Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts

Separation and Purification Technology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts Ai He, Bing Dong, Xueting Feng, Shun Yao ⇑ School of Chemical Engineering, Sichuan University, Chengdu 610065, China

a r t i c l e

i n f o

Article history: Received 14 March 2017 Received in revised form 19 May 2017 Accepted 23 May 2017 Available online xxxx Keywords: Ionic liquids Aqueous two-phase systems Tropine Quinoline Ginsenosides

a b s t r a c t In this research, two types of aqueous two-phase systems (ATPS) based on n-alkyl-tropinium and n-alkylquinolinium bromide ionic liquids (ILs) + salt + water have been developed and studied experimentally at different temperature. The binodal curves are fitted to nonlinear Merchuk relationship. For both investigated ILs, the longer the cation alkyl chain is, the greater its ability for ATPSs formation is; meanwhile the ability of the potential salts used for phase separation was compared and sorted, and high temperature was not beneficial to form these tropinium/quinolinium ILs–based ATPSs. On the basis of above results, the new method of extraction for ginsenosides from crude extracts of Panax Ginseng C. A. Mey using ILATPS was developed for the first time. Under appropriate conditions, a relatively high extraction efficiency (99.5%), partition coefficient (K = 651) and selective enrichment for total saponins could be obtained. Furthermore, the target ginsenosides can be recovered from the IL-rich phase and the enrichment is conducive to improve their antioxidant activity. Above research is expected to provide meaningful reference for the separation of similar bioactive compounds. Ó 2017 Published by Elsevier B.V.

1. Introduction The root of Panax ginseng C. A. Mey is a very famous traditional medicine in East Asian countries with immunopotentiation, cardio-protective, antidiabetic and anthypnotic effects [1–3], which has been used to treat various human diseases for thousands of years. One of its main bioactive fractions is composed of dammarane-type triterpene saponins including ginsenosides Rg1, Re, Rf, Rb1, Rc, Rb2, Rb3 and Rd, etc. Generally, there are a large amount of coexisting non-saponins in the crude extracts of ethanol-water, which inevitably increase the burden of reprocessing for obtaining pure ginsenosides. These constituents have been found as vitamins, amino acids, proteins, peptides and (poly)saccharides, which have similar polarity, solubility and other prosperities with target saponins. Macroporous resin [4], solid phase extraction [5], expanded bed adsorption [6], hydrophilic interaction chromatography (HILIC) [7], centrifugal partition chromatography (CPC) [8], high-speed counter-current chromatography (HSCCC) [9] and supercritical CO2 technology [10] have been employed to solve the separation problem of saponins . It is still expected for more and more efficient methods using in their

⇑ Corresponding author.

selective extraction and enrichment, which can provide researchers more new options. Aqueous two-phase system (ATPS) is one of the new separation methods to extract and purify bioactive substances, which can be operated continuously in a mild condition and enlarged easily [11–13]. The selective partition of a target product between the two phases is the basis of a two-phase system separation. As a new class of purely ionic salt-like materials, ionic liquids (ILs) are low melting-point materials composed entirely of ions and have negligible volatility, low flammability, chemical stability, good environmental benignity and comprehensive solubility for organic compound. A large number of possible variations in cation and anion allow the fine-tuning of their properties [14–16]. So ILs can be effectively designed for various applications, especially for ATPS as a non-aqueous phase. Compared with traditional aqueous two-phase systems, IL-ATPS combines the advantages of both efficiently. Gutowski and his co-workers first used hydrophilic ionic liquid 1-butyl-3-methyl imidazolium chloride ([Bmim]Cl) and K3PO4 to form aqueous two-phase system with IL-rich upper phase and salt-rich lower phase [17]. This system could achieve high product purity as well as high yield, while maintaining the biological activity of the molecules. Up to now, versatile applications of ILs have emerged in ATPS for enrichment and purification of active ingredients from diverse natural products [18–23]. The results have already proven that the investigated ILs will not affect

E-mail address: [email protected] (S. Yao). http://dx.doi.org/10.1016/j.seppur.2017.05.041 1383-5866/Ó 2017 Published by Elsevier B.V.

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

2

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

significantly structural integrity and activity of study objects. Moreover, as a kind of ‘‘green” solvent, ionic liquids could minimize environmental impacts resulting from the use of volatile organic compounds and could be recovered through the addition of a second hydrophobic IL or ion-exchange resins [24–26]. On the other hand, phase diagrams and phase equilibrium data are essential for the development, optimization, and scale-up of extraction process using ATPS [27]. Liquid-liquid equilibrium (LLE) data for imidazolium/pyridinium ionic liquids composed of 1-alkyl-3-methylimidazolium ([Cnmim]+) or n-alkyl-pyridinium

([CnPy]+) as cation and BF
4 , Br , Cl as anion have been measured by researchers [17,28–33]; and various salts including Na2CO3, K2CO3, K3PO4, KOH, KCl, NaCl, K2HPO4, and Na2HPO4 are frequently used to form IL-ATPS. According to preliminary study, a series of ionic liquids with tropine/quinoline as cationic nucleus and bromide as anion were selected in the following IL-ATPS study. The former represents the series of saturated alicyclic ILs and the latter is the typical representative of planar conjugated ILs, and both of them are seldom applied in aqueous two-phase systems for separation of saponins. Moreover, their bromine salts were found to have better phaseforming ability according to preliminary experiments. So they were firstly synthesized and used to form ATPSs with K3PO4, K2CO3, K2HPO4, K3C6H5O7, NaH2PO4, Na3C6H5O7 and water. Their phase diagrams, liquid–liquid equilibrium and tie-line data were measured for the first time, and obtained results are necessary for the optimization and design of extraction processes and provide essential basis for the prediction of phase composition when related data are not available. Finally, the new systems were used in extraction for ginsenosides from crude extracts of Panax Ginseng C. A. Mey. Appropriate conditions, extraction efficiency together with the results of separation and bioactivity assay were introduced in detail. 2. Experimental 2.1. Reagents and materials Tropine was purchased from Wenhua Chemical Co. Ltd (Zhengzhou, China); commercial cation-exchange resins were provided by Aladdin Chemical Co. Ltd (Shanghai, China), and their H/ Na type was transformed through the reported method [34]; quinoline, n-ethyl/propyl/butyl/amyl/hexyl bromide and the other salts/reagents were purchased from Kelong Chemical Co. Ltd (Chengdu, China). Except for chromatographic grade acetonitrile used for quantitative analysis, all reagents and solvents were analytical pure grade and were used without further purification if not stated otherwise. Crude powders of Ginseng extracts were selfmade, which was obtained by the extraction of dried roots of Panax ginseng C. A. Mey (source area: Jilin province of China, provided by Sengfu Biotech Co. Ltd, Xi’an, China) with 90% ethanol and concentration under vacuum according to previous study [35]. Standard compounds of Ginseng saponins (Rg1, Re, Rb1 and Rd) and 2,2diphenyl-1-picrylhydrazyl were purchased from Mansite BioTechnology Co., Ltd (Chengdu, China), and all of their purity was above 98.5%. Deionized water was obtained by the UPR-I-5T water purification system (0.4 mm filter) from Ulupure Technology Co. Ltd (Chengdu, China). 2.2. Synthesis for two series of ionic liquids Based on the previous study [36], tropine (7.05 g, 0.05 mol) was dissolved in 70 mL ethyl acetate (for n-ethyl, propyl, butyl bromide) or toluene (for n-amyl bromide), and then n-alkyl bromide (0.06 mol) was added into the solution. The reaction system was

refluxed at 323 K under vigorous stirring for 12 h (for n-ethyl, propyl bromide), or 348 K for 24 h (for n-butyl bromide), or 353 K for 24 h (for n-amyl bromide). At the end of the reaction, the mixture was filtered under reduced pressure to remove the solvent and the residual alkyl bromide. The product was washed with ethyl acetate/toluene for several times, and then the white solids were dried under vacuum at 323 K and recrystallized to give n-alkyltropinium bromide. Similarly, quinoline (6.45 g, 0.05 mol) and n-alkyl bromide (0.055 mol) were mixed in a 100 mL one-necked flask and reacted in dark under vigorous stirring for 2 days at 333 K (for n-ethyl, propylbromide) or 343 K (for n-amyl, hexyl bromide). Then the reaction products were dissolved in ethyl acetate to extract the unreacted quinoline and n-alkyl bromide for several times. Finally, the purple solids were dried under vacuum at 313 K for several days and n-alkyl-quinolinium bromide ionic liquids were obtained after recrystallization [37]. The purities of above ILs were finally analyzed by a highperformance liquid chromatography (HPLC) system, which was composed of a LC6A isocratic pump (Shimadzu, Japan), a 2000ES evaporative light scattering detector (Alltech, USA) and an in-line Rheodyne injection valve with a 20 lL sample loop. The LC column was a Waters symmetry C18 column (150 mm  3.9 mm, I.D., 5 lm), and mobile phase was acetonitrile-water solution (12: 88  27:73, V/V; no buffer added) with the flow rate of 1 mL/ min. As the result, their purities were determined as 98.2  99.0% mass fraction. In order to eliminate the influence of water and obtain accurate data, these ILs were thoroughly dried at 105 °C for 8 h before the following measurements. 2.3. Phase diagrams The binodal curves were determined by the titration method (cloud point method) [38]. An analytical balance with a standard uncertainty of ±0.0001 g was used to weigh all the samples. IL aqueous solution of known mass fraction was placed into the glass vessel and stirred vigorously, and the salt solution of known mass fraction was added dropwisely until the mixture became turbid or cloudy. The maximal standard uncertainty of the IL and salt in the compositions was estimated to be ±0.003 in mass fraction. Then drops of deionized water were added continuously into the vessel to form a clear one-phase system, and the procedure was repeated until the system cannot become turbid or cloudy again. Finally, the vessel was immersed in a jacketed glass vessel, and the temperature of the system was kept at 298.15 K, 308.15 K and 318.15 K using a CY20A water thermostat (Boxun Industry & Commerce Co., Ltd., Shanghai, China) in extrinsic cycle pattern with a standard uncertainty of ±0.05 K. The composition of the mixture for each cloud point on the binodal curve was calculated by mass using an electrical analytical balance. The data were fitted according to the empirical nonlinear equation developed by Merchuk [39]:

  0:5 w1 ¼ aexp bw2
cw32

ð1Þ

where w1 and w2 are the mass fractions of ILs and salt, respectively; and a, b, c are fitting parameters. 2.4. Tie-lines After binodal curves are determined, a mass fraction of ILs and salt was chosen which could form two phases for each ILs, salts and temperature. Then related systems of this mass fraction were built for different ILs, salts and temperature. Corresponding vessels filled with these systems were placed into the thermostatic bath (T = ±0.05 K) at a certain temperature for 12 h until equilibrium.

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

3

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

Both top phase (mt) and bottom phase (mb) were weighted respectively. Each individual tie-line was determined by application of the lever rule to the relationship between the top mass phase composition and the overall system composition [40]. Meanwhile, each experimental binodal curve was fitted using Eq. (1) as previously described. On the basis of Eq. (1) and lever rule, the equilibrium compositions were calculated by MATLAB software using Eqs. (2) (5) as follows [41]:

wt1 ¼ aexp½bðw2 t Þ

0:5


3
c wt2 

ð2Þ

0:5


3
c wb2 

ð3Þ

wb1 ¼ aexp½bðw2 b Þ w1 t
w1 mb ¼ mt w1
wb1

ð4Þ

w2 t
w2 mb ¼ mt w2
wb2

ð5Þ

where wt1, wb1, wt2, and wb2 represent the equilibrium compositions (percentage in mass fraction) of IL (1) and salt (2), in the top (t), and bottom (b) phases respectively; w1 and w2 represent the total compositions (percentage in mass fraction) of IL (1) and salt (2) respectively. After above values were determined, the tie line length (TLL) and the slope of the tie line (S) at different compositions and temperatures could also be calculated respectively using Eqs. (6) and (7) as follows: 2

2 0:5

TLL ¼ ½ðw1 t
wb1 Þ þ ðw2 t
wb2 Þ 

ð6Þ

S ¼ ðw1 t
wb1 Þ=ðw2 t
wb2 Þ

ð7Þ

2.5. Extraction of ginsenosides with ATPS A certain amount of crude extract powders of ginsenosides were dissolved in water under stirring, and then a certain amount of IL and salt were added into this solution. Different temperature, extraction time, type of salts, ILs and mass fraction of ginseng extract + IL + salt were investigated to obtain higher partition coefficient (K) and extraction efficiency (E, %), which can be calculated using Eqs. (8) and (9) as follows:

E¼

Ct  V t  100% Ct  V t þ Cb  V b

ð8Þ

K¼

Ct Cb

ð9Þ

where Ct and Cb (mg/mL) represent the concentrations of total ginsenosides in the top (t) and bottom (b) phases respectively. Vt and Vb (mL) represent the volumes of the top (t) and bottom (b) phases respectively. The standard uncertainty of the mass fraction of ginseng extract in the sample was ±0.002. The concentrations of individual ginsenosides were determined by HPLC method. Detailedly, the chromatographic analysis was performed on a 250  4.6 mm Diamonsil C18 column packed with 5 lm particles (Dikma Technologies Inc., Lake Forest, USA) at 25 ± 1 °C. The mobile phases consisted of A (acetonitrile) and B (0.1% phosphoric acid aqueous solution). The program of gradient elution was as follows: 0  35 min, 19% A; 35  55 min, 19  29% A; 55  70 min, 29% A; 70  100 min, 29  40% A. All of solvents were filtered through 0.22 lm filter prior to use. The flow rate was constantly kept at 1.1 mL/min; the sample injection volume was 15 lL and the detection wavelength was 203 nm. The chromatogram of standard mixture solutions and crude extracts of Ginseng is shown in

Fig. 1. The HPLC chromatograms of the standard mixture solution (a) and Ginseng extracted sample (b).

Fig. 1. The standard curves for individual ginsenosides were constructed by plotting the peak area (y) and the concentration of standard substance (x, mg/mL). The linear regression equations, linearity ranges and correlation coefficients (R2) are shown in Table 1. Their contents in crude extracts were determined as 7.47% (Rg1), 19.42% (Re), 9.56% (Rd) and 2.68% (Rb1) with a standard uncertainty of ±0.02%, respectively. Determination of total saponins content in extracts was based on the reported UV–visible spectroscopic method at 555 nm on a TU-1810 spectrometer (Purkinje General Instrument Co., Ltd., China) [41], and related linear equation was obtained as y = 0.00374x
0.00449 (linear range: 28.8–360 lg, R2 = 0.9993); the amount of extracted ginsenosides in IL phase was calculated by deduction of the residual amount in water phase from total amount before extraction.

Table 1 The standard curves of three ginsenosides. Ginsenoside

Regression equation

Linearity range (mg/mL)

R2

Rg1 Re Rd Rb1

y = 5156.273x + 18.927 y = 3783.266x + 15.041 y = 3763.686x + 21.730 y = 3826.014x + 18.141

0.0097–0.970 0.0099–0.990 0.0103–1.030 0.0099–0.990

0.9999 0.9999 0.9998 0.9999

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

4

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

2.6. Antioxidation assay 0.1 mg/mL was selected as the initial concentration of 2,2diphenyl-1-picrylhydrazyl (DPPH) radicals for all of the elementary radical reactions in the study. Certain amount of crude extract or total ginsenosides extracted by ATPS was dissolved by ethanol and then 4 mL DPPH ethanol solution and these sample solutions with the same volume were mixed and kept in the dark for 30 min at ambient temperature before analysis. The decrease in absorbance at 517 nm was measured against a blank of pure ethanol to estimate the radical scavenging capacity of each antioxidant sample, and here Vitamin C was selected as the standard reference. DPPH radical scavenging capacity was calculated with Eq. (10) [42].

DPPH Radical scavenging capacityð%Þ ¼ ½ðA0
A1 Þ=A0   100 ð10Þ where A0 and A1 are the UV–vis absorbance values of the DPPH radical solution before and after the addition of the sample, respectively. All tests were run in triplicate and the average value was calculated.

3. Results and discussion 3.1. Study of phase behaviors The binodal curve data for n-alkyl-tropinium bromide ionic liquid(1) + salt(2) + water(3) ATPSs are represented in Fig. 2(a), (c) an d (e), meanwhile those for n-alkyl-quinolinium bromide ionic liquid(1) + salt(2) + water(3) ATPSs are shown in Fig. 2(b), (d) and (f). And the parameters for the Merchuk equation were determined by least-squares regression of the cloud point data were summarized in Tables 2–4. Compared with those semi-empirical and semitheoretical thermodynamic models just like Pitzer and Cabezas, Merchuk empirical equation is simpler in form and has a broader range of applications, which has been successfully used to correlate binodal curve data of polymer/salt or ionic liquid/salt ATPS in recent years [43]. On the basis of the obtained R2 and SD, it can be concluded that Eq. (1) shows enough satisfactory accuracy in binodal data fitting for the investigated systems. Based on the results in Fig. 2(a), it can be concluded the binodal curve will be closer to the axis when the alkyl chain on cations of nalkyl-tropinium bromide ionic liquids is longer. In general, over high hydrophilic nature of the IL is not beneficial for the formation of ATPS. The appropriate increase in the length of non-polar alkyl chain can lead to a strengthened hydrophobic nature of ILs and therefore result in a weaker affinity with water [44], which make ILs need less salt to promote separation of two phases. Under this condition the binodal curve is closer to the axis and a larger biphasic region can be achieved. Above results indicate the ILs with larger molar volume and longer alkyl chain on their cations have stronger ability for ATPS formation, which is ruled by entropic contributions as a direct result of the formation of water
ion hydration complexes and the increased surface tension of water. As for n-alkyl-quinolinium bromide ionic liquid, it can be observed in Fig. 2(b) which shows the similar trend with the series of tropinium ILs. The result accords with the finding of S.P.M. Ventura and J.A.P. Coutinho et al. [45] when they compared ILs with aromatic and aliphatic cores, no major differences could be observed demonstrating that the H-bonding of different cations with H2O was not the driving force during the formation of ATPS. Related parameters of the quinolinium ILs for Eq. (1) were listed in Table 2, and the results also can be well fitted with Merchuk equation (all of R2 > 0.99).

Temperature can change the miscibility of different phases and have influence on the ATPS formation. In general, lower temperature is favorable to form IL-ATPS, which is because the higher the temperature, the higher are the IL and salt concentrations required for phase separation. Here the results of its effect are shown in Fig. 2(c) and (d) in the range from 298.15 K to 318.15 K, and related values of parameters of Eq. (1) for ILs + salts + water at different temperature are shown in Table 3. The effect of temperature is very similar in the study for the two different kinds of IL-ATPSs. The distance to axis and the proportion of biphasic region demonstrate that the two-phase area is indistinctively expanded with a decrease in temperature, and the reason is that a little decrease in solubility of ILs makes phase-forming ability for this ATPS system improved to a small extent, at this time the solvation of [CnTr]Br and [CnQn]Br was weaken and the ionic interaction was enhanced. On the other hand, hydrogen bond weakening with the increase of the temperature led to weakening of salt solvation. More meaningfully, phase behavior is relatively stable in these ATPSs following the temperature change within this range, which is very beneficial for their potential application in separation field because of good thermal adaptability. In the reported study both inorganic and organic salts have been used in the ATPS, and potassium is used more frequently than sodium between these two most popular salts. Moreover, the larger the two-phase region is, the stronger ability of the salts to induce the ATPS formation has. The effects of ions on biological and chemical processes in aqueous solution usually depend on the particular ions involved. These specific ion effects have been analyzed and summarized by researchers, just like Hofmeister phenomena/Hofmeister series. Here, the binodal data obtained at 298.15 K and atmospheric pressure for various salts (K3PO4, K2CO3, K2HPO4, K3C6H5O7, NaH2PO4 and Na3C6H5O7; KH2PO4 and Na2HPO4 were not selected for poor solubility) are shown in Fig. 2(e) and (f), and related values of parameters of Eq. (1) for ILs + salts + water at 298.15 K are shown in Table 4. In particular, n-alkyl-quinolinium bromide ionic liquids are found unstable in K3PO4 and K2CO3 solution of high concentration, so the two salts were not applied in their systems. By comparison, n-alkyltropinium bromide ionic liquids bear more universal applicability for better chemical stability. As can be observed, the ability of the salts investigated in their ATPSs for phase separation follows the order: K3PO4 > K2HPO4 > K2CO3 > Na3C6H5O7 > K3C6H5O7, and the performance of NaH2PO4 is a little unique; for ATPSs of n-alkyl-quinolinium bromide ILs, it follows the order of K2HPO4 > Na3C6H5O7 > NaH2PO4 > K3C6H5O7. So the sequence of these three kinds of salts is identical in two series of ATPSs. In potential molecular mechanism, above investigated salts (except Na3C6H5O7) share a common cation but contain different anions, so it is easily concluded the ability of anions for ATPS formation follows the 2
2
3
order of PO3
4 > HPO4 > CO3 > C6H5O7 , which is same as the conclusions in reported studies [46,47] and closely related with their Gibbs free energy of hydration (the higher absolute value of DhydG is, the stronger salting-out ability is). The interaction between salt and water becomes the driving force for ATPS formation. As for two different cations, sodium (Na3C6H5O7) has better salting-out than potassium (K3C6H5O7) for ATPS formation, which is accorded with Hofmeister’s rule and previous findings [48,49]. 3.2. Tie-lines The comparison of tie-lines can be made through the data in Tables 5–7, which show the effects of alkyl chain length on IL cation/temperature/salts on the phase compositions. It can be concluded that the formation of ATPSs is closely related with both the polarity/salting-out ability of IL and salt ions. Moreover, the obtained data could prove that the tie-line of IL-ATPS conforms

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

5

Fig. 2. Effect of IL alkyl chain of [CnTr]Br + K2CO3 + water at 298.15 K (a) and [CnQn]Br + K2HPO4 + water at 298.15 K (b), temperature of [CnTr]Br + K2CO3 + water (c) and [CnQn]Br + K2HPO4 + water (d) (h: 298.15 K, s: 308.15 K, 4: 318.15 K) and salts in [CnTr]Br (e) or [CnQn]Br (f) + salt + water at 298.15 K (h: K3PO4, s: K2HPO4, 4: K2CO3, 5: Na3C6H5O7, }: K3C6H5O7, q: NaH2PO4) on binodal curves.

to the lever rule, and applying the Merchuk equation to IL-ATPS reproduces our experimental results satisfactorily. Here the tielines of [C2Tr]Br or [C2Qn]Br + salt + water were taken for two examples, which were depicted as a schematic diagram in Fig. 3 (a) and (b). Each point on tie-line has same composition (mass percentage of IL, salt and water) in top/bottom phase as other points on this tie-line, but different volume of top/bottom phase. It can be observed that the slope (absolute value) of tie-line increases with an increase in alkyl chain length or decrease in temperature. And the sequence of slope (absolute value) of tie-line follows the order as K3PO4 > K2HPO4 > K2CO3 > Na3C6H5O7 > K3C6H5O7 (see

Fig. 3(c)), which is same as the order of their ability for ATPS formation. The possible reason is that a longer alkyl chain /lower temperature for IL-salt ATPSs/stronger salting-out ability of salts can prompt water to transfer preferably from the IL-rich phase to the salt-rich phase; as the result the IL concentration at the IL-rich phase increases while the required amount of salts will be decreased. Moreover, this sequence of slope (absolute value) is the same as the order of phase-separation abilities of ILs/temperature/salts. As for TLL, it has a same order like that of slope and shows the difference from top phase to bottle phase; the larger TLL is, the greater discrepancy of two phases will have. Finally,

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

6

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

Table 2 Values of parameters of Eq. (1) for various ILs + salts + water under 298.15 K at atmosphere pressure (p = 0.1 MPa).a

a

n

a

b

c

R2

SD

[CnTr]Br + K2CO3 + water 2 3 4 5

99.5183 100.6027 101.7313 98.3039


0.2957
0.3264
0.3415
0.3394

3.20E-05 3.41E-05 4.63E-05 6.48E-05

0.9988 0.9997 0.9996 0.9995

0.4484 0.1856 0.2204 0.2943

[CnQn]Br + K2HPO4 + water 2 108.6858 3 101.2622 4 103.0390 5 241.1889 6 185.7237


0.3724
0.4044
0.4241
0.7472
0.8566

5.08E-05 6.26E-05 8.77E-05 5.25E-05 4.72E-05

0.9996 0.9991 0.9976 0.9978 0.9982

0.2557 0.3805 0.6165 0.3482 0.6228

Standard uncertainties u are u(T) = ±0.05 K and u(p) = ±2 kPa.

Table 3 Values of parameters of Eq. (1) for ILs + salts + water under different temperature at atmosphere pressure (p = 0.1 MPa).a

a

T

a

b

c

R2

SD

[C3Tr]Br + K2CO3 + water 298.15 K 308.15 K 318.15 K

100.6027 104.6782 101.2580


0.3264
0.3303
0.3125

3.41E-05 3.46E-05 3.29E-05

0.9997 0.9996 0.9996

0.1856 0.2870 0.3215

[C3Qn]Br + K2HPO4 + water 298.15 K 308.15 K 318.15 K

101.2622 106.9082 109.3714


0.4044
0.3972
0.3908

6.26E-05 5.10E-05 5.13E-05

0.9991 0.9993 0.9973

0.3805 0.3254 0.7092

Standard uncertainties u are u(T) = ±0.05 K and u(p) = ±2 kPa.

Table 4 Values of parameters of Eq. (1) for ILs + various salts + water under 298.15 K at atmosphere pressure (p = 0.1 MPa).a

a

Salt

a

b

c

R2

SD

[C3Tr]Br + salts + water K3PO4 K2CO3 NaH2PO4 K2HPO4 Na3C6H5O7 K3C6H5O7

95.4668 100.6027 65.0416 90.9875 85.5091 96.1365


0.3652
0.3264
0.1978
0.3316
0.2297
0.2399

7.14E-05 3.41E-05 1.39E-05 4.29E-05 3.43E-05 1.63E-05

0.9988 0.9997 0.9973 0.9992 0.9994 0.9991

0.4721 0.1856 0.6011 0.3244 0.2409 0.3282

[C3Qn]Br + salts + water NaH2PO4 K2HPO4 Na3C6H5O7 K3C6H5O7

113.8052 101.2622 92.6187 85.8769


0.3225
0.4044
0.2881
0.2318

2.64E-05 6.26E-05 5.05E-05 2.90E-05

0.9988 0.9991 0.9997 0.9934

0.3997 0.3805 0.1509 1.2611

Standard uncertainties u are u(T) = ±0.05 K and u(p) = ±2 kPa.

Table 5 Tie line data of ILs + salts + water under 298.15 K at atmosphere pressure (p = 0.1 MPa).a,b w1b

wt2

w2b

S

TLL

[CnTr]Br + K2CO3 + water 2 61.635 3 63.225 4 64.799 5 77.382

3.363 2.452 1.181 1.301

2.620 2.023 1.743 0.497

36.795 37.012 37.156 33.193


1.705
1.737
1.797
2.327

67.555 70.126 72.810 82.809

[CnQn]Br + K2HPO4 + water 2 64.184 3 67.188 4 72.027 5 74.993 6 80.066

0.769 0.357 0.173 0.517 0.274

1.998 1.029 0.713 2.441 0.965

37.470 37.060 35.336 32.858 32.564


1.788
1.855
2.075
2.449
2.525

72.662 75.926 79.761 80.448 85.821

n

a b

wt1

Standard uncertainties u are u(T) = ±0.05 K, u(w) = ±0.003 and u(p) = ±2 kPa. Standard uncertainty u was calculated using standard deviation (SD).

compared with TLL values reported in imidazolium IL-sodium citrate/tartrate/acetate ATPS [50], greater mass fraction differences

of [CnTr]Br/[CnQn]Br and salt between the two phases led to higher values of TLL in this study (see Tables 5–7). It is obvious the result

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

7

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx Table 6 Tie line data of ILs + salts + water under different temperature at atmosphere pressure (p = 0.1 MPa).a,b

a b

T

wt1

w1b

wt2

w2b

S

TLL

[C3Tr]Br + K2CO3 + water 298.15 K 308.15 K 318.15 K

63.225 62.397 61.703

2.452 2.393 2.758

2.023 2.448 2.508

37.012 37.096 37.221


1.737
1.732
1.698

70.126 69.288 68.406

[C3Qn]Br + K2HPO4 + water 298.15 K 67.188 308.15 K 66.249 318.15 K 65.677

0.357 0.732 0.760

1.029 1.450 1.702

37.060 36.935 36.974


1.855
1.846
1.840

75.926 74.509 73.881

Standard uncertainties u are u(T) = ±0.05 K, u(w) = ±0.003 and u(p) = ±2 kPa. Standard uncertainty u was calculated using standard deviation (SD).

Table 7 Tie line data of ILs + salts + water under 298.15 K at atmosphere pressure (p = 0.1 MPa).a,b

a b

T

wt1

w1b

wt2

w2b

S

TLL

[C3Tr]Br + salts + water K3PO4 K2CO3 NaH2PO4 K2HPO4 Na3C6H5O7 K3C6H5O7

67.252 63.225 48.265 65.403 60.452 46.935

0.263 2.452 2.035 1.243 3.188 10.133

0.920 2.023 2.273 0.991 2.275 8.668

37.178 37.012 52.659 37.504 37.950 36.591


1.848
1.737
0.918
1.757
1.605
1.318

76.172 70.126 68.381 73.822 67.468 46.196

[C3Qn]Br + salts + water NaH2PO4 K2HPO4 Na3C6H5O7 K3C6H5O7

55.492 67.188 66.115 54.814

3.511 0.357 1.314 2.671

4.918 1.029 1.368 3.725

38.290 37.060 36.760 40.922


1.558
1.855
1.831
1.402

61.771 75.926 73.836 64.051

Standard uncertainties u are u(T) = ±0.05 K, u(w) = ±0.003 and u(p) = ±2 kPa. Standard uncertainty u was calculated using standard deviation (SD).

Fig. 3. Tie-lines of [C2Tr]Br + K2CO3 + water (a) and [C2Qn]Br + K2HPO4 + water (b) at 298.15 K (j: tie-line, s: binodal curve) and comparison among salts for the sequence of slope (absolute value) of tie-lines (c).

is more beneficial for the recovery of IL. In consequence, the concentrations of the ions become higher with the increase of TLL and there is be less water to hydrate the ions completely, thus cation–anion interactions begin to dominate. 3.3. Separation behavior of ginsenosides in under various IL-ATPS conditions 3.3.1. Comparison of various ILs In the separation application for target constituents, the requirement for ATPS should include not only easy formation, but also ideal extraction efficiency. According to previous studies, the extraction performance of IL is determined by their dissolving capacity for target compounds, meanwhile it also can be affected by the cluster phenomenon originating from IL itself or combined with the extracted objects [51]. As known to all, various ILs have different dissolving and cluster behaviors. In order to find out the

optimal IL between the series of [C2
4Tr]Br and [C2
5Qn]Br to extract total ginsenosides from crude powders of Ginseng saponins, the partition coefficients (K) and extraction efficiency (E) mentioned in Section 2.5 were compared in the same ATPS composition of 40 wt% IL + 15 wt% NaH2PO4 + 1 wt% ginseng extract + 44 wt% water. It should be noted that K3PO4 and K2HPO4 were replaced by NaH2PO4 here because they had not ideal solubility in the investigated separation system. Obviously, the addition of ginsenosides increases the complexity of the system. The comparison results were showed in Fig. 4(a). According to the variation of both partition coefficients and extraction efficiency obtained by different IL-ATPSs, it can be observed that [C4Tr]Br is the optimal IL, and the tropine-type ILs have stronger ability than quinoline-type ILs to extract total ginsenosides no matter how long alkyl chain is on their cation. It is probably because that tropine-type and quinoline-type ILs have different multi-interactions or affinity with ginsenosides.

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

8

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

Fig. 4. Comparison on extraction results (a) of various ILs (4: [CnTr]Br, 5: [CnQn]Br) and H-bond interactions between IL and ginsenoside (b).

Generally, H-bond and hydrophobic effects are two main interactions existing between these saponins and potential agents because the former is always consisted of nonpolar terpene polycyclic system and polyhydroxy sugar chains. Hydroxyl group in tropinum cation can interact with ginsenosides through hydrogen bonding and quinolinium cation has not such kind of strong interaction. In order to discover the intermolecular mechanism, a molecular simulation of two ionic liquids and saponin molecular was performed with molecular dynamics method by Hyperchem 8.0 software to investigate the interactions and sites between them. The stereo structures of the molecules were firstly obtained with drawing tools, and then energy minimization of related structures was carried out by MM+ calculation and molecular dynamics method. Finally, the micro molecules were taken for the docking procedure and the Ground State method was adopted to optimize the molecular. It can be proved from the three-dimensional diagram (see Fig. 4(b)) that H-bonding exist between tropine nuclear and ginsenosides when the combination of [C2Tr]Br and Rg1 was taken as a example; and a-OH on C-3 of cation can interact with the hydroxyl group on C-6 of glucose and anion interacts with b-OH on C-3 of aglycone. In the previous study for the dissolution mechanism of cellulose in ionic liquids, C-6-OH on sugar chains was also inferred as the interaction site with IL cation [52]. Furthermore, in the comparison for the performance of [C2
4Tr] Br, when the length of the alkyl chain become longer, the level of both K and E are improved simultaneously. It is because that [CnTr]Br with shorter alkyl chain has stronger surface tension and polar interaction, which result in difficulty for extracting ginsenosides into IL-rich phase. 3.3.2. Selection for sample and IL concentrations Firstly, five different concentrations of crude extracts have been studied for extraction performance. As shown in Fig. 5(a), high extraction efficiency (E) above 99% can be achieved under each concentration in the range from 1 to 5 wt%. With the increasing of sample concentration, K of the total ginsenosides becomes higher significantly until the concentration reaches 3 wt%, which can reflect the migration ability of these saponins between the two phases and the separation performance of extractant from another side. Higher initial sample concentration promotes the ginsenosides into the IL-rich phase, which has the stronger affinity interaction with them. Because the amount of IL is given, K would not increase further. Meanwhile, the effect of IL concentration on extraction of ginsenosides could be observed from the results in Fig. 5(b). It is distinct that 35 wt% IL had optimal effect for extracting ginsenosides. Higher IL concentration leads to higher K value,

but overhigh IL concentration make no contribution for improvement of K because of the reducing of ginsenosides concentration in IL-rich phase when the amount of ginsenosides remains unchanged. 3.3.3. Investigation of salt type and concentration Different salts could influence the distribution of IL in top/bottom phase, meanwhile K and E of target constituents would be changed. As mentioned in Section 3.3.1, some potential salts cannot be considered in the separation process because of poor solubility in ATPS after the addition of extracts, such as K3PO4 and K2HPO4. Among those remaining, it is obvious that NaH2PO4/ K2CO3 in ATPSs have more ideal K and E than sodium citrate and potassium citrate according to Fig. 5(c). And NaH2PO4 has better solubility in ATPS than K2CO3, so the former is chosen in further investigation. Besides the salt type, its concentration also can play an important role. In ATPS, more salt always concentrates in bottom phase, and higher concentration of salt will have more influence on the existence of IL and ginsenosides in bottom phase for it can increase the hydrophobicity of this layer. The larger number of salt ions competes for H2O molecules with target saponins while more salvation spheres surrounding the ginsenosides groups are removed and stronger salting-out effect takes place. As the result, the IL-ginsenosides complex will be driven from the salt-rich phase to the IL-rich phase. So it could be observed in Fig. 5(d) that K value is significantly improved with the increase of salt concentration. Nonetheless, NaH2PO4 could not dissolve totally when its concentration reaches 22.5 wt.%. In the investigated range of concentration, extraction efficiency can be maintained at a very high level all the time. 3.3.4. Study on extraction time and temperature The partition coefficients (K) and extraction efficiency (E) that change over time were acquired from 30 min to 3 h and presented in Fig. 5(e). Obviously, the whole process includes two stages, and the first is rapider than the second; 60 min is enough to extract total ginsenosides, and main separation can be completed within 30 min. Little research of extraction kinetics was focused on in the those reported literatures of IL-ATPS, here it was investigated for further understanding the rapid stage of separation process (within the first 15 min). The results indicated that the relationship of initial concentration (C0), the concentration in IL phase at specific time (Ct) and time (t) met the following equation: ln C0 – ln Ct = kt, k (representative rate factor) = 1.27  10
3 s, R2 = 0.9997. So the rapid extraction process follows first order kinetics and concentration difference between two phases is the

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

9

Fig. 5. Effects of the initial sample concentration (a), IL concentration (b), salt type (c), salt concentration (d), extraction time (e) and temperature (f) on separation performance.

main driving force. Furthermore, the results of effect of temperature on partition behavior of ginsenosides in IL-ATPS are shown in Fig. 5(f). When temperature becomes higher, the K value changes slightly. It can indicate that temperature is not a significant factor influencing extraction, which means this kind of ATPS is not extremely thermosensitive. The studied ATPS was suitable for enrichment of ginsenosides at a relative wide range of temperature. Finally, under the ideal extraction conditions (35 wt% [C4Tr]Br + 20 wt% NaH2PO4 + 3 wt% ginseng extracts + 42 wt% water, 1 h, room temperature), total E (Extraction efficiency, 99.5%) and K (Partition coefficient, 651) for total saponins were obtained. Here it should be noted that Rb1 was selected as one of representative studied saponins because its content was very low according to the HPLC chromatogram. Besides those main compounds, the study was also aimed to investigate the extraction performance of new ATPS for the saponins with low content. Moreover, the number of sugar units in its structure is more than that of the other

three ginsenosides, the difficulty for its extraction from aqueous phase should be greater because of higher polarity and water solubility. As the result, ginsenoside Rb1 was enriched by above ATPS with satisfied E% value together with the other three major saponins, which could reach to 99.1%; total content of ginsenosides in the extracts increased from 48.3% to 97.9%, which indicated the system had excellent selective separation ability for saponins and most of coexisting constituents (mainly including vitamins, amino acids, peptides, and saccharides) have been removed from IL-rich phase. It can be used as a useful technology with high specificity for ginsenosides and their analogues. 3.4. Separation of saponins from ATPS system In the similar studies, the reprocessing of target compounds is always overlooked. Considering the recovery of saponins and reusing of IL, the feasible separation way for them was explored in the

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

10

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx

section. Previously, Clark-Lubs buffer solution (KH2PO4-NaOH) with appropriate pH was used in the reuse of hydrophilic imidazolium ionic liquid after the extraction of tetracycline with its ATPS [53], and the back-extraction with organic solvent was also considered because they could be easily removed for the low-boiling point. However, the extraction with n-butanol, ethyl acetate or chloroform was found to be low efficiency after our repeated experiments for both of saponins and IL have similar polarity and solubility, and the heavy use of volatile organic extractant would affect the ‘‘green” feature of the separation process. Finally, ionexchange static adsorption method was selected and various resins (732H/Na type, AmberjetÒIMAC HP1110 H/Na type, AmberliteÒIR120 H/Na type) were compared under the same conditions as follows: solid-liquid ratio of 3:50 (g/mL), 288.15 K and 12 h. The upper phase after ATPS extraction was diluted by deionized water (dilution multiple = 100) and absorbed by these resins in constanttemperature shaker. Commonly, IL can be exchanged onto the sorbent and more saponins tend to remain in the solution. The experimental results indicated that (1) both [C4Tr]Br and ginsenosides cannot be absorbed on three kinds of Na type resins; (2) both [C4Tr]Br and ginsenosides can be adsorbed on 732 H type and HP1110 H type resins at different extent; (3) only ginsenosides can be selectively adsorbed on AmberliteÒIR-120 H type resin meanwhile the adsorption rate of [C4Tr]Br < 0.1%, which was found as a new function of this resin for the first time; so the last one was used to separate target saponins from upper phase of ATPS under above conditions, and IL still stayed in it. Furthermore, the suitable elution conditions were also explored, and 96.3% saponins could be desorbed by pure ethanol with solid-liquid ratio of 1:10 (g/mL) at room temperature. Through methanol also could result in high desorption efficiency (D, %), it was finally replaced with ethanol which has not obvious toxicity. Moreover, the D (%) of those frequently used eluents of inorganic acids (5  10% HCl or H2SO4) for ion-exchange resins was found lower than 5%, because adsorption and desorption mechanism of neutral ginsenosides was not based on the exchange of ions. 3.5. Activity analysis and comparison As known to all, some inappropriate separation methods can affect the bioactivity of target products, which will restrict their actual application to a great extent. Saponins from Panax ginseng play an important role in scavenging free radicals, improving myocardial ischemia and nerve cell protection by their antioxidant function. DPPH method is popularly used to evaluate free radical scavenging capacity for Ginseng saponins according to many previous reports, which was employed in the study to analyze the change of antioxidant activity of ginsenosides before and after ATPS extraction. Generally, the higher inhibition rate indicates the stronger antioxidant capacity of saponins. As the result, the inhibition rate of crude extract and enriched ginsenosides (both as ethanol solution of 7 mg/mL) were 51.86% to 67.40% with a standard uncertainty of ±0.03% after ATPS extraction, respectively. This result can prove the developed method does not weaken related activity but is beneficial for its improvement. Moreover, the latter (67.4%) is higher than that of total saponins (50  60%) enriched with n-butanol [54], which is regarded as the most commonly used method to extract saponins from the aqueous solution of crude extract; so it is another advantage of new IL-ATPS method compared with traditional ones. Finally, it is also found that the growth extent of antioxidant activity is lower than that of ginsenosides content, which indicates some other removed non-saponin constituents also have contribution to the activity. This conclusion is supported by previous researches [55–58], and these compounds with higher antioxidant activity have been mentioned in Section 3.3.1.

4. Conclusions This work shows that [C2
5Tr]Br/[C2
6Qn]Br are able to induce aqueous phase separation in the presence of several types of salts and thus to form ATPS. Phase diagrams and tie-lines at different temperature were also determined and presented. It is shown that ILs with longer alkyl chain have stronger ability for ATPS formation. And the order of ability of salts for forming IL-ATPS was as following: K3PO4 > K2HPO4 > K2CO3 > Na3C6H5O7 > K3C6H5O7. With a decrease in temperature, the phase-forming ability would increase in the investigated system. Meanwhile, the slope (absolute value) of tie-line had a similar rule to ability for ATPS formation. The stronger ability for ATPS formation is, the larger slope (absolute value) of tie-line is. After the investigation of phase behavior, this study also explored the extraction of ginsenosides from crude extracts of ginseng roots using IL-ATPS, and the optimal conditions (extracted with 35 wt% [C4Tr]Br + 20 wt% NaH2PO4 + 3 wt% ginseng extracts + 42 wt% water for 60 min under R.T.) were explored. The results can prove that IL-ATPS method is an effective approach for the ginsenosides (Rg1, Re, Rd and Rb1) extraction with a relatively high extraction efficiency above 99% and good selectivity. Recovery of saponins can be realized from IL-rich phase and the new extraction method is found beneficial for bioactivity improvement of ginsenosides. Above research is expected to provide meaningful reference for the separation of similar bioactive compounds. At present, the most important problem to be solved in the industrialization of aqueous two-phase extraction is the relatively high expense of ATPS system; meanwhile the separation between target product and phase-forming substance is not easy. So simplification of synthesis and recovery of IL will be expected in largescale application of this method in the future study. Currently, there is little research on system hydrodynamics and interphase mass transfer, and the present studies basically rely on the experimental method and most of the cases cannot be extended for lack of enough understanding of key phase-forming together with separation process. Finally, the integrated optimization of aqueous two-phase extraction system is still urgently needed by academia and industry including its design, development, after processing and theoretical calculation. Acknowledgements Preparation of this paper was supported by the National Natural Science Foundation of China (No. 81373284, 81673316). References [1] K. Sun, C.S. Wang, J. Guo, Y. Horie, S.P. Fang, F. Wang, Y.Y. Liu, L.Y. Liu, J.Y. Yang, J.Y. Fan, J.Y. Han, Protective effects of ginsenoside Rb1, ginsenoside Rg1, and notoginsenoside R1 on lipopolysaccharide-induced microcirculatory disturbance in rat mesentery, Life Sci. 81 (2007) 509–518. [2] C.Y. Yang, J. Wang, Y. Zhao, L. Shen, X. Jiang, Z.G. Xie, N. Liang, L. Zhang, Z.H. Chen, Anti-diabetic effects of Panax notoginseng saponins and its major antihyperglycemic components, J. Ethnopharm. 130 (2010) 231–236. [3] H. Dong, L.P. Bai, V.K.W. Wong, H. Zhou, J.R. Wang, Y. Liu, Z.H. Jiang, L. Liu, The in vitro structure-related anti-cancer activity of ginsenosides and their derivatives, Molecules 16 (2011) 10619–10630. [4] Y.N. Zhao, Z.L. Wang, J.G. Dai, L. Chen, Y.F. Huang, Preparation and quality assessment of high-purity ginseng total saponins by ion exchange resin combined with macroporous adsorption resin separation, Chin J. Nat. Med. 12 (2014) 382–392. [5] X.L. Shi, Y.R. Jin, J.B. Liu, H.Y. Zhou, W. Wei, H.Q. Zhang, X.W. Li, Matrix solid phase dispersion extraction of ginsenosides in the leaves of Panax ginseng CM Mey, Food Chem. 129 (2011) 1253–1257. [6] J.N. Mi, M. Zhang, H.Y. Zhang, Y.R. Wang, S.K. Wu, P. Hu, Coupling of ultrasound-assisted extraction and expanded bed adsorption for simplified medicinal plant processing and its theoretical model: extraction and enrichment of ginsenosides from Radix Ginseng as a case study, J. Sep. Sci. 36 (2013) 593–601.

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041

A. He et al. / Separation and Purification Technology xxx (2017) xxx–xxx [7] D. Yu, Z. Guo, A. Shen, J. Yan, X. Dong, G. Jin, Z. Long, L. Liang, X. Liang, Synthesis and evaluation of sulfobetaine zwitterionic polymer bonded stationary phase, Talanta 161 (2016) 860–866. [8] Y.C. Zhang, J.X. Zhang, C.M. Liu, M. Yu, S.N. Li, Extraction, isolation, and aromatase inhibitory evaluation of low-polar ginsenosides from Panax ginseng leaves, J. Chromatogr. A 1483 (2017) 20–29. [9] Y.C. Zhang, L.P. Guo, C.M. Liu, Y.J. Qi, D.P. Li, S.N. Li, Combination of pressurized liquid extraction with countercurrent chromatography for systematical isolation of ginsenosides from Panax ginseng, Sep. Purif. Tech. 116 (2013) 240–247. [10] I. Yu, Z. Yu, M. Koo, B. Wang, A continuous fractionation of ginsenosides and polysaccharides from Panax ginseng using supercritical carbon dioxide technology, J. Food Process. Pres. 40 (2016) 743–748. [11] P.A. Albertsson, Partition of cell particles and macromolecules, third ed., Wiley, New York, 1986. [12] M.T.H. Cavalcanti, T.S. Porto, B.B. Neto, J.L. Lima-Filho, A.L.F. Porto, A. Pessoa, Aqueous two-phase systems extraction of a-toxin from Clostridium perfringens type A, J. Chromatogr. B 833 (2006) 135–140. [13] H. Walter, D.E. Brooks, D. Fisher, Partitioning in Aqueous Two-phase Systems: Theory, Methods, Uses, and Applications to Biotechnology, Academic Press, Orlando, 1985. [14] A.F.M. Cláudio, M.G. Freire, C.S.R. Freire, A.J.D. Silvestre, J.A.P. Coutinho, Extraction of vanillin using ionic-liquid-based aqueous two-phase systems, Sep. Purif. Technol. 75 (2010) 39–47. [15] P. Sun, D.W. Armstrong, Ionic liquids in analytical chemistry, Anal. Chim. Acta 661 (2010) 1–16. [16] C.F. Poole, S.K. Poole, Extraction of organic compounds with room temperature ionic liquids, J. Chromatogr. A 1217 (2010) 2268–2286. [17] K.E. Gutowski, G.A. Broker, H.D. Willauer, J.G. Huddleston, R.P. Swatloski, J.D. Holbrey, R.D. Rogers, Controlling the aqueous miscibility of ionic liquids: aqueous biphasic systems of water-miscible ionic liquids and waterstructuring salts for recycle, metathesis, and separations, J. Am. Chem. Soc. 125 (2003) 6632–6633. [18] L. Liu, Y.S. Dong, Z.L. Xiu, Three-liquid-phase extraction of diosgenin and steroidal saponins from fermentation of Dioscorea zingibernsis C. H. Wright, Process Biochem. 45 (2010) 752–756. [19] X.Y. Wu, L.H. Liang, Y. Zou, T. Zhao, J.L. Zhao, F. Li, L.Q. Yang, Aqueous twophase extraction, identification and antioxidant activity of anthocyanins from mulberry (Morus atropurpurea Roxb.), Food Chem. 129 (2011) 443–453. [20] Y. Jun, W. Raza, Q.W. Huang, Q.R. Shen, Quantification of the antifungal lipopeptide iturin A by high performance liquid chromatography coupled with aqueous two-phase extraction, J. Chromatogr. B 879 (2011) 2746–2750. [21] K.S. Nascimento, P.A.J. Rosa, B.S. Cavada, A.M. Azevedo, M.R.A. Barros, Partitioning and recovery of Canavalia brasiliensis lectin by aqueous twophase systems using design of experiments methodology, Sep. Purif. Technol. 75 (2010) 48–54. [22] A. Haghtalab, B. Mokhtarani, The new experimental data and a new thermodynamic model based on group contribution for correlation liquid– liquid equilibria in aqueous two-phase systems of PEG and (K2HPO4 or Na2SO4), Fluid Phase. Equilibr. 215 (2004) 151–161. [23] S. Park, R.J. Kazlauskas, Biocatalysis in ionic liquids–advantages beyond green technology, Curr. Opin. Biotechnol. 14 (2003) 432–437. [24] X. Han, D.W. Armstrong, Ionic liquids in separations, Acc. Chem. Res. 40 (2007) 1079–1086. [25] A.B. Pereiro, A. Rodríguez, Azeotrope-breaking using [BMIM] [MeSO4] ionic liquid in an extraction column, Sep. Purif. Technol. 62 (2008) 733–738. [26] B.D. Ribeiro, M.A.Z. Coelho, L.P.N. Rebelo, I.M. Marrucho, Ionic liquids as additives for extraction of saponins and polyphenols from mate (Ilex paraguariensis) and tea (Camellia sinensis), Ind. Eng. Chem. Res. 52 (2013) 12146–12153. [27] J. Han, Y. Wang, C. Chen, W. Kang, Y. Liu, K. Xu, L. Ni, (Liquid + liquid) equilibria and extraction capacity of (imidazolium ionic liquids + potassium tartrate) aqueous two-phase systems, J. Mol. Liq. 193 (2014) 23–28. [28] C. He, S. Li, H. Liu, K. Li, F. Liu, Extraction of testosterone and epitestosterone in human urine using aqueous two-phase systems of ionic liquid and salt, J. Chromatogr. A 1082 (2005) 143–149. [29] A. Soto, A. Arce, M.K. Khoshkbarchi, Partitioning of antibiotics in a two-liquid phase system formed by water and a room temperature ionic liquid, Sep. Purif. Technol. 44 (2005) 242–246. [30] N.J. Bridges, K.E. Gutowski, R.D. Rogers, Investigation of aqueous biphasic systems formed from solutions of chaotropic salts with kosmotropic salts (salt–salt ABS), Green Chem. 9 (2007) 177–183. [31] Y. Deng, J. Chen, D. Zhang, Phase diagram data for several salt + salt aqueous biphasic systems at 298.15K, J. Chem. Eng. Data 52 (2007) 1332–1335. [32] M.T. Zafarani-Moattar, S. Hamzehzadeh, Liquid-Liquid equilibria of aqueous two-phase systems containing 1-butyl-3-methylimidazolium bromide and potassium phosphate or dipotassium hydrogen phosphate at 298.15K, J. Chem. Eng. Data 52 (2007) 1686–1692. [33] Y. Pei, J. Wang, L. Liu, K. Wu, Y. Zhao, Liquid–liquid equilibria of aqueous biphasic systems containing selected imidazolium ionic liquids and salts, J. Chem. Eng. Data 52 (2007) 2026–2031.

11

[34] A. He, B. Dong, X.T. Feng, S. Yao, Recovery of benzothiazoliumionic liquids from the coexisting glucose by ion-exchange resins, J. Mol. Liq. 227 (2017) 178–183. [35] L. Uhr, Y.C. Chen, D. Sit, P.C.H. Li, Ginsenosides in commercial ginseng products analyzed by liquid chromatography-tandem mass spectrometry, ISRN Anal. Chem. 2014 (2014), http://dx.doi.org/10.1155/2014/486842. [36] J. Lu, H. Song, Y. Yang, G. Qian, L. Nie, S. Yao, Synthesis and characterization of novel hexafluorophosphate salts with tropine-type cations, J. Mol. Liq. 209 (2015) 648–656. [37] X. Peng, Y. Hu, H. Chu, C. Miao, Z. Li, Z. Wang, Prediction of density and viscosity of ternary systems [C2q]Br+[C4q]Br+H2O, [C2q]Br+[C6q]Br+H2O, and [C4q]Br+[C6q]Br+H2O at different temperatures using their binary subsystems data, Pet. Sci. 9 (2012) 242–250. [38] T. Yao, S. Yao, Magnetic ionic liquid aqueous two-phase system coupled with high performance liquid chromatography: a rapid approach for determination of chloramphenicol in water environment, J. Chromatogr. A 1481 (2017) 12– 22. [39] J.C. Merchuk, B.A. Andrews, J.A. Asenjo, Aqueous two-phase systems for protein separation: studies on phase inversion, J. Chromatogr. B 711 (1998) 285–293. [40] L. Morandeira, M.S. Álvarez, F.J. Deive, M.Á. Sanromán, A. Rodríguez, Contriving to selectively separate drugs with a hydrophilic ionic liquid, Sep. Purif. Technol. 174 (2017) 29–38. [41] Y. Zhao, X.P. Lai, H.Y. Yao, R. Zhao, Y.N. Wu, G. Li, Optimization of dissolution process for superfine grinding technology on total saponins of Panax ginseng fibrous root by response surface methodology, J. Chin. Med. Mater. 37 (2014) 494–498. [42] X.S. Fang, J.H. Wang, Y.Z. Wang, X.K. Li, H.Y. Zhou, L.X. Zhu, Optimization of ultrasonic-assisted extraction of wedelolactone and antioxidant polyphenols from Eclipta prostrate L using response surface methodology, Sep. Purif. Technol. 138 (2014) 55–64. [43] M.T. Zafarani-Moattar, S. Emamian, S. Hamzehzadeh, Effect of temperature on the phase equilibrium of the aqueous two-phase poly(propylene glycol) +tripotassium citrate system, J. Chem. Eng. Data 53 (2008) 456–461. [44] M.G. Freire, C.M.S.S. Neves, P.J. Carvalho, R.L. Gardas, A.M. Fernandes, I.M. Marrucho, L.M.N.B.F. Santos, J.A.P. Coutinho, Mutual solubilities of water and hydrophobic ionic liquids, J. Phys. Chem. B 111 (2007) 13082–13089. [45] S.P.M. Ventura, S.G. Sousa, L.S. Serafim, Á.S. Lima, M.G. Freire, J.A.P. Coutinho, Ionic-liquid-based aqueous biphasic systems with controlled pH: the ionic liquid anion effect, J. Chem. Eng. Data 57 (2012) 507–512. [46] M.T. Zafarani-Moattar, A. Zaferanloo, Measurement and correlation of phase equilibria in aqueous two-phase systems containing polyvinylpyrrolidone and di-potassium tartrate or dipotassium oxalate at different temperatures, J. Chem. Thermodyn. 41 (2009) 864–871. [47] L. Wang, H. Zhu, C.H. Ma, Y.S. Yan, Q.W. Wang, Measurement and correlation of phase diagram data for aqueous two-phase systems containing 1-ethyl-3methylimidazolium dimethyl phosphate+K3PO4/K2HPO4/K2CO3+H2O ionic liquids at (298.15, 308.15, and 318.15) K, J. Chem. Eng. Data 57 (2012) 681– 687. [48] M. Matsumoto, R. Okuno, K. Kondo, Extraction of 2,3-butanediol with aqueous two phase systems formed by water-miscibleorganic solvents and inorganic salts, Solvent Extra. Res. Dev. 21 (2014) 181–190. [49] Y. Yang, Y. Liu, J. Han, S. Hu, Application of water-miscible alcohol-based aqueous two-phase systems for extraction of dyes, Sep. Sci. Technol. 46 (2011) 1283–1288. [50] J. Han, C.L. Yu, Y. Wang, X.Q. Xie, Y.S. Yan, G.W. Yin, W.X. Guan, Liquid–liquid equilibria of ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate and sodium citrate/tartrate/acetate aqueous two-phase systems at 298.15 K: experiment and correlation, Fluid Phase Equilibria. 295 (2010) 98–103. [51] X. Lin, Y. Wang, Q. Zeng, X. Ding, J. Chen, Extraction and separation of proteins by ionic liquid aqueous two-phase system, Analyst 138 (2013) 6445–6453. [52] K. Dong, S.J. Zhang, Hydrogen bonds: a structural insight into ionic liquids, Chem. Eur. J. 18 (2012) 2748–2761. [53] C.H. Ma, L. Wang, Y.S. Yan, G.B. Che, Y.S. Yin, R.Z. Wang, D.Y. Li, Extraction of tetracycline via ionic liquid two-phase system, Chem. Res. Chin. Univ. 25 (2009) 832–835. [54] J.S. Kim G.S. Moon H.O. Kim Y.S. Lee Antioxidant properties of ginseng (P. ginseng C.A. Meyer) extracts by organic solvent fractionation J. Food Sci. Nutr. 12 (2007) 267 272. [55] J.J. Wee, J.D. Park, M.W. Kim, H.J. Lee, Identification of phenolic antioxidant components isolated from Panax ginseng, J. Korean Agric. Chem. Soc. 32 (1989) 50–56. [56] Z.Q. Liu, Chemical insights into ginseng as a resource for natural antioxidants, Chem. Rev. 112 (2012) 3329–3355. [57] E.J. Cho, L.P. Xiang, M.H. Jang, S.H. Baek, H.Y. Kim, K.S. Kang, S.W. Kwon, J.K. Park, The effect of steaming on the free amino acid contents and antioxidant activity of Panax ginseng, Food Chem. 107 (2008) 876–882. [58] D. Luo, B. Fang, Structural identification of ginseng polysaccharides and testing of their antioxidant activities, Carbohyd. Polym. 72 (2008) 376–381.

Please cite this article in press as: A. He et al., Extraction of bioactive ginseng saponins using aqueous two-phase systems of ionic liquids and salts, Separ. Purif. Technol. (2017), http://dx.doi.org/10.1016/j.seppur.2017.05.041