graphene oxide

2 downloads 3 Views 1MB Size Report
and enrichment of evodiamine and rutaecarpine in Evodiae fructus ... selectivity, as well as simultaneous enrichment and separation abilities for evodiamine.

A molecular imprinted polymer on the surface of superparamagnetic Fe3O4–graphene oxide ([email protected]@GO) for simultaneous recognition and enrichment of evodiamine and rutaecarpine in Evodiae fructus Jie-Ping Fan,1,2 Dan-Dan Liao,1 Yan-Long Xie,1 Bing Zheng,1 Jia-Xin Yu,1 Ya-Hui Cao,1 Xue-Hong Zhang,3 Hai-Long Peng1 1

Department of Chemical Engineering, Nanchang University, Nanchang 330031, China Key Laboratory of Poyang Lake Ecology and Bio-Resource Utilization of Ministry of Education, Nanchang University,

2

Nanchang 330031, China 3

School of Foreign Language, Nanchang University, Nanchang 330031, China

Correspondence to: J.-P. Fan (E-mail: [email protected])

A composite material ([email protected]@GO) based on molecular imprinted polymer (MIP), superparamagnetic Fe3O4 particles and graphene oxide (GO) was prepared by the chemical coprecipitation method, and used to simultaneously separate and enrich two alkaloids (evodiamine and rutaecarpine) in the extract of Evodiae fructus. The as-prepared [email protected]@GO was characterized by Fourier transform infrared spectroscopy (FT-IR), transmission electron microscopy (TEM), and vibrating sample magnetometer (VSM). The adsorption kinetics, isotherms, competitive adsorption, and reusability of [email protected]@GO were also investigated. The results revealed that [email protected]@GO was sensitive to the magnetic field and could be easily separated using an external magnet; [email protected]@GO showed good recognition selectivity, as well as simultaneous enrichment and separation abilities for evodiamine C 2016 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2016, 133, 44465. and rutaecarpine in the extract of E. fructus with satisfactory recoveries. V ABSTRACT:

KEYWORDS: adsorption; magnetism and magnetic properties; molecular recognition; separation techniques

Received 9 June 2016; accepted 14 September 2016 DOI: 10.1002/app.44465 INTODUCTION

Evodiae fructus (Chinese name: Wuzhuyu) is a traditional Chinese medicine for the treatment of gastrointestinal disorders, headache, and postpartum hemorrhage,1,2 cultivated in East Asian nations. Evodiamine (EVO) and rutaecarpine (RUT) (shown in Figure 1) are two main active alkaloids in E. fructus, and they have exhibited many pharmacological effects.3–8 Therefore, it is very important to developed an adsorbent for selective and simultaneous separation or enrichment of the two alkaloids to completely utilize the bioresource of E. fructus. In this work, a composite material ([email protected]@GO) based on molecular imprinted polymer (MIP), superparamagnetic Fe3O4 particles and graphene oxide (GO) was prepared to separate and enrich EVO and RUT in the extract of E. fructus, because it combined all of the advantages from MIP, superparamagnetic Fe3O4 and GO.9–12 First, MIP is a synthetic material with the capability of selective molecular recognition of targeted compounds because it has specific binding sites with complementary sizes, shapes, and functional groups to the targeted compounds, and involve a mechanism of interaction based on molecular recognition.13 Compared

with conventional separation material, MIP is of high selectivity, low cost, excellent reusability, and ease of preparation.13,14 Second, GO has unique mechanical properties and extremely large specific surface area, which made GO be a good supported material for preparing surface MIPs to improve accessibility of the recognition sites, binding capacity, and binding kinetics.11,12 Moreover, compared with graphene, GO is easier to immobilize the nanoparticles onto its surfaces due to its surface functional groups. Third, compared with the traditional solid support substrate, the magnetic support can be easily isolated from the samples using a magnet without additional centrifugation or filtration, meanwhile, the superparamagnetic particles can be easily re-dispersed by the removal of magnetic field and simple ultrasonication.11,12 So combining three promising concepts (GO, MIP and magnetic separation) would provide the composite material with high selectivity, high adsorption capacity, and easy manipulation. EXPERIMENTAL

Reagents and Materials The standards (98%) of EVO and RUT were purchased from Shaanxi Sciphar Biotechnology Co., Ltd., Xi’an, China; Limonin

C 2016 Wiley Periodicals, Inc. V

WWW.MATERIALSVIEWS.COM

44465 (1 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Figure 1. Molecular structures of evodiamine (EVO), rutaecarpine (RUT), and limonin (LIM).

(LIM) was purchased from National Institutes for Food and Drug control, China; E. fructus was purchased from Zhangshu Tianqitang Traditional Chinese Medicine Yinpian, Co., LTD, Jiangxi, China; Azobisisobutyronitrile (AIBN) and methacrylic acid (MAA) were purchased from Tianjin Damao Chemical Reagents Co., Tianjin, China; Ethylene glycol dimethacrylate (EGDMA) and graphite powder (450 mesh) were purchased from Shanghai Aladdin Reagent Company, Shanghai, China; (NH4)2Fe(SO4)26H2O and NH4Fe(SO4)212H2O were purchased from Tianjin Hengxing Chemical Reagent Co., Ltd., Tianjin, China. Neutral alumina was purchased from Sinopharm chemical Reagent Co., Ltd., Shanghai, China. Reverse-phase silica gel (RP-18) was supplied by Sepax Technologies, Inc., Suzhou, China. Macroporous resin (D-101) was supplied by Anhui Wandong Resin Technology Co., Ltd, Bengbu, China. Silica gel was purchased from the Qingdao Haiyang Chemical Co., Ltd, Qingdao, China. All the other chemicals were analytical grade reagents and purchased from Tianjin Damao Chemical Reagents Co., Ltd, China. Preparation of the Superparamagnetic [email protected] Nanocomposites GO was first synthesized from graphite powder by a modified Hummers method,15,16 and then the superparamagnetic [email protected] nanocomposites were prepared by chemical coprecipitating ferric and ferrous salts in the presence of GO. In a typical procedure, 10 mL mixture solution containing 1.7 g (NH4)2Fe(SO4)26H2O, 2.51 g NH4Fe(SO4)212H2O and 1 mL PEG was dropwise added into a 400 mL GO suspension solution (containing 500 mg GO) with continuously vigorous stirring at the room temperature, and the mixture was stirred for 12 h. With the assistance of ultrasound (KQ-600DB, Ultrasonic Instrument Ltd. Kunshan, China) at 277.15 K, an appropriate amount of NH3H2O was dropwise added into the mixture to reach a pH of 11, and then the mixture was aged at the constant stirring for another 1 h. The resulting black precipitate ([email protected]) was separated by an external magnetic field and then rinsed repeatedly with water and ethanol. Preparation of the Superparamagnetic [email protected]@GO Nanocomposites MIP was loaded on the surface of [email protected] in the following procedures. Briefly, EVO (0.16 mmol, template) and MAA (0.60 mmol, functional monomer) were dissolved in 50 mL methanol and ultrasonicated for 1 h at 303.15 K. Then EGDMA (4.0 mmol, crosslinker), AIBN (30 mg, initiator), and

WWW.MATERIALSVIEWS.COM

[email protected] (15 mg) were subsequently added into the mixtures, and the obtained pre-polymerization solution was ultrasonicated for another 1 h at 303.15 K. After that, the pre-polymerization solution was sealed and deoxygenated with a stream of nitrogen, and stirred continuously at 289.15 K for 24 h under N2 atmosphere. [email protected]@GO was magnetically separated and then rinsed repeatedly with methanol–acetic acid (9:1, v/v) solution until no template was detected by HPLC in the rinsing solution. After acetic acid was removed by methanol, [email protected]@GO was dried under vacuum for 12 h. The non-molecularly imprinted polymer (NMIP) grafted on the surface of [email protected] nanocomposites ([email protected]@GO) was prepared using the same procedure but in the absence of the template (EVO). Characterization of the Superparamagnetic Nanocomposites The transmission electron microscopy (TEM) images of [email protected]@GO were measured by a JEOL JEM-2100 transmission electron microscopy. Fourier transform infrared spectroscopy (FTIR) spectra were obtained on a Nicolet 5700 instrument (Nicolet, Thermo Company, Westerville, USA) in the range of 4000–400 cm21 with the KBr pellet method. The magnetic properties were performed on a vibrating sample magnetometer (VSM) (Lakeshore 7407, Westerville, USA) at 300.15 K. Sample Solution Preparation Pulverized E. fructus (1.0 g) in 50 mL methanol was ultrasonicated for 50 min with an ultrasonic power of 420 W. After the ultrasonic assisted extraction, the extract was centrifuged for 20 min at 6000 rpm. An aliquot (1 mL) of the supernatant was diluted to 10 mL with methanol and this diluted solution was used for sample solution. HPLC Conditions According to our previous work,17 the analytes were determined by HPLC (Agilent 1100, Agilent Technologies, Santa Clara, USA) with a 4.6 mm 3 200 mm, i.d., 5 lm, Hypersil BDS C18 column (Elite, Dalian, China) at 303.15 K, and the analytes were detected at 225 nm. The mobile phase was composed of methanol and water (methanol–water 5 68:32, v:v) at a flow rate of 1.0 mL min21. Binding Experiments To investigate the binding capacities of the adsorbents, the static binding tests were performed in methanol media. Typically, 20 mg of [email protected]@GO, [email protected]@GO or other conventional adsorbents (i.e., macroporous resin D-101, silica gel,

44465 (2 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Figure 2. Scheme of preparation of [email protected]@GO nanocomposites. [Color figure can be viewed at wileyonlinelibrary.com]

reverse phase silica gel (RP-C18), and neutral alumina) were added into 20 mL of the analyte solution with different concentrations, and then the mixtures were shaken in a thermostatic shaker for an appropriate length of time; subsequently, the adsorbents in the solution were separated by an external magnetic field or centrifugation, and then the concentration of the free analyte in the supernatant was measured by HPLC. According to the concentrations of analyte before and after adsorption, the adsorption amount (Q) of the analyte on the adsorbent at equilibrium (Qe, mg g21) and time t (Qt, mg g21) is calculated using the following eqs. (1) and (2).18 The relative standard deviations for all data were within 10.0%. ðC0 2Ce ÞV m ðC0 2Ct ÞV Qt 5 m

Qe 5

(1) (2)

where C0 (mg mL21), Ce (mg mL21), and Ct (mg mL21) are the initial concentration of analyte, the concentration of analyte at equilibrium, and at time t, respectively; V (mL) is the volume of the sample solution; and m (g) is the mass of the used adsorbent. RESULTS AND DISCUSSION

Protocol for Preparation of Superparamagnetic [email protected]@GO Nanocomposites Figure 2 illustrated the principle for preparation of [email protected] [email protected] First, GO was prepared by the classic Hummers method.15,16 Then, [email protected] was synthesized by chemical co-precipitating ferric and ferrous salts in the presence of GO.

WWW.MATERIALSVIEWS.COM

Subsequently, in the presence of template (EVO), functional monomer (MAA), crosslinker agent (EGDMA) and [email protected] particles, the polymerization occurred on the surface of [email protected], and then the EVO-based [email protected]@GO was collected by an external magnetic field. Finally, after removal of the template molecule with methanol–acetic acid (9:1, v/v), [email protected]@Fe3O4 was obtained. The preparation method of [email protected]@GO was the same as that for [email protected]@GO, but the former was prepared without EVO as template molecule. In this work, we aimed to selectively and simultaneously separate or enrich both alkaloids (EVO and RUT), in the procedure of preparing [email protected]@GO EVO was selected as the template for two reasons. First, EVO had similar structure to RUT; second, as shown in Figure 1 the molecular volume of RUT was smaller than EVO, which made it easily get into or out of the imprinting cavity formed in MIP. Characterization. The FTIR spectra of GO, [email protected] and [email protected]@GO are shown in Figure 3(a–c). The FeAO stretching peak approximately at 583 cm21 was observed for [email protected] and [email protected]@GO, indicating that Fe3O4 nanoparticles were successfully anchored onto the GO sheets; however, the strength of FeAO stretching peak significantly decreased after MIP coating, indicating that the MIP preparation was successful. The intensity of [email protected] stretching (1720 cm21) remarkably increased when [email protected] was covered by MIP with a large number of [email protected] groups, which also indicated the formation of MIP on the [email protected] surface. The peak at 1630 cm21 could be attributed to the stretching vibration of [email protected] groups. Therefore, the results showed that [email protected]@GO was prepared satisfactorily. In [email protected]@GO, the peaks at around

44465 (3 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Qt of [email protected]@GO and [email protected]@GO increased rapidly, then Qt gradually reached equilibrium. Qe of [email protected] [email protected] was almost four times more than that of [email protected] [email protected], which can be explained by that [email protected]@GO has the specific binding sites matching EVO and [email protected] [email protected] has only poor nonspecific adsorption.

Figure 3. FTIR spectra of (a) GO, (b) [email protected], and (c) [email protected]@GO. [Color figure can be viewed at wileyonlinelibrary.com]

1157 and 1250 cm21 were assigned to CAOAC symmetric and asymmetric stretching vibrations; the peaks at around 1390 and 1460 cm21 were assigned to ACH3 deformation vibration and ACH2-A bending vibration, respectively. The TEM image of [email protected]@GO [Figure 4(A)] showed that GO had a typical flake-like shape with some wrinkles. Moreover, from Figure 4(B,C) it could be distinctly seen that Fe3O4 nanoparticles and MIP were grafted onto GO surfaces and no isolated nanoparticles were observed beyond the GO. The HRTEM image of [email protected]@GO [Figure 4(D)] revealed that the Fe3O4-nanoparticles (oriented-clear fringes) were attached on the surface of GO matrix (random fringes in the background). The magnetic properties of the Fe3O4 nanoparticles, [email protected] and [email protected]@GO were investigated by VSM. Figure 5(a–c) showed the magnetic hysteresis loops of the Fe3O4, [email protected] and [email protected]@GO at 300.15 K, respectively. The absence of magnetic hysteresis indicated that all these particles had superparamagnetic behavior which facilitated magnetic separation and reusability. The saturation magnetization values of the Fe3O4, [email protected], and [email protected]@GO decreased from 54.6 emu g21 to 39.9 emu g21 and 3.0 emu g21 due to the increasing coatings shielding the magnetite effectively. These results further confirmed that [email protected] and [email protected]@GO were prepared successively. Although the addition of nonmagnetic portion (MIP) led to decreased saturation supermagnetizations, the saturation magnetization value of [email protected]@GO was high enough for an easy and quick magnetic separation from the suspension with a conventional magnet as observed in Figure 5 (upper inset). Adsorption Kinetics of [email protected]@GO and [email protected]@GO Nanocomposites For the kinetic study, the effect of adsorption time (t) on EVO adsorption was investigated in 20 mL EVO solution (0.2 mmol L21) containing fixed adsorbent amounts (20 mg) at 293.15 K. The adsorption kinetic curves of [email protected]@GO and [email protected] [email protected] are presented in Figure 6. During the first 100 min

WWW.MATERIALSVIEWS.COM

Figure 4. TEM images of (A) and (B) [email protected]@GO and (C) highresolution TEM (HRTEM) image of the selected area in [email protected]@GO.

44465 (4 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

indicated that EVO binding on the [email protected]@GO and [email protected]@GO followed predominantly the pseudo-secondorder kinetics model, and the overall adsorption process was controlled by chemisorption.20

Figure 5. Magnetization curves at 300.15 K of (a) Fe3O4, (b) [email protected], and (c) [email protected]@GO. The upper inset picture shows the good response of [email protected]@GO under an external magnetic field. [Color figure can be viewed at wileyonlinelibrary.com]

Adsorption Isotherms of [email protected]@GO and [email protected]@GO To evaluate the effect of temperature on Qe for EVO binding onto [email protected]@GO and [email protected]@GO, the adsorption isotherm experiments were performed in 20 mL solutions containing fixed sorbent amounts (20 mg) at the different initial EVO concentrations and different temperatures. As shown in Figure 7, Qe of both [email protected]@GO and [email protected]@GO increased with the increasing of the initial EVO concentration. However, Qe reduced with the increasing temperature, indicating the exothermic nature of the adsorption processes. This performance was attributed to the hydrogen bond interaction between adsorbent and EVO, and the hydrogen bond force was weak at a higher temperature.22 Moreover, under the same

To investigate the kinetic mechanism of the adsorption process, the pseudo-first-order model [eq. (3)] and the pseudo-secondorder model [eq. (4)] were used to fit the experimental data.19–21 Qt 5Qe 2Qe e2K1 t

(3)

Qt 5K2 Qe2 t=ð11K2 Qe tÞ

(4)

where K1 is the pseudo-first-order rate constant, and K2 is the pseudo-second-order rate constant. As shown in Figure 6, the experimental data were correlated by eqs. (3) and (4) in the form of nonlinear regression and the relative parameters are presented in Table I. For [email protected]@GO and [email protected]@GO, the correlation coefficient (R2) values of eq. (4) were higher than those of eq. (3). The results

Figure 6. Adsorption kinetic curves for EVO on [email protected]@GO (䊏) and [email protected]@GO (w) with the fitting to the pseudo-first-order model (–) and pseudo-second-order model (—) in the form of nonlinear regression (20 mg adsorbents in 0.2 mmol L21 EVO solution (20 mL) at 293.15 K). [Color figure can be viewed at wileyonlinelibrary.com]

WWW.MATERIALSVIEWS.COM

Figure 7. Adsorption isothermic curves for EVO on [email protected]@GO and [email protected]@GO at various temperatures and with the fitting to (A) the Langmuir model and (B) the Freundlich model in the form of nonlinear regression (20 mg adsorbents in 20 mL EVO solution). [Color figure can be viewed at wileyonlinelibrary.com]

44465 (5 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Table I. Adsorption Kinetic Parameters for the Adsorption of EVO on [email protected]@GO and [email protected]@GO Pseudo-first order 21

Qe (lmol g

)

21

K1 (min

)

Pseudo-second-order model R

2

21

Qe (lmol g

)

K2 (g lmol21 min21)

R2

[email protected]@GO

16.0132

0.0240

0.9692

17.9392

0.0019

0.9728

[email protected]@GO

4.5248

0.0218

0.9628

5.0297

0.0058

0.9641

conditions [email protected]@GO exhibited a higher Qe than [email protected]@GO in all cases.

For [email protected]@GO, DGo was calculated by eq. (8),28 because the best fit was obtained from the Langmuir model.

It was important to correlate Qe with either theoretical or empirical equations for the design and operation of adsorption systems. In this work, two classical isotherm models, i.e., the Langmuir model [eq. (5)] and Freundlich model [eq. (6)] were employed to fit the equilibrium adsorption data.23–25 The equilibrium adsorption data of [email protected]@GO and [email protected]@GO and nonlinear regression plots of the models are presented in Figure 7, and the relevant parameters calculated from the two isotherm equations are given in Table II.

DG o 52RTlnKL

Qe 5KL Qmax Ce =ð11KL Ce Þ

(5)

Qe 5KF Ce 1=n

(6)

where Qmax is the theoretical maximum monolayer capacity for EVO, and KL is the Langmuir constant related to the affinity of the adsorption sites; n and KF are Freundlich constants, which are related to the adsorption favorability and adsorption capacity, respectively. For [email protected]@GO, over the whole temperature range the correlation coefficient values (R2) of the Freundlich model were a little higher than those of the Langmuir model. KF decreased with the increasing the temperature, implying a weaker adsorption driving force at a higher temperature. Meanwhile, all values of 1/n were less than 1 which meant that [email protected]@GO had a favorable adsorption towards EVO. These results were able to be explained by the follow reasons: (1) [email protected]@GO could interact with EVO on two groups of sites with different affinities, (2) the active site distribution of [email protected]@GO was heterogeneous.25,26 For [email protected]@GO, at all temperatures the correlation coefficient values (R2) of the Langmuir model were higher than those of the Freundlich model. The reason was that [email protected]@GO had only nonspecific interaction with EVO and the active site distribution of [email protected]@GO was homogeneous.23,25 According to the data in Figure 7, three thermodynamic parameters, i.e., standard Gibbs free energy change (DGo, kJ mol21), standard enthalpy change (DHo, kJ mol21) and standard entropy change (DSo, kJ mol21 K21) for the adsorption of EVO on [email protected]@GO and [email protected]@GO were calculated and listed in Table III. Because the best fit was obtained from the Freundlich model, DGo for [email protected]@GO was calculated by eq. (7) as recommended by Huang et al.27 DG o 52nRT

(7)

where n is the characteristic constant in Freundlich model. R (8.314 J mol21 K21) is the gas constant, and T (K) represents the absolute temperature.

WWW.MATERIALSVIEWS.COM

(8)

21

where KL (L mol ) is the constant of Langmuir equation, The relation between DGo and DHo and DSo could be described by Van’t Hoff correlation in eq. (9). Therefore, DHo and DSo were calculated from the slope and intercept of Van’t Hoff plots, respectively. DG o 5DH o 2T DS o

(9)

As shown in Table III, for [email protected]@GO and [email protected] [email protected] the values of DGo were negative over the whole temperatures, indicating that the adsorption processes on both adsorbents were thermodynamically feasible and could occur spontaneously. The negative values of DHo at all temperatures showed the adsorption processes of [email protected]@GO and [email protected]@GO had the exothermic nature, which confirmed that Qe for EVO binding on both [email protected]@GO and [email protected]@GO reduced with the increasing temperature. For [email protected]@GO, the negative DSo reflected the decrease of randomness on the solid-solute interface.28 For [email protected] [email protected], the positive values of DSo suggested that the randomness increased on the solid–solution interface during the adsorption of EVO on [email protected]@GO.28 Desorption and Reusability of [email protected]@GO To achieve the recycle of the adsorbents, EVO loaded on the adsorbents should be desorbed efficiently without reducing the performance of the adsorbents. First, EVO adsorption was investigated in 10 mL solutions containing 10 mg [email protected]@GO at 293.15 K. The mixed solution was oscillated at 293.15 K for 8 h, and then the adsorbents were separated from the mixture under a magnetic field. Second, the recovered [email protected]@GO was eluted by acetic acid-methanol (1:9, v/v) at 293.15 K repeatedly until EVO was no longer detected by HPLC in the extraction media. Third, the eluted [email protected]@GO was reused for another adsorption cycle. Such adsorption–desorption cycles were repeated five times, and the results are shown in Figure 8. The results showed that the changes of Qe for EVO on [email protected]@GO were very small after five regeneration cycles, suggesting that [email protected]@GO kept its good affinity and excellent reusability during several adsorption–desorption cycles. Recognition Selectivity of [email protected]@GO and [email protected]@GO To further evaluate the recognition selectivity of [email protected]@GO and [email protected]@GO, Qe for RUT and LIM as reference compounds were measured in 10 mL solutions with

44465 (6 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Table II. Parameters of Isotherm Models of the Adsorption for EVO Adsorption onto [email protected]@GO and [email protected]@GO at Various Temperatures Langmuir model 21

Adsorbents

Temperatures (K)

Qmax (lmol g

[email protected]@GO

293.15

15.5819

303.15

14.4226

[email protected]@GO

)

Freundlich model R

n

KF

R2

68.5680

0.9268

6.9447

16.5269

0.9273

42.6539

0.9251

5.6399

15.3665

0.9576

21

KL (mL lmol

)

2

313.15

12.8705

23.1189

0.9427

5.1573

13.6657

0.9494

293.15

5.1225

33.5390

0.9877

12.2024

5.2985

0.8897

303.15

4.8273

25.1067

0.9638

8.3079

5.0465

0.9050

313.15

4.8248

19.171

0.9846

4.8047

5.0723

0.9374

0.1 mmol L21 analyte containing 10 mg adsorbent at 293.15 K. Meanwhile, Qe on four conventional sorbents, i.e., macroporous resin D-101, silica gel, reverse phase silica gel (RP-C18) and neutral alumina, was also measured. Qe for EVO, RUT and LIM on the six adsorbents are presented in Figure 9. Among the six adsorbents, [email protected]@GO had the highest Qe for all of the three analytes. Compared with [email protected]@GO, [email protected] [email protected] had a much higher adsorption capacity for EVO and RUT because of the specific recognition sites in [email protected] [email protected] On [email protected]@GO, Qe for EVO and RUT was much higher than that for LIM, which suggested that the [email protected]@GO had good binding selectivity for the template molecule; Qe for RUT was higher than that towards EVO, which could be explained that RUT had a similar and smaller molecular structure compared with EVO, making it easy to enter the imprinting cavity and be adsorbed onto [email protected]@GO. These results showed that [email protected]@GO had higher adsorption capacity and recognition selectivity, so it could be used to simultaneously enrich both EVO and RUT. Simultaneous Separation and Enrichment of EVO and RUT from the Mixture Solution Containing EVO, RUT, and LIM Standards After the evaluation of the adsorption efficiency, simultaneous separation and enrichment of EVO and RUT with [email protected] [email protected] and [email protected]@GO from the mixture solution containing EVO, RUT and LIM standards were further investigated. The enrichment process included three steps. On the first step (adsorption), 10 mg of [email protected]@GO or [email protected] [email protected] was oscillated in the standard mixture solution (0.1 mmol L21 EVO, 0.1 mmol L21 RUT and 0.1 mmol L21 LIM) at 293.15 K for 8 h, and then the superparamagnetic adsorbent in the mixture was separated by an external magnetic field; On the second step (rinsing), the adsorbent dried by nitrogen current was rinsed by 10 mL methanol–water (1:1, v/v); On the last

step (eluting), the adsorbent was eluted by 10 mL methanol/ acetic acid mixture (9:1, v/v). After the enrichment, the solution on each step was analyzed by HPLC, and the HPLC chromatograms are presented in Figure 10. For [email protected]@GO, compared those in Figure 10A(a), after adsorption the peak heights of EVO, RUT and LIM in Figure 10A(b) had a very small change, indicating that the adsorption amounts of EVO, RUT and LIM on [email protected]@GO were very small. After being rinsed with methanol–water (1:1, v/v) Figure 10A(c) showed that EVO, RUT and LIM were rinsed off from [email protected]@GO. Finally, as shown in Figure 10A(d), nearly neither EVO nor RUT was recovered after being eluted by methanol/acetic acid (9:1, v/v). The results of Figure 10(A) indicated that EVO, RUT and LIM had similar and weak affinity on [email protected]@GO, and [email protected]@GO had no significant selectivity and enrichment of EVO and RUT. For [email protected]@GO, the peak heights of EVO and RUT in Figure 10B(b) became clearly lower than those in Figure 10B(a), indicating that [email protected]@GO had higher adsorption capacity than [email protected]@GO. Figure 10B(d) showed that there were mainly EVO and RUT in the eluting solution, while nearly no LIM was presented in the solution. Furthermore, after the eluting step 81.25% of EVO and 72.36% of RUT were recovered from the initial mixtures. Compared with the recovery of RUT, the higher recovery of EVO also verified the selectivity of [email protected]@GO for EVO because EVO was used as a template for preparing MIP. These results suggested that EVO and RUT had a stronger affinity and higher imprinting efficiency on [email protected]@GO than LIM due to the selective molecular recognition, and [email protected]@GO could be used to simultaneously enrich both EVO and RUT. Application of [email protected]@GO on the Extracts of E. fructus Finally, [email protected]@GO was employed to simultaneously enrich and separate EVO and RUT in the extract of E. fructus.

Table III. The Thermodynamic Parameters of the Adsorption Process for EVO on [email protected]@GO and [email protected]@GO at Various Temperatures [email protected]@GO Thermodynamic parameters o

21

DG (kJ mol

[email protected]@GO

293.15 K

303.15 K

313.15 K

214.2171

213.4293

)

216.9288

DHo (kJ mol21)

267.9105

DSo (kJ mol21 K21)

20.1750

WWW.MATERIALSVIEWS.COM

293.15 K

303.15 K

313.15 K

225.4016

225.5380

225.6780

221.3484 0.0138

44465 (7 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

Figure 8. Reusability of the [email protected]@GO for EVO (10 mg adsorbents in 0.1 mmol L21 EVO solution (10 mL) at 293.15 K). [Color figure can be viewed at wileyonlinelibrary.com]

The enrichment process was similar to that for the standard mixture solution. As shown in Figure 11(a) the components in the untreated extract of E. fructus were very complicated, and a lot of unknown components in addition to EVO, RUT and LIM existed in the extract. After being adsorbed on [email protected] [email protected], Figure 11(b) showed that the content of EVO and RUT decreased significantly, while the content of LIM had a small change, indicating the specific affinity of [email protected]@GO for EVO and RUT. In Figure 11(d), the chromatogram of eluting solution showed that EVO and RUT were enriched in the eluting solution; the amount of the unknown components significantly decreased compared with the untreated extract solution; and the recovery of EVO and RUT were 82.17% and

Figure 9. Adsorption capacities toward EVO and reference compounds (RUT and LIM) on six sorbents ([email protected]@GO, [email protected]@GO, silica gel, reverse phase silica gel (RP-C18), neutral alumina, and macroporous resin D-101) (10 mg adsorbent, 0.1 mmol L21 EVO, RUT, and LIM solution (10 mL), respectively, at 293.15 K).

WWW.MATERIALSVIEWS.COM

Figure 10. HPLC chromatograms of the standard mixtures of EVO, RUT, and LIM separated by [email protected]@GO (A) and [email protected]@GO (B). (a) Untreated standard mixture before being adsorbed; (b) fraction after being adsorbed; (c) fraction rinsed from the adsorbent by methanol–water (1:1, v/v); (d) fraction eluted from the sorbent by methanol–acetic acid (9:1, v/v). [Color figure can be viewed at wileyonlinelibrary.com]

68.24%. The higher recovery of EVO also verified the selectivity of [email protected]@GO for EVO. Therefore, the [email protected]@GO had achieved the desired result that EVO and RUT were simultaneously enriched and separated from the initial extracts.

Figure 11. HPLC chromatograms of (a) the untreated extract of E. fructus; (b) the extract of E. fructus after being adsorbed by [email protected]@GO; (c) fraction rinsed from [email protected]@GO by methanol–water (1:1, v/v); (d) fraction eluted from [email protected]@GO methanol–acetic acid (9:1, v/v). [Color figure can be viewed at wileyonlinelibrary.com]

44465 (8 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465

ARTICLE

WILEYONLINELIBRARY.COM/APP

CONCLUSIONS

To sum up, a nanocomposite ([email protected]@GO) for simultaneous recognition and enrichment of EVO and RUT was prepared using [email protected], MAA and EVO as the supporting material, monomer and template molecule, respectively. The asprepared [email protected]@GO was characterized by FTIR spectra, TEM and VSM, and the results confirmed the formation of the superparamagnetic nanocomposite. The adsorption kinetic study showed that the adsorption process of [email protected]@GO was well fitted by the pseudo-second-order model. The isotherm study showed that the adsorption behavior on [email protected]@GO was thermodynamically feasible and could occur spontaneously, and the equilibrium adsorption data were well fitted by the Freundlich model. The selective recognition and enrichment experiments indicated that [email protected]@GO could be used to simultaneously enrich both EVO and RUT. In addition, [email protected]@GO displayed satisfactory recoveries for EVO and RUT, when they were used for enrichment of EVO and RUT from the extract of E. fructus. [email protected]@GO could provide an excellent platform for pretreatment and enrichment of EVO and RUT from the extract of E. fructus. ACKNOWLEDGMENTS

Financial support from the National Natural Science Foundation of China (grant numbers 21366019, 20806037, and 20876131), Jiangxi Province Young Scientists (Jinggang Star) Cultivation Plan (grant number 20112BCB23002), Jiangxi Province Higher School Science and Technology Landing Plan Projects (grant number KJLD13012), Special Funds for Graduate Student Innovation in Jiangxi Province (grant number YC2014-S013), and Jiangxi Province Undergraduate Innovation and Entrepreneurship Training Program (grant number 201310403040) are gratefully acknowledged.

6. Ko, H. C.; Wang, Y. H.; Liou, K. T.; Chen, C. M.; Chen, C. H.; Wang, W. Y.; Chang, S.; Hou, Y. C.; Chen, K. T.; Chen, C. F.; Shen, Y. C. Eur. J. Pharmacol. 2007, 555, 211. 7. Sheu, J. R. Cardiovasc. Drug Rev. 1999, 17, 237. 8. Jayakumar, T.; Sheu, J. R. J. Exp. Clin. Med. 2011, 3, 63. 9. Chen, L.; Li, B. Food Chem. 2013, 141, 23. 10. Pan, S. D.; Shen, H. Y.; Zhou, L. X.; Chen, X. H.; Zhao, Y. G.; Cai, M. Q.; Jin, M. C. J. Mater. Chem. A 2014, 2, 15345. 11. Ning, F.; Peng, H.; Li, J.; Chen, L.; Xiong, H. J. Agric. Food Chem. 2014, 62, 7436. 12. Han, S.; Li, X.; Wang, Y.; Chen, S. Chem. Eng. J. 2015, 27, 187. 13. Fan, J. P.; Zhang, L.; Zhang, X. H.; Huang, J.; Tong, S.; Kong, T.; Tian, Z. Y.; Zhu, J. H. Anal. Bioanal. Chem. 2012, 402, 1337. 14. Pan, Y.; Zhang, Q.; Li, Z.; Wang, Q.; Ren, C.; Li, S.; Li, H.; Zhao, Q. J. Appl. Polym. Sci. 2016, DOI: 10.1002/APP.43806. 15. Qiu, H.; Luo, C.; Sun, M.; Lu, F.; Fan, L.; Li, X. Carbon 2012, 50, 4052. 16. Li, Y.; Li, X.; Dong, C.; Qi, J.; Han, X. Carbon 2010, 48, 3427. 17. Fan, J. P.; Yang, X. M.; Xu, X. K.; Xie, Y. L.; Zhang, X. H. J. Chem. Thermodyn. 2015, 83, 85. 18. Fan, J. P.; Li, L.; Tian, Z. Y.; Xie, C. F.; Xue, Y.; Xie, Y. L.; Xu, R.; Qin, Y.; Zhang, X. H.; Zhu, J. H. Sep. Sci. Technol. 2014, 49, 258. 19. Fan, J. P.; Zheng, B.; Qin, Y.; Yang, D.; Liao, D. D.; Xu, X. K.; Zhang, X. H.; Zhu, J. H. Appl. Surf. Sci. 2016, 364, 332. 20. Song, R.; Hu, X.; Guan, P.; Li, J.; Qian, L.; Wang, C.; Wang, Q. Appl. Surf. Sci. 2015, 332, 159. 21. Fan, J. P.; Tian, Z. Y.; Tong, S.; Zhang, X. H.; Xie, Y. L.; Xu, R.; Qin, Y.; Li, L. Food Chem. 2013, 141, 3578. 22. Cao, Y.; Liu, L.; Xu, W.; Wu, X.; Huang, W. J. Appl. Polym. Sci. 2014, 13, 1378.

REFERENCES

23. Fan, J. P.; Li, L.; Tian, Z. Y.; Xie, C. F.; Song, F. T.; Zhang, X. H.; Zhu, J. H. J. Memb. Sci. 2014, 467, 13.

1. Jiang, J.; Hu, C. Molecules 2009, 14, 1852. 2. Chinese Pharmacopoeia Committee. Pharmacopoeia of the People’s Republic of China; Beijing: Chemical Industry Press, 2010, Vol. 1, p 160. 3. De Petrocellis, L.; Schiano Moriello, A.; Fontana, G.; Sacchetti, A.; Passarella, D.; Appendino, G.; Di Marzo, V. Br. J. Pharmacol. 2014, 171, 2608. 4. Liu, A. J.; Wang, S. H.; Chen, K. C.; Kuei, H. P.; Shih, Y. L.; Hou, S. Y.; Chiu, W. T.; Hsiao, S. H.; Shih, C. M. Chem. Biol. Interact. 2013, 205, 20. 5. Huang, Y. C.; Guh, J. H.; Teng, C. M. Life Sci. 2004, 75, 35.

WWW.MATERIALSVIEWS.COM

24. Li, H.; Xu, M.; Wang, S.; Lu, C.; Li, Z. Appl. Surf. Sci. 2015, 328, 649. 25. Yang, W.; Zhou, W.; Xu, W.; Li, H.; Huang, W.; Jiang, B.; Zhou, Z.; Yan, Y. J. Chem. Eng. Data 2012, 57, 1713. 26. Al-Asheh, S.; Banat, F.; Al-Omari, R.; Duvnjak, Z. Chemosphere 2000, 41, 659. 27. Huang, J.; Deng, R.; Huang, K. Chem. Eng. J. 2011, 171, 951. 28. Fan, J. P.; Xu, X. K.; Xu, R.; Zhang, X. H.; Zhu, J. H. Chem. Eng. J. 2015, 279, 567.

44465 (9 of 9)

J. APPL. POLYM. SCI. 2016, DOI: 10.1002/APP.44465