A formaldehyde carbonyl groups-modified self-crosslinked polystyrene resin: Synthesis, adsorption and separation properties

Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

A formaldehyde carbonyl groups-modified self-crosslinked polystyrene resin: Synthesis, adsorption and separation properties Xiaoting Li a,b , Yongfeng Liu a,c , Duolong Di a,c,∗ , Gaohong Wang a,b , Yi Liu a,c a Key Laboratory of Chemistry of Northwestern Plant Resources and Key Laboratory for Natural Medicine of Gansu Province, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, 18 Tianshui Middle Road, Lanzhou 730000, People’s Republic of China b Graduate University of the Chinese Academy of Sciences, 19A Yuquan Road, Beijing 100049, People’s Republic of China c Centre of Resource Chemical and New Material, 36 Jinshui Road, Qingdao 266100, People’s Republic of China

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

• Novel formaldehyde carbonyl modi-

• • • •

fied self-crosslinked polystyrene resins were synthesized and employed to adsorb glabridin. The molecular structure and diameter of glabridin were optimized by Gaussian 09D01. The resins have much improved adsorption properties for glabridin than the BMKB-1. Detailed comparative studies on adsorption equilibrium and kinetics of glabridin with BMKB-1. The adsorption mechanism mainly ascribed to a synergistic effect of molecular sieving effect and multiple interactions.

a r t i c l e

i n f o

Article history: Received 26 November 2015 Received in revised form 9 March 2016 Accepted 25 March 2016 Available online 26 March 2016 Keywords: Self-crosslinked resin Microporous Glabridin Adsorption kinetics Adsorption isotherm

a b s t r a c t Based on our previous work and calculational chemistry, the pore-size distribution is the key factor which influences glabridin uptakes on resins. In addition, micropores and mesopores in the range of 1.8–5.4 nm are proper for glabridin adsorption. Based on this, a series of novel microporous and mesoporous self-crosslinked polystyrene resins (named as XT2) were synthesized by Friedel–Crafts reaction after different reaction time and utilized to adsorb glabridin in aqueous solution. The adsorption behaviors of the synthetic resins were studied systematically in terms of adsorption capacity, equilibrium time, isotherm adsorption and regeneration properties and compared with a commercial adsorbent (BMKB-1) which was screened out as the optimal macroporous adsorption resin (MAR) for glabridin based on our previous work. The characteristic methods of BET surface area, pore size distribution, Fourier transform infrared spectroscopy and scanning electron microscopy were investigated to analyze the resins and adsorption process. The glabridin uptakes on XT2-10 were remarkably larger than those of macroporous adsorption resin BMKB-1. The maximum adsorption capacity of XT2-10 is up to 43.69 mg/g for glabridin. The adsorption isotherms could be well described by the Freundlich model, and the adsorption kinetics

∗ Corresponding author at: Key Laboratory of Chemistry of Northwestern Plant Resources and Key Laboratory for Natural Medicine of Gansu Province, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, 18 Tianshui Middle Road, Lanzhou 730000, People’s Republic of China. E-mail address: [email protected] (D. Di). http://dx.doi.org/10.1016/j.colsurfa.2016.03.061 0927-7757/© 2016 Elsevier B.V. All rights reserved.

2

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

were fitted by both pseudo-second-order kinetic equation and intra-particle diffusion model. The adsorption mechanism was a synergistic effect of specific surface area, molecular sieving effect, and multiple adsorption interactions including hydrogen bonding and ␲–␲ stacking. The glabridin uptakes decreased to approximately 93.23% after five cycles of adsorption–desorption, exhibiting an excellent reusability and remarkable regeneration. Based on these results, this research not only opens up the possibility of synthesizing microporous and mesoporous self-crosslinked polystyrene resins, but also provides guidance for understanding the adsorption mechanism of purification of flavones from herbal plants. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Glabridin is an isoflavonoid originally isolated from the roots of Glycyrrhiza glabra L. Glabridin is widely considered to be a phytoestrogen and has been associated with numerous biological properties ranging from anti-atherogenic effects, antiinflammatory, neuroprotective, antioxidant, to the regulation of energy metabolism, but also including anti-tumorigenic, antinephritic, antibacterial and skin-whitening activities [1]. Glabridin and, to a lesser extent, have a growing impact not only on cosmetics but also on the food and dietary supplements (DSs) market. Although high-purity glabridin can be obtained, methods widely used at present are not appropriate for industrial process because of high-cost, time-consuming processing and environmental pollution. Adsorption resins have received considerable attentions due to their excellent performance in isolation and purification of fine chemicals, pharmaceuticals and food additives [2]. The adsorption resins have been widely applied in enriching the glabridin from crude G. glabra L. extracts in China, Japan, and other Asia countries. However, the main problem in the attempt to separate specific component is the poor adsorption selectivity of adsorption resins. In order to obtain better adsorption selectivity for a specific compound, chemical modification of ordinary adsorbents is often adopted by introducing some special functional groups onto the adsorbent matrix [3]. Polymeric adsorbents are widely applied in modern adsorption separation technology due to their favorable physicochemical stability, large adsorption capacity, excellent selectivity and structural diversity [4–6]. Their geometric pore structure can be easily regulated by adjusting the crosslinking reagents and porogens and their surface chemistry can be changed with ease by using co-monomers with desired functional groups in copolymerization or by a particular chemical reaction of the synthesized polymer [7,8]. Recently, a new kind of hypercrosslinked polymeric adsorbent (HCP) was found to be very effective for removing aromatic compounds from aqueous solutions [9,10]. This kind of hypercrosslinked polymeric adsorbent was firstly prepared by Davankov and Tsyurupa and is usually synthesized from a linear poly (styrene-co-divinylbenzene) (PS) or a low cross-linked PS by adding bi-functional cross-linking reagents such as monochloromethylether, p-dibenzenylchloride and pdichlomethylbenzene, and Friedel–Crafts catalysts including anhydrous zinc chloride, iron(III) chloride and stannic(IV) chloride [11–13]. They can also be prepared from a macroporous low crosslinked chloromethylated PS via its self Friedel–Crafts reaction. The structure characterization of this kind of polymeric adsorbent possessed a high crosslinking degree (40–500%), small average pore diameter (about 3 nm), a high BET surface area (700–1300 m2 /g) and a large pore volume (0.4–0.8 cm3 /g) with a predominant micro/mesopores [12–14]. The adsorption experiments indicate that HCPs have very high capacity toward various substances from gas phase media and aqueous solutions, which are far superior to macroporous polystyrene adsorbents [15,16].

In the present study, a self-crosslinked polymeric adsorbent modified with formaldehyde carbonyl groups, XT2 was synthesized from chloromethylated polystyrene and used as the adsorbent. A kind of natural product, glabridin, as previously mentioned, was chosen as the representative adsorbate. The adsorption characteristics, kinetics and thermodynamics of XT2 for glabridin were elucidated in aqueous solution. The main goal of this report is to compare the adsorption performance of XT2 and BMKB-1 and examine the effects of the pore structure and functional groups of the polymeric adsorbent surface on the adsorption properties. 2. Experimental 2.1. Materials Glabridin extract with 90% purity used as the adsorbate in this study was purchased from Nanjing Zelang Medical Technology Co., Ltd. (Nanjing, China). Acetonitrile used for HPLC analysis was of chromatographic grade and purchased from Tianjin Chemical Reagent Co., Inc. (Tianjin, China). Acetic acid was purchased from Shandong Yuwang Industrial Co., Ltd. (Shandong, China) and distilled water used was obtained in our laboratory. All solutions prepared for HPLC were subjected to filtration through 0.45 ␮m nylon membranes before use. 2.2. Synthesis of formaldehyde carbonyl groups-modifed self-crosslinked resins As described in Fig. 1 , formaldehyde carbonyl groups-modifed self-crosslinked resins were fabricated by one step. 20 g of chloromethylated polystyrene beads was firstly swollen in 75 mL of nitrobenzene at 298 K overnight. Under mild mechanical stirring, 1.875 g of anhydrous zinc chloride was added as the catalysts into the reaction mixture as quickly as possible at 298 K. After the added zinc chloride was completely dissolved, the mixture was evenly heated to 388 K within 1.5 h using linear temperature program with gradients of 1 ◦ C/min. After refluxing reaction mixture at 388 K for 4, 6, 8, 10, and 12 h, respectively, the self-crosslinked polystyrene resin beads labeled as XT2-4, XT2-6, XT2-8, XT2-10, and XT2-12 were obtained. To remove residual nitrobenzene and zinc chloride after the reaction, the polymeric beads were subjected to rinsing with 1% hydrochloric acid solution and ethanol respectively till the effluent was transparent, and finally washed with deionized water until neutral pH. The polymeric beads were extracted with ethanol in Soxhlet apparatus for 10 h and then dried under vacuum at 323 K for 6 h. 2.3. Adsorbent characterization and analytical methods The chorine content of the resins was measured according to a Volhard method [17]. The specific surface area and pore volume of the resins were determined by nitrogen adsorption–desorption isotherms at 77 K using a Micromeritics ASAP 2020 surface area and porosity analyzer (Micromeritics Instrument Corp., Norcross,

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

3

Fig. 1. Schematic illustration for the synthesis procedure of XT2 resin series.

Fig. 2. SEM images of (a) the chloromethylated polystyrene and (b) XT2-10.

the glabridin aqueous solutions was sampled at preset intervals to analyze the residual concentration. The adsorption capacities of the resins to glabridin were evaluated using the following equation: qe = (Co −Ce ) ×

V0 W

(1)

where qe is the adsorption capacity (mg/g dry resin) toward glabridin at adsorption equilibrium; C0 and Ce are the initial and equilibrium concentrations of the glabridin solutions (mg/mL), respectively; V0 is the adsorption solution volume (mL); and W is the dry weight of the resins (g).

2.5. Adsorption isotherms

Fig. 3. Comparison of equilibrium glabridin uptakes on XT2-4, XT2-6, XT2-8, XT2-10 and XT2–12 at 288 K.

USA). The adsorbents were outgassed at 333 K for 24 h on the degas port of the analyzer prior to the BET surface area measurement. The total specific surface area and pore volume were calculated according to BET, Langmuir and Barrett–Joyner–Halenda (BJH) models while the pore size distribution was calculated by applying BJH method to the nitrogen desorption data. The infrared spectra of the resins were obtained using Fourier-transform infrared spectrometry (FTIR) with a spectrophotometer (Thermo Nicolet, NEXUS, USA) via the potassium bromide technique with the wavenumbers ranged in 500–4000 cm−1 . The morphology of the resins were carried out by JSM–6701 scanning electron microscopy (SEM). 2.4. Adsorption kinetics The adsorption kinetics of glabridin on BMKB-1 and XT2-10 were conducted as follows: 100 mL (ethanol:water = 20:80, v/v) glabridin solution with a known concentration (0.45 mg/mL) was allowed to come in contact with the resins (equal to 1.0 g dry resin) in a 250 mL conical flask. The flask was continuously shaken at 288, 303, and 318 K for 12 h, separately. Subsequently, 1.0 mL of

The equilibrium adsorptions of glabridin were conducted at three different temperatures (288, 298, and 308 K) with 100 mL glabridin solutions for 8 h at 120 rpm. The residual concentrations were determined via HPLC.

2.6. HPLC analysis of glabridin Glabridin concentrations in aqueous solution were analyzed with an Agilent 1200 series. Chromatographic separation of glabridin and its interfering compounds was performed on a reversed-phase C18 column (Sinochrom ODS-BP, 250 mm × 4.6 mm, i.d., 5 ␮m) (Elite Analytical Instruments, Dalian, China). The temperature of the column was maintained at 303 K. The mobile phase was composed of acetic acid–water (0.5:99.5, v/v, A) and acetonitrile (B). The mobile phase proportion was 40:60 (A:B, v/v). The flow rate was set at 1.0 mL/min. The detection wavelength was set at 282 nm (190–400 nm full scan). The injection volume was 20 ␮L for each run. All the mobile phase, standard and sample solutions prepared for HPLC were subjected to filtration through a 0.45 ␮m membrane filter before use. The results indicated that the working calibration curve based on glabridin standard solutions showed good linearity over the range of 0–0.6 mg/mL. The regression equation for glabridin is y = 67071.1905x + 26.0143 (R2 = 0.9871, n = 8), where y is the peak area of glabridin and x is glabridin concentration (mg/mL).

4

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

3. Results and discussion 3.1. Characteristics of the resins As listed in Table 1, after the Friedel–Crafts reaction, the chlorine content of the self-crosslinked polystyrene resins sharply decreased from 2.03 mmol/g to 0.86 mmol/g, indicating that the chlorine of XT2 was consumed in the reaction process and a longer reaction time lead to lower chlorine content. These results also demonstrated that self-crosslinking was successful comparative to chloromethylated polystyrene resin. It was found that the optimum ratio of the pore diameter of the adsorbent to the molecular size of the adsorbate was 2–6 for polymeric adsorbents [18]. The molecular diameter of glabridin is optimized to be 0.9 nm by Gaussian 09D01 software package and the pore diameter of XT2 series resins is determined to be about 3.0 nm, which is suitable for the adsorption. The BET surface area of XT2-4, XT2-6, XT2-8, XT2-10 and XT212 rapidly increased from 531.4 m2 /g to 567.5, 600.4, 730.8 and 870.5 m2 /g, respectively, meaning that a large quantity of methylene crosslinking bridges with conformationally rigid links were formed between the polystyrene chains in the Friedel–Crafts reaction and a longer reaction time brought about more methylene crosslinking bridges. In particular, the t-plot surface area of the self-crosslinked polystyrene resins accounted for more than half of the BET surface area, which demonstrated that a large number of micropores were produced in the reaction process. Fig. S1 describes the N2 adsorption/desorption isotherms of the resins at 77 K. As shown in Fig. S1a, the N2 adsorption capacity of the XT2-10 was much larger than chloromethylated polystyrene at the same relative pressure, indicating XT2-10 has larger specific surface area. Fig. S1 illustrated that both the N2 adsorption isotherms seem close to type-II classification [19,20]. As shown in Fig. S1a, at a relative pressure below 0.05 and above 0.95, the N2 uptakes of the XT2-10 resin increased sharply with increasing of the relative pressure, demonstrating that micropores are predominant and macopores are also existent for the XT2-10 resin while the chloromethylated polystyrene resin possesses mainly macopores. Furthermore, the visible hysteresis loops of desorption isotherms of the XT2-10 resin indicate that adsorbent contain mesopores and with a low pore connectivity [21,22]. The N2 adsorption/desorption isotherms of the XT2 series resins in Fig. S1b present the same trend as XT2-10. These analyses agree with results of the pore width distribution of the resins in Fig. S2. The Friedel–Crafts reaction results in a great transfer for the pore width distribution of the resins, meso/macropores are the main pores for the chloromethylated polystyrene, while mesopores in the range of 2–5 nm play a dominant role for the self-crosslinked polystyrene resins. The FTIR spectra in Fig. S3 displays that after the Friedel–Crafts reaction, all of the vibrations of the chloromethylated polystyrene remained while the two characteristic vibrations related to C Cl stretching of CH2 Cl group at 1265.1 and 669.2 cm−1 are weakened greatly, in accordance with the sharp decrease in chlorine content. The spectrum of chloromethylated polystyrene characterizes its polystyrene-type structure based on the representative vibrations at 3019, 2930, 1603, 1510, and 1448 cm−1 . This result suggested that a self-crosslinked resin was successfully formed. Interestingly, one strong band with frequency at 1704.8 cm−1 for the obtained self-crosslinked resin appears and this band can be assigned to the carbonyl groups (C O) stretching of benzaldehyde due to the oxidation of benzyl chloride by nitrobenzene and oxygen in the reaction system [23]. The SEM images of the chloromethylated polystyrene and XT210 are presented in Fig. 2. It is reasonable to observe that the surface of the chloromethylated polystyrene (see Fig. 2a) is much smoother and more compact with many particle and lower porosity. By

Fig. 4. Adsorption isotherms of glabridin on BMKB-1 and XT2–10 at 288 K.

contrast, the XT2-10 (see Fig. 2b) displays a more porous structure after Friedel–Crafts reaction. This is consistent with the fact that the BET surface area of XT2-10 (730.8 m2 /g) is much higher than that of the chloromethylated polystyrene (15.0 m2 /g).

3.2. Adsorption selectivity Section 3.1 indicated that the five XT2 series resins had the different pore structure and BET surface area, indicative of their different adsorption selectivity. The adsorption capability of an adsorbent is one of the most important aspects to be concerned if the adsorbent can be really applied in industry. The adsorption capacities of glabridin on XT2-4, XT2-6, XT2-8, XT2-10 and XT2-12 at 288 K are compared in Fig. 3. Evidently, the glabridin uptakes on the resins firstly increased and then decreased with increment of the Friedel–Crafts reaction time and XT2-10 had the largest glabridin uptakes among the five resins. Hence, XT2-10 was employed as a specific polymeric adsorbent in this study. The commercial resin (BMKB-1) was screened out as the optimal macroporous adsorption resin for glabridin based on our previous work. Thus, the isotherm of glabridin on XT2-10 is then compared with that on BMKB-1 at 288 K (see Fig. 4). As shown in the figure, equilibrium adsorption capacity of glabridin on XT2-10 is obviously larger than that on BMKB-1 at the same equilibrium concentration. The larger adsorption capacity of XT2-10 may result from the synergistic effect (see Fig. 5) of the size matching between the pore diameter of XT2-10 and the molecular size of glabridin, a larger ␲–␲ interaction, and hydrogen bonding between the formaldehyde carbonyl group of XT2-10 and glabridin. To interpret the possible hydrogen bonding between the carbonyl groups of XT2-10 and glabridin, the FTIR measurement was employed to examine the representative vibrational shifts of the carbonyl groups on XT2-10. The FTIR results of XT2-10 before and after glabridin adsorption were revealed in Table S1. The polystyrene-type structure based on the representative vibrations at 3019, 2930, 1603, 1510, and 1448 cm−1 are not changed at the tested two states. However, the characteristic vibration which is concerned with C O stretching with frequency at 1704.8 cm−1 before glabridin adsorption appears at 1715 cm−1 (red-shifted by 10.2 cm−1 ) after glabridin adsorption. Hence, we deduce that hydrogen bonding is formed between the carbonyl groups on XT2-10 and the phenolic hydroxyl groups of glabridin, and hydrogen bonding is an important mechanism for the glabridin adsorption on XT2-10.

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

5

Table 1 Physical properties of the self-crosslinked polystyrene resins.

BET surface area (m2 /g) Langmuir surface area (m2 /g) t-Plot micropore surface area (m2 /g) Pore volume (cm3 /g) t-Plot micropore volume (cm3 /g) Average pore width (nm) Particle size (nm) Chlorine content (mmol/g)

XT2-4

XT2-6

XT2-8

XT2-10

XT2-12

531.4 715.2 273.2 0.46 0.12 3.47 0.3–1.2 2.03

567.5 763.1 295.3 0.49 0.13 3.46 0.3–1.2 1.88

600.4 807.9 317.7 0.51 0.14 3.40 0.3–1.2 1.35

730.8 982.2 409.0 0.59 0.19 3.24 0.3–1.2 1.18

870.5 1168.6 477.7 0.69 0.22 3.15 0.3–1.2 0.86

Fig. 5. Schematic drawing of multiple interactions between glabridin and XT2 resin series with formaldehyde carbonyl as functional group.

Fig. 6. Adsorption kinetic curves of glabridin on (a) BMKB-1 and XT2–10 at 288 K; (b) XT2–10 at 288, 303 and 318 K.

3.3. Adsorption kinetics Adsorption kinetics studies described the dynamics of the adsorption and evidently this rate controlled the residence time of adsorbate uptake at the solid–liquid interface including the diffusion process. Fig. 6a displays the adsorption kinetics curves for glabridin on macroporous adsorption resin BMKB-1 and selfcrosslinked polystyrene resin XT2–10 at 288 K. And Fig. 6b displays the adsorption kinetics curves for glabridin on XT2–10 at 288, 303, and 318 K, respectively. As shown in the figure, the uptakes of glabridin on XT2-10 is much higher than BMKB-1. Also it can be seen that the required time from the beginning to the equilibrium for BMKB-1 is about 400 min at the initial glabridin concentration of 0.45 mg/mL, shorter than 600 min for XT2–10 at the same concentration. This is reasonable that a good deal of micropores (<2 nm) in

XT2-10 increase the diffusion resistance of glabridin, which causes the diffusion rate to decrease for XT2-10. The well-known Lagergren pseudo-first order [24], Mckay pseudo-second order [25] and intra-particle diffusion models [6] were employed to analyze all the kinetic data to further explain the adsorption mechanisms. Pseudo-first-order rate equation: ln(qe −qt ) = −k1 t+lnqe

(2)

Pseudo-second-order rate equation: 1 1 t = t+ qt qe k2 q2e

(3)

6

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

Fig. 7. Plotting of lnk2 versus 1/T for the adsorption kinetic curves of glabridin adsorption on XT2-10.

Fig. 8. The fitted kinetic curves for the adsorption of XT2-10 toward glabridin according to the intra-particle diffusion model.

The intra-particle diffusion equation: qt = kid t1/2 +C

(4)

where k1 (1/min), k2 (g/mg min) and kid (mg/g min1/2 ) are rate constants of the pseudo-first-order, pseudo-second-order and intra-particle diffusion model, respectively. Parameters qt and qe (mg/g) are the adsorption capacity at contact time t and equilibrium, respectively. The constant C (mg/g) represents the boundary layer thickness. Plotting ln(qe − qt ) against t, t/qt via t, and qt versus t1/2 would get straight lines, respectively. Corresponding values of kinetic parameters are summarized in Table 2 and Table 3. By checking the values of linear regression correlation coefficients, it was found that adsorption on macroporous adsorption resin BMKB-1 can be fitted by the pseudo-second-order equation while the pseudo-secondorder kinetics and intra-particle diffusion equation provided a good correlation for the adsorption of glabridin onto XT2-10. Furthermore, the rate constant k2 for BMKB-1 is greater than the one of XT2–10 at the same condition, consistent with the shorter required equilibrium time for BMKB-1. This can be explained from the pore structure of the two resins. Both the average pore diameter and the pore distribution of BMKB-1 are more favorable for the fast diffusion of glabridin from the solution phase to the adsorption site of the resin. Additionally, the k2 was also applied to estimate the activation energy (Ea, kJ/mol) for the adsorption as: lnk2 = −

Ea +lnk0 RT

(5)

where Ea is the Arrhenius activation energy for the adsorption and it represents the minimum energy that the reactants must have if the reaction can be proceeded successfully, k0 is the Arrhenius factor. By plotting lnk2 versus 1/T, a straight line is obtained (see Fig. 7), and Ea can be calculated from the slope of straight line to be 23.08 kJ/mol. Generally, a low Arrhenius activation energy (5–40 kJ/mol) is characteristic of a physical adsorption while a higher Arrhenius activation energy (40–800 kJ/mol) suggests a chemical reaction has occured. The present Arrhenius activation energy for the adsorption indicates that the process may be related to a physical process. The image in Fig. 8 shows the plot of qt versus t1/2 for adsorption on XT2-10. As can be seen from Fig. 8, these plots at three different temperatures show similar characters having two linear segments followed by a plateau. The first stage gives a linear relationship and the straight lines do not pass through the origin, implying that

Fig. 9. Adsorption isotherms of glabridin on XT2–10 at the temperature of 288, 298 and 308 K, respectively.

the intra-particle diffusion is not the sole rate-limiting step. The kid,1 (see Table 3) evaluated are 2.26, 2.14 and 1.97 (mg/g min1/2 ) at 288, 303 and 318 K, respectively, proving that the intra-particle diffusion rate at a lower temperature is higher, and which is in accordant with the deduction of the glabridin molecule diffuses faster at a lower temperature [26]. At the second stage, plotting of qt versus t1/2 also yields a linear relationship but does not pass through the origin, revealing that multi-diffusion mechanisms are involved. At the third stage, adsorption reaches equilibrium because of the extremely low concentration of glabridin left in solution. 3.4. Equilibrium adsorption isotherms Based on the adsorption kinetics study, we next evaluated the adsorption isotherms of XT2-10. Equilibrium adsorption isotherms of an adsorbent are a general relationship between the equilibrium adsorption capacity of the adsorbate on the adsorbent (qe ) and the equilibrium concentration of the adsorbate (Ce ) at a desired temperature. As shown in Fig. 9, the equilibrium adsorption isotherms of XT2-10 were investigated at three different temperatures of 288, 298, and 308 K. It is obvious that adsorption capacity increases with increasing concentrations of glabridin and low temperature

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

7

Table 2 Kinetic parameters for adsorption of glabridin onto MARs. T (K)

MARs

288 288 303 318

BMKB-1 XT2-10 XT2-10 XT2-10

Pseudo-first order model k1 (1/min)

qe (mg/g)

R2

Pseudo-second order model k2 (g/mg min)

qe (mg/g)

R2

0.0138 0.0075 0.0069 0.0071

64.25 47.36 44.71 47.18

0.9473 0.7890 0.9594 0.9195

0.0013 0.0002 0.0003 0.0005

42.79 48.88 49.26 48.78

0.9976 0.9953 0.9954 0.9957

Table 3 Intraparticle diffusion model parameters for the adsorption of glabridin on XT2-10 resin. T (K)

kid,1 (mg/g min1/2 )

C1

kid,2 (mg/g min1/2 )

C2

kid,3 (mg/g min1/2 )

C3

288 303 318

2.26 2.14 1.97

−0.63 0.51 0.09

1.09 1.36 1.34

17.24 10.31 9.01

0.03 0.17 0.10

42.98 37.38 37.83

is favorable for the adsorption of glabridin onto XT2-10, indicating that the adsorption is an exothermic process. The experimental results are fitted to the well-known Langmuir and Freundlich model for the comparison. It is well known that the Langmuir model is usually used with an ideal assumption of an entirely homogeneous adsorption surface, whereas the Freundlich model is appropriate for a heterogeneous surface [27]. The Langmuir and Freundlich model can be written as [28]: Langmuir model: Ce Ce KL = + qe qm qm

(6)

Freundlich model: lnqe =

1 lnCe +lnKF n

(7)

where Ce is the equilibrium concentration (mg/mL), qe is the equilibrium adsorption capacity (mg/g), qm is the saturated adsorption capacity (mg/g), KL , KF and n are the characteristic constants. Both the Langmuir and Freundlich equations are adopted to describe the adsorption isotherms and the corresponding equation parameters are listed in Table 4. It is seen that Freundlich equation is obviously more suitable for characterizing the isotherm data due to a higher correlation coefficients R2 (>0.97) given by Table 4, suggesting that the present adsorption system is a surface energy heterogeneity. qm of glabridin on XT2-10 is much larger than BMKB-1. In particular, KL of XT2-10 is greater than BMKB1, implying that the adsorption affinity of glabridin on XT2-10 is stronger. 3.5. Adsorption thermodynamics Adsorption enthalpy H (kJ/mol), adsorption free energy G (kJ/mol) and adsorption entropy S (J/mol K) are basic aspects for the adsorption process [29,30]. G is a fundamental criterion of adsorption spontaneity and an adsorption occurs spontaneously at a given temperature if G is negative. H is the interaction between the adsorbent and the adsorbate and it presents endothermic or exothermic character of an adsorption process. S is defined as the degree of chaos for a given system. A positive S reflects an increasing randomness at the solid/liquid interface during the adsorption while a negative S implies that the whole system is more ordered after the adsorption. In the present study, the three thermodynamic parameters G, H, and S could be estimated by the following equations: K = M/KL

(8)

G = −RTlnK

(9)

lnK = −H/RT+S/R

(10)

where M is the molecular weight of glabridin (324.28 g/mol), R is universal gas constant (8.314 J/mol K) and T is the absolute temperature (K). Then a plot of ln K as a function of 1/T yielded a straight line. The values of H and S were calculated from the slope and intercept of Eq. (10), respectively. As can be seen from Table 5, the overall G (−18.23 kJ/mol at 288 K, −17.49 kJ/mol at 298 K, −16.86 kJ/mol at 308 K) during the adsorption process is negative for the experimental temperature range, which suggested that the adsorption processes of glabridin on XT2-10 were spontaneous. The negative values of H (−60.88 kJ/mol) indicated that the adsorption processes were exothermic. The results were in accordance with the tendency that adsorption capacity decreased with the increase of temperature [31]. The present negative S reflects the ordered arrangement of the system after the adsorption of glabridin on XT2-10. 3.6. Reusability The regeneration and reuse of XT2-10 resin for glabridin adsorption are quite crucial for economic costs. It was found that the spent XT2-10 resin could be well desorbed by 100 mL 95% ethanol solution. Fig. 10 shows the glabridin removal percentage in five adsorption–desorption cycles. It can be seen that only 6.77% loss in the adsorption capacities of glabridin was observed after five cycles of adsorption–desorption. The results indicated that XT2-10 resin could be regenerated and exhibited remarkable reusability on separation applications. In addition, the adsorbents which have been used once should be stored by soaking in ethanol in general for later use. However, the completely exhausted adsorbent has to be safely disposed by common methods of disposal of wastes of similar category. And the disposal of waste adsorbent should conform to the laboratory waste disposal system and should possess no harm to the environment. 4. Conclusions A series of microporous and mesoporous self-crosslinked polystyrene resins named as XT2-4, XT2-6, XT2-8, XT2-10 and XT212 were synthesized by controlling the Friedel–Craft reaction time and they were shown to possess different BET surface area, pore textural property and surface functionality, indicative of their different adsorption performances toward glabridin. The glabridin uptakes on the resins firstly increased and then decreased with increment of the Friedel–Crafts reaction time and XT2-10 had the largest glabridin uptakes among the five resins. The linear fitting methods indicated that the Freundlich isotherm model was more suitable for fitting the equilibrium data than the Langmuir one. The kinetic data could be fit better by the pseudo second-order rate equation and intraparticle diffusion model, and the Arrhenius

8

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9

Table 4 Langmuir and Freundlich isotherm parameters of glabridin onto MARs. T (K) 288 288 298 308

MARs BMKB-1 XT2-10 XT2-10 XT2-10

Langmuir isotherm model qm (mg/g)

KL (L/g)

R2

Freundlich isotherm model KF [(mg/g)(L/mg)1/n ]

n

R2

45.39 54.11 71.68 65.32

0.09 0.16 0.36 0.35

0.9524 0.8494 0.9797 0.9771

54.84 76.09 68.39 60.40

2.5 1.59 1.54 1.61

0.9927 0.9738 0.9907 0.9939

Fig. 10. Effect of the regeneration cycles on the adsorption capacity of glabridin on XT2–10 at 288 K.

Table 5 Thermodynamics parameters for the adsorption of glabridin onto XT2-10.

References

T (K)

−G (kJ/mol)

−H (kJ/mol)

−S (J/mol K)

R2

288 298 308

18.23 17.49 16.86

60.88

147.88

0.9995

activation energy was calculated to be 23.08 kJ/mol. The XT2-10 resin displays higher adsoption capacity than that of macroporous adsorption resin BMKB-1 because of the size matching between the pore diameter of XT2-10 and the molecular size of glabridin, a larger ␲–␲ interaction and hydrogen bonding between the formaldehyde carbonyl group of XT2-10 and glabridin. The XT2-10 resin could be well desorbed by 100 mL 95% ethanol solution and the glabridin uptakes decreased to approximately 93.23% after five cycles of adsorption–desorption process. Acknowledgments This research project was financially supported by the National Natural Sciences Foundation of China (NSFC No. 20974116 and 21374128), the “Hundred Talents Program” of the Chinese Academy of Sciences (CAS), and the National High Technology Research and Development Program of China (863 No. 2014AA022203). The authors declare no competing financial interest(s). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfa.2016.03. 061.

[1] C. Simmler, G.F. Pauli, S.N. Chen, Phytochemistry and biological properties of glabridin, Fitoterapia 90 (2013) 160–184. [2] J. Bretag, D.R. Kammerer, U. Jensen, R. Carle, Adsorption of rutin onto a food-grade styrene–divinylbenzene copolymer in a model system, Food Chem. 114 (2009) 151–160. [3] X.M. Wang, R.J. Deng, X.Y. Jin, J.H. Huang, Gallic acid modified hyper-cross-linked resin and its adsorption equilibria and kinetics toward salicylic acid from aqueous solution, Chem. Eng. J. 191 (2012) 195–201. [4] B. Ge, Z.Z. Zhang, X.T. Zhu, X.H. Men, X.Y. Zhou, A superhydrophobic/superoleophilic sponge for the selective absorption oil pollutants from water, Colloids Surf. A: Physicochem. Eng. Asp. 457 (2014) 397–401. [5] W.M. Zhang, Q. Du, B.C. Pan, L. Lv, C.H. Hong, Z.M. Jiang, D.Y. Kong, Adsorption equilibrium and heat of phenol onto aminated polymeric resins from aqueous solution, Colloids Surf. A: Physicochem. Eng. Asp. 346 (2009) 34–38. [6] S. Yang, L.Y. Li, Z.G. Pei, C.M. Li, J.T. Lv, J.L. Xie, B. Wen, S.Z. Zhang, Adsorption kinetics, isotherms and thermodynamics of Cr(III) on graphene oxide, Colloids Surf. A: Physicochem. Eng. Asp. 457 (2014) 100–106. [7] H.T. Li, Y.C. Jiao, M.C. Xu, Z.Q. Shi, B.L. He, Thermodynamics aspect of tannin sorption on polymeric adsorbents, Polymer 45 (2004) 181–188. [8] J.H. Huang, K.L. Huang, S.Q. Liu, Q. Luo, M.C. Xu, Adsorption properties of tea polyphenols onto three polymeric adsorbents with amide group, J. Colloid Interface Sci. 315 (2007) 407–414. [9] A.L. He, J.H. Huang, C. Yan, J.B. Liu, L.B. Deng, K.L. Huang, Adsorption behaviors of a novel carbonyl and hydroxyl groups modified hyper-cross-linked poly(styrene-co-divinylbenzene) resin forb-naphthol from aqueous solution, J. Hazard. Mater. 180 (2010) 634–639. [10] A.C. Pan, W. Du, W.M. Zhang, X. Zhang, Q.R. Zhang, B.J. Pan, L. Lv, Q.X. Zhang, J.L. Chen, Improved adsorption of 4-nitrophenol onto a novel hyper-crosslinked polymer, Environ. Sci. Technol. 41 (2007) 5057–5062. [11] V.A. Davankov, M.P. Tsyurupa, Structure and properties of hypercrosslinked polystyrene-the first representative of a new class of polymer networks, React. Polym. 13 (1990) 27–42. [12] M.P. Tsyurupa, V.A. Davankov, Hypercrosslinked polymers: basic principle of preparing the new class of polymeric materials, React. Funct. Polym. 53 (2002) 193–203. [13] M.P. Tsyurupa, V.A. Davankov, Porous structure of hypercrosslinked polystyrene: state-of-the-art mini-review, React. Funct. Polym. 66 (2006) 768–779. [14] J.H. Huang, X.Y. Jin, J.L. Mao, B. Yuan, R.J. Deng, S.G. Deng, Synthesis, characterization, and adsorption properties of diethylenetriamine-modified hypercrosslinked resins for efficient removal of salicylic acid from aqueous solutions, J. Hazard. Mater. 217–218 (2012) 406–415.

X. Li et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 500 (2016) 1–9 [15] G.Q. Xiao, L.C. Fu, A.M. Li, Enhanced adsorption of bisphenol A from water by acetylaniline modified hyper-cross-linked polymeric adsorbent: effect of the cross-linked bridge, Chem. Eng. J. 191 (2012) 171–176. [16] A.M. Li, Q.X. Zhang, G.C. Zhang, J.L. Chen, Z.H. Fei, F.Q. Liu, Adsorption of phenolic compounds from aqueous solutions by a water-compatible hypercrosslinked polymeric adsorbent, Chemosphere 47 (2002) 981–989. [17] C.P. Wu, C.H. Zhou, F.X. Li, Experiments of Polymeric Chemistry, Anhui Science and Technology Press, Hefei, 1987. [18] S. Kasaoka, Y. Sakata, E. Tanaka, R. Naitoh, Design of molecular-sieve carbon. Studies on the adsorption of various dyes in the liquid phase, Int. Chem. Eng. 29 (1989) 734–742. [19] J.H. Huang, X.M. Wang, K.L. Huang, Adsorption of p-nitroaniline by phenolic hydroxyl groups modified hyper-cross-linked polymeric adsorbent and XAD-4: a comparative study, Chem. Eng. J. 155 (2009) 722–727. [20] J.H. Huang, R.J. Deng, K.L. Huang, Equilibria and kinetics of phenol adsorption on a toluene-modified hyper-cross-linked poly(styrene-co-divinylbenzene) resin, Chem. Eng. J. 171 (2011) 951–957. [21] J.H. Huang, Molecular sieving effect of a novel hyper-cross-linked resin, Chem. Eng. J. 165 (2010) 265–272. [22] G. Zgrablich, S. Mendioroz, L. Daza, J. Pajares, V. Mayagoitia, F. Rojas, W.C. Conner, Effect of porous structure on the determination of pore size distribution by mercury porosimetry and nitrogen sorption, Langmuir 7 (1991) 779–785. [23] G.H. Meng, A.M. Li, W.B. Yang, F.Q. Liu, X. Yang, Q.X. Zhang, Mechanism of oxidative reaction in the post crosslinking of hypercrosslinked polymers, Eur. Polym. J. 43 (2007) 2732–2737.

9

[24] Y.J. Zhang, J.L. Ou, Z.K. Duan, Z.J. Xing, Y. Wang, Adsorption of Cr(VI) on bamboo bark-based activated carbon in the absence and presence of humic acid, Colloids Surf. A: Physicochem. Eng. Asp. 481 (2015) 108–116. [25] X.H. Zhao, F.P. Jiao, J.G. Yu, Y. Xi, X.Y. Jiang, X.Q. Chen, Removal of Cu(II) from aqueous solutions by tartaric acid modified multi-walled carbon nanotubes, Colloids Surf. A: Physicochem. Eng. Asp. 476 (2015) 35–41. [26] C.T. Miller, W.J. Weber Jr., Sorption of hydrophobic organic pollutants in saturated soil systems, J. Contam. Hydrol. 1 (1986) 243–261. [27] H. Zhang, M. Xu, H.J. Wang, D. Lei, D. Qu, Y.J. Zhai, Adsorption of copper by aminopropyl functionalized mesoporous delta manganese dioxide from aqueous solution, Colloids Surf. A: Physicochem. Eng. Asp. 435 (2013) 78–84. [28] X. Zhang, P.Y. Zhang, Z. Wu, L. Zhang, G.M. Zeng, C.J. Zhou, Adsorption of methylene blue onto humic acid-coated Fe3 O4 nanoparticles, Colloids Surf. A: Physicochem. Eng. Asp. 435 (2013) 85–90. [29] J.H. Huang, K.L. Huang, S.Q. Liu, A.T. Wang, C. Yan, Adsorption of Rhodamine B and methyl orange on a hypercrosslinked polymeric adsorbent in aqueous solution, Colloids Surf. A: Physicochem. Eng. Asp. 330 (2008) 55–61. [30] Z.X. Luo, M.L. Gao, S.F. Yang, Q. Yang, Adsorption of phenols on reduced-charge montmorillonites modified by bispyridinium dibromides: mechanism kinetics and thermodynamics studies, Colloids Surf. A: Physicochem. Eng. Asp. 482 (2015) 222–230. [31] J.J. Xiao, N. Di, G.Y. Liu, H. Zhong, The interaction of N-butoxypropyl-N -ethoxycarbonylthiourea with sulfide minerals: scanning electrochemical microscopy diffuse reflectance infrared Fourier transform spectroscopy, and thermodynamics, Colloids Surf. A: Physicochem. Eng. Asp. 457 (2014) 203–210.