. Introduction Glycyrrhiza glabra a leguminous shrub having a height of 70-200 cm occurs mainly in subtropical regions where it grows wild and also under cultivation [1]. There are several species

ip t

of Glycyrrhiza glabra (licorice), including Glycyrrhiza glabra L. var. typica Reg. et Herd., G.glabra L. var. pallida Boiss., G. glabra L. var. glandulifera Reg. et Herd. and G. glabra L.

cr

var. violacea Boiss. (Persian or Turkish liquorice) [2].

us

Licorice root and rhizomes are extensively used in herbal medicine for their demulcent, antacid, antiulcer [3], anti-inflammatory, expectorant, tonic, diuretic, laxative and sedative

an

properties [4, 5].They also possess antipyretic [6], antimicrobial, antiheroes [7], antiviral, anti-allergic, antioxidant and anti-cancerous activities [8]. It is widely used worldwide in food

M

confectionery and pharmaceutical products, such as cough syrups, herbal supplements, chewing gums, drinks and candy. In the traditional system of medicine, G. glabra is

d

recommended for the treatment of epilepsy [3]. Studies have shown that the extract has

te

estrogenic activity and may help regulate the estrogen-progesterone ratio [9, 10] and growthinhibitory effect on breast cancer cells [11, 12]. Many of the phenolic compounds isolated

Ac ce p

from licorice root may also help to protect low density lipoprotein (LDL) and red blood cells from oxidative damage [13-15]. Licorice root extract has also been shown to be beneficial for the liver. It has been used in Japan for more than twenty years as a treatment for chronic hepatitis, and studies with licorice root have shown a significant reduction of serum aminotransferase and a significant improvement in liver histology [16-18]. It has also been found that the licorice juice cause differences in the salivary pharmacokinetics of paracetamol if consumed with paracetamol [19]. Chemical constituents of the root include several bioactive compounds, such as glycyrrhizin (about 16%), different sugars (up to 18%), flavonoids, saponins, sterols, starches, amino acids, gums and essential oils. Kitawa [20] reported the detail structure of 33 constituents in 2 Page 2 of 35

licorice root and their sweetness. Glycyrrhizin is a pentacyclic triterpenoid glycoside. The glycoside usually occurs in a combined calcium or potassium salt form of Glycyrrhizic acid (GA) which is a weak acid containing three carboxyl and five hydroxyl groups. Molecular

ip t

structure of GA is shown in Fig 1. Moreover, GA is a polar compound. Glycyrrhizic acid the most studied active constituent in licorice is a sweet tasting material. It

cr

is 50 times sweeter than sugar and is widely used as a sweetening additive in the food industry [21, 22]. GA has anti-inflammatory, anti-ulcer, anti-oxidant, anti-hepatotoxic and

us

anti-virus activities [23-26]. There have been reports that GA has cancer chemo protective

an

function [27] and that it has been used clinically in patient with acids [28]. GA is also used as an additive in some foods and toothpaste. Therefore GA extraction from licorice root and its

M

characterization is considered to be an important research area.

Technology for the extraction of GA from licorice has been the subject of recent reports.

d

Among the different existing techniques, single pot extraction (SPE) in batch mode is the

te

most widely practical in herbal industry [29, 30]. Solvent extraction is not affordable due to the high consumption of energy and solvent, also complete solvent recovery is not possible

Ac ce p

which is very important in food and drugs application. In addition, organic solvents particularly chlorinated ones that are used in traditional separation methods are very harmful to humans, environment and the ozone layer so they are rejected in the view of green chemistry [31]. Thus it is necessary to develop an effective and optimal extraction method. Application of various novel techniques such as microwave assisted extraction [32], microwave assisted micellar extraction [33], ultrasound assisted extraction [34] and multi stage countercurrent extraction [35] for the extraction of GA from licorice root has been already reported in literatures. Disadvantages of these methods could be eliminated by replacing the toxic liquid solvent with supercritical CO2 (SC-CO2), because CO2 is inexpensive, nontoxic, nonflammable with low critical temperature and pressure and

3 Page 3 of 35

environmentally benign. Attentions to the supercritical technology are growing due to the unique properties of supercritical fluids, such as high selectivity, liquid-like densities, gas-like viscosities and low surface tension. The low polarity is the only major drawback of CO2 that

ip t

leads to low extraction recovery of GA. Thus co-solvents can be used to change the polarity of SC-CO2 and increase its solvation power to desired analyte [36]. The enormous advantages

cr

of supercritical fluids have caused their widely utilization in different industries [37-39].

The main objective of this research was GA extraction from Persian licorice by modified SC-

us

CO2 in periodic static-dynamic procedure for pharmaceutical application. The optimization

an

was carried out by response surface methodology (RSM). The RSM is useful for modeling, problem analysis and optimization when a response (i.e., extraction recovery) is influenced by

M

several variables such as pressure, temperature, flow rate of SC-CO2, co-solvent concentration (type of polar modifier) and dynamic time. In current work the other effective variable (static

d

time) was fixed at optimum value of 30 min that was obtained by experiment. Furthermore, a

Ac ce p

2. Experimental

te

binary entrainer (methanol-water) was used to enhance the recovery of GA.

2.1. Materials

Persian Licorice root was purchased from agricultural research center. The samples were ground and screened with mesh size of 20–35 (0.841–0.507 mm). The primary samples were stored inside a sealed bag and were maintained in a cold and dry place for extraction experiments. Methanol (purity ≥ 99.9%, Merck), pure acetic acid (Merck) deionized water and industrial grade carbon dioxide (purity > 99%, Ardestan) were utilized for Soxhlet and supercritical fluid extraction and HPLC analysis. Glycyrrhizic acid ammonium salt (>95%, Sigma-Aldrich) was used in the HPLC analysis (Jasco HPLC equipped with a UV detector) to obtain a standard spectrum.

4 Page 4 of 35

2.2. Soxhlet extraction For Soxhlet extraction, 2 g of ground licorice root was weighted and set in a Soxhlet apparatus and then continuously extracted for 8 hours using methanol as solvent. After

ip t

extraction, the solvent was evaporated by rotary vacuum evaporator (30 ◦C), and the extract was dried at 70◦C to remove residual solvent to desired amount. Then the amount of GA was

cr

determined by HPLC method. The results showed that Soxhlet extraction provided a yield of

calculating the GA recovery by SC-CO2 extraction.

us

138 mg GA/g licorice root, which was considered as the total extractable GA content while

an

2.3. Supercritical fluid extraction: apparatus and procedure

To carry out the objectives of this study, the supercritical extraction system shown in Fig. 2

M

was used. To increase the purity of the CO2, which is stored in a CO2 cylinder (1), it was passed through a column of molecular sieve (Merck, Molecular Sieve 5A-K28751105148)

d

and metal porous filter (Mott Metallurgical, 1003630-01-050) (2). Then, CO2 is cooled down

te

in a chiller and pump head cooler (3, 4) in the range of −10 to -5 ◦C. Then liquefied CO2 is

Ac ce p

charged by a feed pump (reciprocating pump, Jasco, PU-1580, maximum pressure = 35 MPa) (5) through the needle valve (6) and is fed to the oven that controlled the temperature (298– 523 ± 0.5 K) with proportional–integral–derivative controller (PID controller) (11). Two vessels (modifier storage column and extraction column) were placed in the oven. The first one (8) is filled with methanol and water as the binary modifier (10). CO2 is preheated using spring coil preheater (7) before entering the modifier vessel (height = 12.5 cm, inner diameter = 0.9 cm, and outer diameter = 1.3 cm) (8) which is inside the oven. Then CO2 was first charged into the entrainer vessel (8) and saturated with methanol/water mixture. Then it was preheated using a spring coil preheater (9) before entering the extraction column. This column (10) was charged with 2 g of licorice root. Subsequently, modified SC-CO2 was passed through the second vessel and carried out the extraction of GA from licorice root. In this 5 Page 5 of 35

process, the pressure of the extraction vessel is controlled via needle valve (12) and back pressure regulator (BPR) valve (TESCOM, 26-1762-24, maximum pressure = 40.8 MPa), in which the pressure fluctuation can be controlled within 0.2 MPa. (13). Finally, the modified

ip t

SC-CO2 is expanded via passage through the BPR valve and the extracted material in all experiments were collected in bottles containing 10 ml of methanol that were placed in a

cr

vessel containing ice, salt, and acetone (14). For human usage, GA could be recovered from entertainers via (1) freeze dryer and/or (2) oven drier up to the temperature of 70 ◦C (methanol

us

boiling point) which is much lower than degradation temperature of GA.

an

For desired mass transfer, after reaching the corresponding supercritical fluid conditions inside the extraction column, the static time was provided for extraction equilibrating process

M

via valve closure (6). Especially, static extraction is necessary when a liquid as entrainer is used, because it allows the entrainer penetration into the sample matrix. The results of

d

screening experiment in this study indicated that optimum static time is 30 min and increasing

te

static extraction did not have any significant effect on recovery due to gradual depletion of mass transfer driving force. Using dynamic extraction time after static extraction causes

Ac ce p

higher mass transfer driving force and provides a suitable condition for extraction. Thus after the static extraction, the dynamic extraction with constant volumetric flow rate of CO2 was started by opening the valve (12). At this stage, the system pressure was regulated by a back pressure (13). The GA recovery and extraction yield were calculated using the following equations:

6 Page 6 of 35

2.4. HPLC analysis Pure standard GA and extracted GA were analyzed and quantified by a HPLC apparatus. The Venusil MP C18 column was used (length = 250 mm, internal diameter = 4.6 mm and particle characteristic= 5µm, 100Å, 360m2/g ultra-pure (99.999%) silica). A pre-column was also

ip t

used prior to analytical column for the removal of pollutant and residuals from solvent and sample components. Mobile phase, mobile phase velocity, temperature, detector and injected volume were mixture of 41% methanol and 59% water (3% acetic acid), 1 ml/min, ambient

cr

temperature, UV detector with 248 nm wavelength and 20 µL, respectively. The calibration curve was plotted by injection of different concentration of solutions made of standard

us

Glycyrrhizic acid ammonium salt with linear regression R2 = 0.999 and retention time=25 min. The concentrations of GA in extracted samples were obtained by linear calibration curve

an

according to y=1764.6x-4880.8.

Samples were centrifuged for 5 min at the rate of 9000 round per minutes for removal of

M

suspended substance in the liquid before HPLC analysis. 3. Design of experiment

d

Traditionally, optimization has been carried out by checking the influence of one factor at a

te

time on response. In other words, only one parameter is changed while others are kept at a

Ac ce p

constant level. The major disadvantages of this optimization technique were (1) the high number of required experiments which leads to increased time and expenses and (2) the interactive effects among different operating variables are not considered. In order to overcome these problems, the optimization was carried out by using multivariate statistic techniques such as RSM [40]. RSM is combination of mathematical and statistical techniques that are useful for modeling and analysis issues when a response is influenced by several variables. The objective is to simultaneously optimize the levels of these variables to obtain the best response. Stages of RSM application are as follows:

7 Page 7 of 35

1. The selection of independent and dependent variables with the greatest effect on response. Numerous variables may affect the response of the studied system, and it is practically impossible to identify and control all of them. Therefore, it is necessary to choose those

ip t

variables with major effects. 2. The choice of the experimental design, carrying out the experiments according to the

an

us

quadratic variables as well as interaction terms is shown in Eq. 3:

cr

selected experimental matrix and getting answers. The RSM model which includes linear and

Where A0 is the constant, k is the number of variables, Ai represents the coefficients of the

M

linear parameters, Zi represents the variables, Aii is the coefficients of the quadratic parameters, Aij represents the coefficients of the interaction parameters and ε is the residual

d

associated with the experiments.

te

The most known second order symmetrical designs is central composite rotatable design (CCRD) which was selected for five independent variables, each at five levels to fit second-

Ac ce p

order polynomial model. The experimental design is based on coded variables because it enables the investigation of variables with different orders of magnitude. The variables were coded according to the following equation:

Where Zi is the coded value of the independent variable, Xi is the real value, Xi,c.p is the real value at the center point, and is the step change in the variable Xi. 3. Find a suitable model of the response variables based on the fit of a polynomial equation to the experimental data. After acquiring data related to each experimental point, it is necessary to fit a mathematical equation to describe the behavior of the response according to the levels of studied values. In other words, the coefficients of Eq. 3 are obtained via least square 8 Page 8 of 35

method (LSM). LSM is a multiple regression technique used to fit a mathematical model to a set of experimental data with generating the lowest residual possible. 4. The evaluation of the model’s validity. The achieved mathematical model was evaluated by

ip t

the application of analysis of variance (ANOVA), coefficient of determination (R2), and adjusted coefficient of determination (Adj. R2).

cr

5. Obtaining the optimum values of model for each studied variable. Finally, the necessary conditions were obtained for optimum recovery of GA extraction through the first derivative

us

of the mathematical function, which describes the response surface [41-43].

an

Among the different experimental design methods, RSM was chosen in this study to determine the effects of temperature, pressure, dynamic time, flow rate and methanol

M

concentration in binary modifier on the extraction efficiency of GA. Accordingly, five levels central composite rotatable design (CCRD) with 5 independent variables was used.

d

The independent variables in this research were dynamic time (t), pressure (P), temperature

te

(T), flow rate of SC-CO2 (Q), methanol concentration in binary modifier (C) in the range of 40–120 min (step size=20), 10–34 MPa (step size=6), 45–85 ◦C (step size=10), 0.8–2 ml/min

Ac ce p

(step size=0.3) and 0-100% (step size=25%), respectively. Recovery of GA extraction was response variable (dependent variable). The coded and uncoded levels of independent variables are shown in Table 1. Design of experiment was based on coded levels of the independent variables in the CCRD which resulted in 32 experimental runs shown in Table 2. The obtained responses were also shown in this table. Minitab software (version 17) was used for implementation of RSM. 4. Results and Discussion The effect of four different types of modifier was investigated on GA recovery in Fig. 3. As illustrated, the minimum and maximum recovery of GA is corresponded to pure ethanol and methanol-water (50-50 v/v) to be 13.2 and 47%, respectively. The GA recovery for two other 9 Page 9 of 35

pure co-solvents, methanol and water, is obtained to be 15.6 and 34.1%, respectively. These results indicated that the utilization of binary entrainer is more effective in recovery enhancement than pure modifier. The precision of the experimental recovery data presented in

ip t

Fig. 3 with triplicate experiments is average value± 1%. It is important to realize that the main reason for using methanol in the binary entrainer was the fact that pure methanol had a higher

us

4.1. Fitting the model and analysis of experimental data

cr

positive effect on GA recovery in contrast to pure ethanol.

The shown experimental data in Table 2 were used to calculate the coefficients of the second-

an

order polynomial equation (Eq. 5) by the least squares technique. These regression

M

coefficients were summarized in Table 3 and the significance of each coefficient was determined by absolute t-value and p-value which were listed in the same table. For any of the

d

terms in the model, a large absolute value of t and a small p-value would indicate more

te

significant effects on the corresponding response variables. In this research, model terms with p < 0.001, 0.001 ≤ p < 0.05 and p ≥ 0.05 are highly significant, significant and insignificant,

Ac ce p

respectively. The second-order polynomial model in terms of coded variables was obtained for GA recovery (R) as a function of independent variables in Eq. 5:

The results indicated that linear terms of pressure and time were highly significant (p < 0.001) while CO2 flow rate, temperature and methanol concentration in binary entrainer were significant. The quadratic terms of all variables except CO2 flow rate were highly significant (p < 0.001). Moreover, all linear interactions between different variables were nonsignificant. The fitted model was verified via the analysis of variance (ANOVA) shown in Table 4. The

10 Page 10 of 35

calculated coefficient of determination (R2) and adjusted coefficient of determination (Adj. R2) were 98.05% and 94.51%, respectively. These values indicate that the model adequately represented the experimental data and 94.51% of the variations could be covered by the fitted

cr

4.2. Optimization of extraction operating conditions

ip t

model.

us

The maximum GA recovery of 54.4% was predicted by the RSM model at the optimal operating conditions of 29.6 MPa, 68oC, 108 min (dynamic time), 2 ml/min and 46.5%

an

methanol concentration in binary modifier. The accuracy of the modeling maximum GA recovery was validated with triplicate experiments at the optimal operating conditions giving

4.3. Response surface analysis

M

the average extraction recovery of 54±1%.

d

The interactions among different variables and optimum values for attaining the maximum

te

recovery were investigated by a three-dimensional response surface model according to Eq. 5.

Ac ce p

The plots were designed by sketching the response (z-axis) versus two independent variables (x and y coordinates) and the other three independent variables were considered constant at zero levels. Fig. 4 shows the effects of pressure and temperature on GA recovery at the fixed dynamic extraction time (80 min), flow rate of SC-CO2 (1.4 ml/min) and 50% methanol concentration in the binary co-solvent. The SC-CO2 density increases with increasing pressure that leads to improved GA solubility and therefore causes higher recovery. On the other hand, increasing the pressure decreased the GA diffusivity, convective mass transfer coefficient, and, resulting in a lower GA recovery. The trend of Fig. 4 illustrates that increasing operating pressure from 10 MPa (code = −2) to 29 MPa (code = 1.27) increases the extraction recovery because of overwhelming positive

11 Page 11 of 35

effect of increased SC-CO2 density. But with further increasing of pressure from 29 MPa (code = 1.27) to 34 MPa (code = 2), the effect of decreasing permeability and mass transfer coefficient prevails the increasing effect of density and GA recovery decreased.

ip t

As indicated in Table 4, pressure had a positive linear effect on GA recovery at low-pressure levels (code = −2 to 1.27) (p < 0.001), most likely due to the improvement of GA solubility

cr

resulting from the increased CO2 density with the rise of pressure. At higher pressure levels (code = 1.27 to 2), however, the negative quadratic effect of pressure on the GA recovery also

us

became an important factor (p < 0.001). This is probably a reflection of the increased

an

repulsive solute–solvent interactions resulting from the highly compressed CO2 at highpressure levels and reduction of mass transfer and permeability. Thus the highest recovery

M

was obtained at almost 29 MPa (code = 1.27) according to Section 4.2. As summarized in Table 4, the temperature had a positive linear (0.001 ≤ p < 0.01) and

d

negative quadratic (p < 0.001) effect, while the interaction of temperature–pressure displayed

te

nonsignificant effect on GA recovery (p ≥0.05). These two opposing effects can be clearly observed in Fig. 4. As temperature increases, a dual counter effect is observed. On one hand,

Ac ce p

higher temperature increases the GA vapor pressure, resulting in enhanced recovery. On the other hand, further increasing the temperature leads to lower SC-CO2 density and, therefore, reduces the solubility of GA and causes lower recovery. Based on which of these two effects is dominant, the effect of temperature on the extraction will be different. As illustrated in Fig. 4, increasing temperature from 45 ◦C (code = −2) to 68 ◦C (code =0.3) leads to more extraction recovery but further increasing temperature from 68◦C (code=0.3) to 85◦C (code=2) leads to lower recovery of GA extraction. In the aforementioned region (45-68 ◦C), the positive effect of temperature with enhanced mass transfer coefficient and diffusivity is more dominant than the negative effect of temperature with lower SCF density. At the optimal temperature (68 ◦C) the dual counter effects are equal in which the thermodynamic

12 Page 12 of 35

phenomenon of retrograde solubility is reached and beyond that the negative effect of temperature becomes more significant and subsequently reduction of extraction recovery is observed with a steep slope. The optimal operating temperature at 68◦C (retrograde solubility)

ip t

was predicted by the RSM model that shows the incentives of RSM modeling and its application in prediction of optimum conditions in order to obtain the maximum recovery.

cr

Fig. 5 shows the effects of extraction dynamic time and CO2 flow rate on GA recovery at 22 MPa, 55◦C and 50% methanol concentration in water. Increasing the SC-CO2 flow rate from

us

0.8 ml/min (code = −2) reduces the film thickness around the solid particles and this leads to

an

lower external mass transfer resistance around the licorice root and consequently enhanced GA extraction recovery. Table 4 shows a positive linear effect of SC-CO2 flow rate (0.001 ≤ p

M

< 0.05) on the response variable. But at flow rate beyond 1.7 ml/min two opposite effects of lower residence time (i.e., fluid-bed contact time) and higher film mass transfer coefficient

d

with negative and positive effects reduces the slope of GA recovery enhancement. This is due

te

to positive linear (0.001 ≤ p < 0.05) and negative quadratic effect of flow rate on extraction recovery. Ghoreishi and Bataghva [44] reported that the static extraction (25 min) before

Ac ce p

dynamic process improved the SC extraction recovery due to equilibrating effect on fatty acid methyl ester (FAME) transfer from solid particles to fluid phase. In this work the optimal static extraction time was experimentally determined to be 30 min. In this research beside static time, the effect of dynamic extraction time on GA recovery was also investigated. Increasing dynamic time can improve the GA recovery (because fresh solvent passes through licorice root fixed bed) until there is effective mass transfer driving force between the SC-CO2 and licorice root. According to Table 3, the dynamic extraction time had a positive linear effect on GA recovery up to optimum dynamic time of 108 min (p < 0.001). However, beyond 108 min (code = 1.43), the negative quadratic effect also became significant (p < 0.001) and mass transfer driving force gradually is eliminated and therefore no more recovery

13 Page 13 of 35

enhancement is observed. The same trend for dynamic time was obtained by other researches [36, 40, and 43]. In order to enhance the effect of co-solvent on GA recovery, a variety of methanol-water

ip t

mixtures of different concentrations ranging from 0-100% (v/v) were employed as modifiers and the effect of methanol contents in aqueous methanol solution on SFE of GA were

entrainer used in conjunction with supercritical CO2 was 5% (v/v).

cr

examined under identical condition. It is imperative to realize that the maximum amount of

us

Fig. 6 shows the effect of methanol concentration in binary co-solvent on the GA recovery.

an

Increasing methanol concentration from 0% (coded= -2) until concentration of 46.5% (coded= -0.14) enhanced GA extraction recovery. But GA recovery was reduced by increasing the

M

amount of methanol concentration to 100% (coded=2). As shown, the trend of GA recovery versus methanol concentration at different CO2 flow rate (Fig. 6a), dynamic time (Fig. 6b),

d

temperature (Fig. 6c) and pressure (Fig. 6d) are similar. An increasing GA recovery is

te

observed up to a maximum level and then the reduction of GA recovery occurred. The effect of three different levels of SC-CO2 flow rate, dynamic time, temperature and pressure on GA

Ac ce p

recovery is the same observed trend of Fig. 4-5, which is theoretically expected. The increasing SC-CO2 flow rate, dynamic time and pressure indicate an enhancement effect on GA recovery. But due to thermodynamic phenomenon of retrograde solubility, increasing temperature has a positive effect on recovery up to a certain value (retrograde solubility) in which higher GA vapor pressure is dominant and further increase of temperature leads to lower SCF density which causes lower recovery. The aforementioned elucidation of temperature effect is obtained via Fig. 4 and Fig. 6c. 5. Conclusions In this study, optimal conditions for the SFE of Glycyrrhizic acid from licorice root were investigated using modified SC-CO2 with polar binary entrainer. Optimization of this process 14 Page 14 of 35

was carried out via RSM and experimental investigation. Overall, it could be concluded that the SFE of GA from licorice using SC-CO2 modified with aqueous methanol is a much better method than conventional organic solvent extraction when the extraction efficiency and the

ip t

required amount of environmentally non-benign organic solvent are considered. The obtained results of GA recovery indicated that using binary co-solvent is much better than pure

cr

entrainer. Response surface analysis verified that the data were adequately fitted to secondorder polynomial model. The linear terms of temperature, pressure, CO2 flow rate, co-solvent

us

methanol concentration and dynamic time as well as quadratic terms of all variables except

an

CO2 flow rate had significant effects on the obtained RSM model of GA recovery based on coded variables. The predicted and experimental results of this study could be utilized by

M

process engineers for design and scale up of SCF of GA in the pharmaceutical industry.

d

Acknowledgements

te

The financial support provided by Isfahan University of Technology and Iran National

Ac ce p

Science Foundation (INSF) is gratefully acknowledged. 6. Nomenclature and units A0 Ai Aii Aij

constant

Coefficient of linear parameters Coefficient of quadratic parameters Coefficient of interaction parameters

k

Number of variables

P(MPa)

extraction pressure

Q(ml/min)

CO2 flow rate

R (%)

Recovery

15 Page 15 of 35

T(◦C)

Extraction temperature

Xi

Real values

Xi,c,p

Real values at the center point

Y

yield

Zi

Coded value of the independent variable

Step change in variable Xi

ε

Residual associated with experiments

an

ζ

us

Greek letters

7. References

G.R. Fenwick, J. Lutomski, C. Nieman, Liquorice, Glycyrrhiza glabra L. composition,

M

[1]

ip t

Dynamic extraction time

cr

t(min)

uses and analysis, J. Food Chemistry 38 (1990) 119-143. U. Quattrocchi, CRC World Dictionary of Medicinal and Poisonous Plants: Common

d

[2]

te

Names, Scientific Names, Eponyms, Synonyms, and Etymology (5 Volume Set),

[3]

Ac ce p

Taylor & Francis Group, Boca Raton, FL, 2012, pp1862. S.D. Ambawade, V.S. Kasture, S.B. Kasture, Anticonvulsant activity of roots and rhizomes of glycyrrhiza glabra, Indian J. Pharmacology 34 (2002) 251-255.

[4]

G.V. Obolentseva, V.I. Litvinenko, A.S. Ammosov, T.P. Popova, A.M. Sampiev, Pharmacological and therapeutic properties of licorice preparayions: A review, Pharmaceutical Chemistry J. 33(1999) 24-31.

[5]

Lakshmi T. and Geetha R., Glycyrrhiza glabra Linn. commonly known as licorice: A therapeutic review, International J. Pharmacy and Pharmaceutical 3 (2011) 20-25.

[6]

S. Lata, R.S. Saxena, A. Kumar, S. Kakkar, V.K. Srivastava, K.K. Saxena, Comparative antipyretic activity of Ocimum sanctum, glycyrrhiza glabra and aspirin in experimentally induced pyrexia in rat, Indian J. Pharmacology 31(1999) 71-75 . 16 Page 16 of 35

[7]

[7] L.A. Baltina, R.M. Kondratenko, L.A. Baltina, Jr., O.A. Plyasunova, A.G. Pokrovskii, G.A. Tolstikov, Prospects for the creation of new antiviral drug based on glycyrrhizic acid and its derivatives: a review, Pharmaceutical Chemistry J. 43(2009)

[8]

ip t

3-12. V. K. Guptaa, A. Fatimaa, U. Faridia, A.S. Negib, K. Shankerb, J.K. Kumarb, N.

cr

Rahujaa, S. Luqmana, B.S. Sisodiaa, D.Saikiaa, M.P. Darokara, S.P.S. Khanujaa, Antimicrobial potential of Glycyrrhiza glabra roots, J. Ethnopharmacology 116 (2008)

C.H. Costello, E.V. Lynn, Estrogenic substances from plants: I. Glycyrrhiza, J.

an

[9]

us

377-380.

Pharmaceutical Science 39(1950)177-180.

A. Kumagai, K. Nishino, A. Shimomura, Effect of glycyrrhizin on estrogen action,

M

[10]

Endocrinologia Japonica 14 (1967) 34-38.

S. Tamir, M. Eizenberg, D. Somjen, N. Stern, R. Shelach, A. Kaye, J. Vaya,

d

[11]

te

Estrogenic and antiproliferative properties of glabridin from licorice in human breast cancer cells, Cancer Research 60 (2000) 5704-5709. E.H. Jo, S.H. Kim, J.C. Ra, S.R. Kim, S.D. Cho, J.W. Jung, SR Yang, JS Park, J.W

Ac ce p

[12]

Hwang, O.I Aruoma, T.Y Kim, Y.S Lee, K.S Kang, Chemopreventive properties of the ethanol extract of Chinese licorice (Glycyrrhiza uralensis) root: induction of

apoptosis and G1 cell cycle arrest in MCF-7 human breast cancer cells, Cancer Letters 230 (2005) 239-247.

[13]

J. Vaya, P.A. Belinky, M. Aviram, Antioxidant constituents from licorice roots: isolation, structure elucidation and antioxidative capacity toward LDL oxidation, Free Radical Biology and Medicine 23 (1997) 302-313.

17 Page 17 of 35

[14]

P.A. Belinky, M. Aviram, B. Fuhrman, M. Rosenblat, J. Vaya, The antioxidative effects of the isoflavan glabridin on endogenous constituents of LDL during its oxidation, Atherosclerosis 137 (1998) 49-61. H. Haraguchi, H. Ishikawa, K. Mizutani, Y. Tamura, T. Kinoshita, Antioxidative and

ip t

[15]

superoxide scavenging activities of retrochalcones in Glycyrrhiza inflate, Bioorganic

[16]

cr

& Medicinal Chemistry 6 (1998) 339-347.

T.G.V. Rossum, A.G. Vulto, R.A. de Man, J.T. Brouwer, S.W. Schalm, Review

[17]

an

Pharmacology & Therapeutics 12 (1998) 199-205.

us

article: glycyrrhizin as a potential treatment for chronic hepatitis C, Alimentary

Y. Yamamura, H. Kotaki, N. Tanaka, T. Aikawa, Y. Sawada, T. Iga, The

M

pharmacokinetics of glycyrrhizin and its restorative effect on hepatic function in patients with chronic hepatitis and in chronically carbon-tetrachloride-intoxicated rats,

E. Gumpricht, R. Dahl, M.W. Devereaux, R.J. Sokol, Licorice compounds

te

[18]

d

Biopharmaceutics & Drug Disposition 18 (1997) 717-725.

glycyrrhizin and 18beta-glycyrrhetinic acid are potent modulators of bile acid-induced

[19]

Ac ce p

cytotoxicity in rat hepatocytes, J. Biological Chemistry 280 (2005) 10556-10563. N.A Qinna, E.M. Mallah, T.A. Arafat, N.M. Idkaidek, Effect of licorice and grapefruit juice on paracetamol pharmacokinetics in human saliva, J. Biological Chemistry 4 (2012) 158-162.

[20]

I. Kitagawa, Licorice root. A natural sweetener and an important ingredient in Chinese medicine, Pure and Applied Chemistry 74 (2002) 1189-1198.

[21]

S.K. Acharya, S. Dasarathy, A. Tandon, Y.K. Joshi, B.N. Tandon, A preliminary open trial on interferon stimulator (SNMC) derived from glycyrrhiza glabra in the treatment of subacute hepatic failure. Indian J. Medical Research 98( 1993) 69-74.

18 Page 18 of 35

[22]

H. S. Kim, S. Y. Lee, B. Y. Kim, E. K. Lee, J. H. Ryu, G. B. Lim, Effects of modifiers on the supercritical CO2 extraction of glycyrrhizin from licorice and the morphology of licorice tissue after extraction, Biotechnology and Bioprocess Engineering 9 (2004)

[23]

ip t

447-453. Y. Fujisawa, M. Sakamoto, M. Matsushita, T. Fujita, K. Nishioka, Glycyrrhizin

cr

inhibits the lytic pathway of complement: possible mechanism of its antiinflammatory effect on liver, Microbiology and Immunology 44 (2000) 799-804. A.R. Dehpour, M.E. Zolfaghari, T. Samadian, F. Kobarfard, M. Faizi, M. Assari,

us

[24]

an

Antiulcer activities of liquorice and its derivatives in experimental gastric lesion induced by ibuprofen in rats, International J. Pharmaceutics 119 (1995)133-138. J. Cinatl, B. Morgenstern, G. Bauer, P. Chandra, H. Rabenau, H.W. Doerr,

M

[25]

Glycyrrhizin, an active component of liquorice roots, and replication of SARS-

B. Fu, J. Liu, H. Li, L. Li, F.S.C. Lee, X. Wang, The application of macroporous

te

[26]

d

associated coronavirus, The Lancet 361 (2003) 2045-2046.

resins in the separation of licorice flavonoids and glycyrrhizic acid, J.

[27]

Ac ce p

Chromatography A 1089 (2005) 18-24. G. Kuttan, P. Pratheeshkumar, K. A.

Manu, R.

Kuttan, Inhibition of tumor

progression by naturally occurring terpenoids, Pharmaceutical Biology 49 (2011) 9951007.

[28]

T. Hattori, S. Ikematsu, A. Koito, S. Matsushita, Y. Maeda,M. Hada, M. Fujimaki,

K.

Takatsuki, Preliminary evidence for inhibitory effect of glycyrrhizin on HIV

replication in patients with AIDS, Antiviral Research 11(1989) 255-61. [29]

L.Wang, L.C. Weller, Recent advances in extraction of nutraceuticals from plants, Trends in Food Science & Technology 17(2006) 300-312.

19 Page 19 of 35

[30]

C.W. Huie, A review of modern sample-preparation techniques for the extraction and analysis of medicinal plants, Analytical and Bioanalytical Chemistry 373 (2002) 2330. S.M. Ghoreishi, R. Gholami Shahrestani, H.S. Ghaziaskar, Experimental and

ip t

[31]

modeling investigation of supercritical extraction of mannitol from olive leaves,

[32]

cr

Chemical Engineering & Technology 32 (2009) 45–54.

X. Pana, H. Liua, G. Jiaa, Y.Y. Shub, Microwave-assisted extraction of

us

glycyrrhizic acid from licorice root, Biochemical Engineering J. 5 (2000) 173-

[33]

an

177.

S. Chen, X. Yuchun , L. Huizhou, Microwave-assisted Micellar Extraction and

M

Determination of Glycyrrhizic Acid and Liquiritin in Licorice Root by HPLC, Chinese J. Chemical Engineering 15(2007) 474 - 477.

T.W. Charpe, V.K. Rathod, Extraction of glycyrrhizic acid from licorice root using

d

[34]

(2012) 37– 41.

Q. Wang, S. Ma, B. Fua, F.S.C. Lee, X. Wang, Development of multi-stage

Ac ce p

[35]

te

ultrasound: Process intensification studies, Chemical Engineering and Processing 54

countercurrent extraction technology for the extraction of glycyrrhizic acid (GA) from licorice (Glycyrrhiza uralensis Fisch) Biochemical Engineering J. 21 (2004) 285–292

[36]

S.M. Ghoreishi, E. Heidari, Extraction of Epigallocatechin-3-gallate from green tea via supercritical fluid technology: Neural network modeling and response surface optimization, J. of Supercritical Fluids 74 (2013) 128– 136.

[37]

N. Esfandiari, S.M. Ghoreishi, Synthesis of 5-Fluorouracil nanoparticles via supercritical gas antisolvent process, The J. Supercritical Fluids 84(2013) 205–210.

20 Page 20 of 35

[38]

Gh. Sodeifiana, J. Azizia, S.M. Ghoreishi, Response surface optimization of Smyrnium cordifolium Boiss (SCB) oil extraction via supercritical carbon dioxide, J. Supercritical Fluids 95(2014) 1–7. S.M. Ghoreishi, , P. Moein, Biodiesel synthesis from waste vegetable oil via

ip t

[39]

transesterification reaction in supercritical methanol, J. Supercritical Fluids 76(2013)

[40]

cr

24–31.

S.M. Ghoreishi, E. Bataghva, Supercritical extraction of evening primrose oil:

us

experimental optimization via response surface methodology, J. American Institute of

[41]

an

Chemical Engineers (AIChE) 57 (2011) 3378–3384.

M.A. Bezerra, R.E. Santelli, E.P. Oliveira, L.S. Villar, L.A. Escaleira, Response

Talanta 76 (2008) 965–977.

S.M. Ghoreishi, E. Bataghva, A.A. Dadkhah, Response surface optimization of

d

[42]

M

surface methodology (RSM) as a tool for optimization in analytical chemistry, J.

te

essential oil and diosgenin extraction from tribulus terrestris via supercritical fluid technology, Chemical Engineering & Technology 34 (2011) 1–10. S.M. Ghoreishi, E. Mardani, H.S. Ghaziaskar, Separation of

Ac ce p

[43]

-linolenic and other

polyunsaturated fatty acids from boraginaceae via supercritical CO2, J. Separation Science 34 (2011) 233–240.

[44]

S.M. Ghoreishi, E.Bataghva, Supercritical extraction of essential oil from Echium amoenum seed: Experimental, modeling and genetic algorithm parameter estimation, Korean J. Chemical Engineering 31 (2014) 1632-1640.

21 Page 21 of 35

ip t cr us an M

[ºC]

[MPa]

Ac ce p

variables

Temperature(T) Pressure(P)

te

Coded

d

Table 1: Uncoded and coded levels of independent variables used in the RSM design

*

Dynamic

Flow rate

methanol

time(t)

of CO2(Q)

concentration in

[min]

[ml/min]

binary entrainer (c)*

-2

45

10

40

0.8

0%

-1

55

16

60

1.1

25%

0

65

22

80

1.4

50%

+1

75

28

100

1.7

75%

+2

85

34

120

2

100%

Maximum amount of entrainer used in conjunction with supercritical CO2 = 5% (v/v).

22 Page 22 of 35

Table 2: Central composite rotatable design for GA extraction and observed responses: GA yield (Y) and recovery (R).

te

Ac ce p

R 24.12 40.89 44.29 32.13 30.79 44.34 48.09 32.92 36.42 33 36.29 40.24 40.1 29.13 36.73 43.35 47.34 37.07 39 48.61 48.48 48.2 45.74 34.45 48.34 31.52 52 37.33 35.25 47.64 35.41 47.5

Y 33.28 56.42 61.12 44.33 42.49 61.18 66.36 45.42 50.25 45.54 50.08 55.53 55.33 40.19 50.68 59.82 65.32 51.15 53.82 67.08 66.9 66.51 63.12 47.54 66.7 43.49 71.76 51.51 48.64 65.74 48.86 65.55

ip t

Q 0 1 -2 1 1 1 0 0 -1 0 -1 1 -1 -1 1 -1 0 0 0 0 2 0 1 0 0 -1 0 1 -1 0 -1 0

cr

T 0 1 0 -1 1 -1 0 0 1 0 -1 1 -1 -1 -1 1 0 0 2 0 0 0 1 -2 0 1 0 -1 -1 0 1 0

us

t 0 -1 0 -1 -1 1 0 0 -1 -2 1 1 1 -1 1 1 0 0 0 0 0 0 1 0 0 -1 2 -1 -1 0 1 0

an

P 0 1 0 -1 -1 1 0 0 1 0 -1 -1 1 -1 -1 1 2 -2 0 0 0 0 1 0 0 -1 0 1 1 0 -1 0

M

C 2 -1 0 -1 1 -1 0 -2 1 0 -1 -1 1 1 1 -1 0 0 0 0 0 0 1 0 0 -1 0 1 -1 0 1 0

d

Number of experiments 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

23 Page 23 of 35

Table 3: Regression coefficients of predicted second-order polynomial model for GA recovery

48.034

0.663

72.49

0.000

C

-1.248

0.339

-3.68

0.004

P

2.988

0.339

8.81

0.000

t

3.614

0.339

10.66

T

0.923

0.339

Q

1.213

0.339

us

C2

-4.857

P2

-1.435

t2

-1.362

T2 Q2

cr

3.58

0.004

0.307

-15.83

0.000

0.307

-4.68

0.001

0.307

-4.44

0.001

-2.805

0.307

-9.15

0.000

-0.390

0.307

-1.27

0.229

0.243

0.415

0.58

0.571

te

d

an

0.020

Ac ce p

0.000

2.72

C×P

ip t

A0

M

Term Coefficient SE Coefficient t-value p-value

C×T

-0.007

0.415

-0.02

0.986

C×T

-0.182

0.415

-0.44

0.669

C×Q

-0.104

0.415

-0.25

0.807

P×t

-0.091

0.415

-0.22

0.830

P×T

0.365

0.415

0.86

0.409

P×Q

0.352

0.415

0.85

0.414

t×T

0.094

0.415

0.23

0.826

t×Q

0.193

0.415

0.46

0.652

T×Q

0.075

0.415

0.18

0.860

24 Page 24 of 35

Table 4: ANOVA of RSM modeling for GA recovery Source

Degree of freedom Sum of squares Mean square f-value p-value 20

1528.91

76.445

27.70

0.000

Linear

5

620.96

124.192

45.00

0.000

C

1

37.40

37.4

13.55

0.004

P

1

214.32

214.323

77.66

0.000

t

1

313.49

313.493

113.60

0.000

T

1

20.46

20.461

7.41

0.020

Q

1

35.28

35.284

12.79

0.004

Square

180.264

65.32

0.000

691.870

250.71

0.000

60.433

21.90

0.001

cr

ip t

Model

901.32

1

691.87

P2

1

60.43

t2

1

54.38

54.382

19.71

0.001

T2

1

230.85

230.852

83.65

0.000

Q2

1

4.47

4.469

1.62

0.229

2-Way Interactions

10

6.62

0.662

0.24

0.984

C×P

1

0.94

0.941

0.34

0.571

C×t

1

0.00

0.001

0.00

0.986

C×T

1

0.53

0.533

0.19

0.669

1

0.17

0.172

0.06

0.807

1

0.13

0.133

0.05

0.830

P×T

1

2.03

2.031

0.74

0.409

P×Q

1

1.99

1.988

0.72

0.414

t×T

1

0.14

0.141

0.05

0.826

t×Q

1

0.59

0.593

0.21

0.652

T×Q

1

0.09

0.090

0.03

0.860

Error

11

30.36

2.760

-

-

Lack of fit

6

29.46

4.911

27.55

0.001

Pure Error

5

0.89

0.178

-

-

Total

31

1559.26

-

-

-

C×Q

an

M d

Ac ce p

P×t

te

C

us

5

2

25 Page 25 of 35

Figure caption Fig. 1: Molecular structure of Glycyrrhizic acid Fig. 2: Schematic diagram of experimental setup for the supercritical extraction system; (1)

ip t

CO2 cylinder, (2) column consisting of molecular sieve and silica gel, (3) chiller, (4) pump head cooler, (5) HPLC pump, (6) needle valve, (7) spring coil preheater, (8) modifier vessel,

(14) extracted sample collection vessel.

us

Fig. 3: The effect of different type of entrainer on GA recovery

cr

(9) spring coil preheater, (10) extraction vessel, (11) oven, (12) needle valve, (13) BPR valve,

an

Fig. 4: The effects of temperature and pressure (coded values) on the GA recovery at CO2 flow rate of 1.4 ml/min, dynamic time of 80 min and co-solvent methanol concentration of

M

50% (a) response surface plot (b) contour plot (c) interaction plot

Fig. 5: The effects of extraction dynamic time and CO2 flow rate (coded values) on GA

te

contour plot (c) interaction plot

d

recovery at 22 MPa, 65 ◦C and 50% methanol concentration (a) response surface plot (b)

Fig. 6: The effect of methanol concentration in binary co-solvent on GA recovery at different

Ac ce p

operating conditions of C: (a) at several values of Q; (b) at different values of t; (c) for various values of T; (d) at different values of P.

26 Page 26 of 35

ip t cr us an M d te Ac ce p

Fig. 1

27 Page 27 of 35

ip t cr us an M d te Ac ce p

Fig. 2

28 Page 28 of 35

ip t cr us an M Ac ce p

te

d

Fig. 3

29 Page 29 of 35

ip t cr us an M d te Ac ce p Fig. 4

30 Page 30 of 35

ip t cr us an M d te Ac ce p Fig. 5

31 Page 31 of 35

32

Page 32 of 35

d

te

Ac ce p us

an

M

cr

ip t

ip t cr us

Ac ce p

te

d

M

an

Fig. 6

33 Page 33 of 35

 SC-CO2 was used for Glycyrrhizic acid (GA) extraction from Glycyrrhiza glabra root.  Response surface method was applied to optimize effective operating conditions.  54.4% GA recovery was obtained from licorice root at optimum operating conditions.

Ac ce p

te

d

M

an

us

cr

ip t

 Modified SC-CO2 with binary methanol-water entrainer was used for GA recovery.

34 Page 34 of 35

Ac

ce

pt

ed

M

an

us

cr

i

*Graphical Abstract (for review)

Page 35 of 35