Salicylic Acid Complex Imprinted Polymer Membranes: Preparation and Separation Characteristics

Abstract

Salicylic acid complex imprinted polymeric membranes have been prepared by thermal polymerization using polyvinilidene fluoride membrane as a support membrane, acrylamide as a functional monomer and ethylene glycol dimethacrylate as a crosslinker. The imprinted membranes are characterized by Fourier Transform Infrared Spectrometer (FT-IR), thermogravimetric analysis (TGA) and scanning electron microscope (SEM). It can be found that the structure of imprinted membranes is different from the supporting membrane, and the surface of the supporting membrane is covered with an imprinted polymer layer after polymerization. Adsorption and permeation binding experiments indicate that imprinted membranes show high specific binding capacity of template molecule, the selectivity factor of the imprinted membrane for the salicylic acid was 26.4, The optimum polymerization condition is 0.3 mmol salicylic acid, 1.2 mmol acrylamide and 6.0 mmol ethylene glycol dimethacrylate in acetonitrile at 60℃ for 24 h.

1. Introduction

Molecularly imprinted technology (MIT) has attracted considerable attention in the last few years¹⁾. This technique requires functional monomers surrounding a template molecule in liquor, and the mixed liquor can be polymerized by thermal initiation or irradiation with UV light. As a result of the reaction, target molecule is surrounded by functional monomers, and a selected interaction is secured. Adding a crosslinker in the polymerization process provides the imprinted polymers with good mechanical strength, stability and rigidity. Finally, template molecule was removed by extraction with suitable solvents. The separation effect of molecular imprinted polymers (MIP) is dictated by the highly specific binding of template molecules²⁾. MIT has been broadly used in various research fields, including sensors^3,4), drug separation^5–7), chromatographic separation^8,9) and biological receptors¹⁰⁾.

Molecularly imprinted membrane (MIM) has the advantages of combining the selectivity and mechanical properties of the support membrane compared with MIP.

MIM which has micro-porous structure can ease template transport and MIM which has transmembrane macropores can be employed as adsorbing material. Typical MIM syntheses which had also been adapted to membrane preparations are based on in situ crosslinking polymerization of functional and crosslinker monomers yielding either self-supported¹¹⁾ or thin layer composite MIM¹²⁾. MIM was firstly prepared by Piletsky et al¹³⁾. And later, various approaches of preparing MIM were reported, such as surface imprinting, in situ polymerization, dry or wet phase separation and particles composite membrane¹⁴⁾. Hong¹⁵⁾ prepared a composite imprinted membrane by photo polymerization on the surface of a microporous alumina support membrane. In their study, the results showed that the flux was dramatically improved to 10⁻⁸ mol/(cm²·h) and high selectivity factor (α = 5.0) was tested for the template relative to its analogue.

Salicylic acid (SA) is used as a topical keratolytic and as an external antiseptic and antifungal has been widely applied in many pharmaceutical and cosmetic formulations. It is very important to ensure the quality of SA in biological fluids, and a number of SA measured methods such as chromatographic method and ultraviolet spectrometry method have been developed. However, there are some disadvantages of these methods such as low selectivity, inconvenience and inefficiency. MIM can be characterized by selective transport of SA and rejection of other molecules, which is easier and more efficient.

The aim of this study is to prepare a molecularly imprinted membrane for selective SA recognition. The membrane was prepared using a polyvinilidene fluoride membrane as the support, SA as the template, acrylamide (AM) as the functional monomer and ethylene glycol dimethacrylate (EGDMA) as the cross-linker. The chemical structures of the MIM were analyzed by FTIR spectrometry, the morphologies of MIM were characterized by SEM, and the thermal stability of MIM was analyzed by TGA. To evaluate the adsorption properties and selectivity of MIM, adsorption experiments were investigated by UV-Vis spectrophotometry. Importantly, the optimal experimental conditions have been clearly studied.

2. Experimental

2.1 Materials

Acrylamide (AM), Salicylic acid (SA) and acetosalicylic acid were purchased from Sinopharm Chemical Reagent Co., Ltd (China), polyvinilidene fluoride (PVDF) membranes with a nominal pore size dn = 0.45 µm and a diameter of 20.5 mm were purchased from Linyi Zhengheng chemical Glass Instrument Co., Ltd (China), EGDMA (98%) was purchased from Sigma–Aldrich (China) and was purified by distillation under vacuum. 2,2-Azobiisobutyronitrile (AIBN) was purchased from Shanghai No.4 Reagent & H.V Chemical Co., Ltd (China). Acetonitrile, methanol and sodium hydroxide (NaOH) were purchased from Xilong Chemical Co., Ltd (China).

2.2 Instruments and apparatus

To measure the UV spectrum of SA in aqueous solution, a UV-5600 spectrometer (Shanghai Yuanxi Analysis Instrument Factory, China) was used. The surface features of membranes were analyzed by an S-3400N SEM (Hitachi, Tokyo, Japan). FTIR spectra (4000–400 cm⁻¹) were recorded using an IS10 FTIR spectrometer (Nicolet, USA). TGA of the samples was performed on a STA 449 F3 analyzer (Netzsch, Germany) over a temperature range of 25–800℃ at 15℃ min⁻¹ under N₂. X-ray diffraction (XRD) patterns were collected on a Rigaku D/max-2700 powder X-ray diffractometer (0.08° step/s).

2.3 Synthesis of the imprinted membrane

Circular polyvinilidene fluoride (PVDF) membranes were precoated in a 0.0150 g/mL solution of AIBN in acetonitrile and dried under vacuum for further use. To a stirred solution of salicylic acid, the monomer (AM) and the crosslinker (EGDMA) in acetonitrile was degassed for 5 min with nitrogen. The PVDF membrane precoated with initiator was coated by the solution and dried under nitrogen. The membrane was removed and clamped between two plates of glass, sealed and heated at 60℃ for 24 h. The membrane was then treated with a methanol/0.1 molL⁻¹ sodium aqueous hydroxide solution (4/1, v/v) solution to remove salicylic acid, AIBN and EGDMA. Finally, the membranes were dried at room temperature. For a comparative purpose non-imprinted membrane (NIM) were also prepared under the same conditions without adding any salicylic acid during preparation. The prepared formulations are described in Table 1.

Table 1 Formulation of the composite imprinted membranes.

NO.	1	2	3	4	5	6
SA (mmol)	0.10	0.15	0.20	0.30	0.40	0.60
AM (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
AIBN (mg)	15	15	15	15	15	15
EGDMA (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
NO.	7	8	4	9	10	11
SA (mmol)	0.30	0.30	0.30	0.30	0.30	0.30
AM (mmol)	1.8	1.5	1.2	0.8	0.6	0.4
AIBN (mg)	15	15	15	15	15	15
EGDMA (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
NO.	12	13	14	4	15	16
SA (mmol)	0.30	0.30	0.30	0.30	0.30	0.30
AM (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
AIBN (mg)	5.0	8.0	12	15	18	22
EGDMA (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
NO.	14	17	18	19	20	21
SA (mmol)	0.30	0.30	0.30	0.30	0.30	0.30
AM (mmol)	1.2	1.2	1.2	1.2	1.2	1.2
AIBN (mg)	12	12	12	12	12	12
EGDMA (mmol)	1.2	3	4.8	6	7.2	7.8

2.4 Swelling ratio of the membrane

Wet membrane was prepared by immersing the dry membrane in deionized water for 24 h at room temperature, and the membrane was centrifuged at 3000 rpm for 2 min and wiped water on filter paper. Swelling ratio (δ) of the membrane was calculated by eq. (1):¹⁶⁾   

\[\delta = \frac{{\rm V}_2 - {\rm V}_1}{{\rm V}_1} \times 100\%\](1)

Where V₁ and V₂ are the volumes of the dry and wet membrane, respectively.

2.5 Porosity of the membrane

Wet membrane was prepared by immersing the dry membrane in deionized water for 72 h at room temperature, the membrane was centrifuged at 3000 rpm for 2 min and weight. The porosity (Pr) of the membrane was calculated by eq. (2):¹⁶⁾   

\[\Pr = \frac{({\rm w}_2 - {\rm w}_1)/\rho_{{\rm H}_2{\rm O}}}{{\rm V}_{\rm MIM} - {\rm V}_{\rm PVDF}} \times 100\%\](2)

Where V_MIM and V_PVDF are volumes of MIM and PVDF, w₁ and w₂ are the weight of the dry and wet membrane, respectively, and $\rho_{{\rm H}_2{\rm O}}$ is water density.

2.6 Binding tests

Static binding experiments were conducted to evaluate the recognition property of the membranes toward the target molecule. Imprinted and blank control membrane samples were immersed in 10 ml acetonitrile containing 0.15 mg/mL of SA to test their capability to SA at room temperature for 24 h. The amount of SA adsorbed by the membrane Q is calculated from eq. (3):   

\[{\rm Q} = \frac{({\rm C}_1 - {\rm C}_2)}{{\rm m}} \times {\rm V}\](3)

Where, C₁ is the concentration of SA in the feed solution, C₂ is the final concentration in the collected permeate, V is the volume and m is the mass of the dried membrane. Blank control membranes were conducted for the same procedure.

The selectivity factor was calculated using the eq. (4)¹⁷⁾:   

\[\alpha = \frac{{\rm R}_{\rm template}}{{\rm R}_{\rm analogue}}\](4)

Where, α is the selectivity factor, R_template is the amount of template retained by imprinted the membrane and R_analogue is the amount of the structural analogue retained by the membrane imprinted with the template.

3. Results and Discussion

3.1 Swelling ratio

From all the data given in the Table 2, the average swelling ratio was 7.55%, which showed that the membrane can be used in distilled water with a long duration. The structure of membrane was not changed, which indicates that the separation ability of the membrane is stable¹⁸⁾.

Table 2 Thickness and swelling of the imprinted membrane around immersion.

NO.	h1/mm	h2/mm	d1/mm	d2/mm	S1/mm2	S2/mm2	V1/mL	V2/mL	δ (%)
1	0.139	0.140	19.96	20.54	313	331	0.0435	0.0464	6.66
2	0.119	0.121	19.48	20.00	298	314	0.0354	0.0380	7.18
3	0.129	0.132	19.68	20.18	304	320	0.0392	0.0422	7.59
4	0.169	0.175	18.96	19.30	282	292	0.0477	0.0512	7.30
5	0.179	0.183	20.10	20.58	317	332	0.0568	0.0608	7.18
6	0.149	0.152	20.02	20.58	315	332	0.0469	0.0505	7.80
7	0.180	0.188	20.54	20.76	331	338	0.0596	0.0636	6.69
8	0.168	0.180	20.04	20.08	315	317	0.0530	0.0570	7.57
9	0.153	0.157	20.24	20.72	322	337	0.0492	0.0529	7.54
10	0.190	0.192	18.94	19.50	282	298	0.0535	0.0573	7.12
11	0.149	0.152	20.22	20.60	321	333	0.0478	0.0506	5.88
12	0.142	0.149	20.18	20.46	320	329	0.0454	0.0490	7.86
13	0.149	0.152	20.24	20.72	322	337	0.0479	0.0512	6.91
14	0.159	0.162	19.92	20.46	311	329	0.0495	0.0532	7.49
15	0.119	0.121	19.48	20.00	298	314	0.0354	0.0380	7.18
16	0.188	0.190	20.50	21.02	330	347	0.0620	0.0659	6.26
17	0.150	0.159	19.88	20.00	310	314	0.0465	0.0499	7.28
18	0.153	0.158	18.98	19.32	283	293	0.0433	0.0463	7.00
19	0.146	0.152	18.96	19.24	282	290	0.0412	0.0442	7.21
20	0.191	0.200	19.96	20.22	313	321	0.0597	0.0642	7.46
21	0.168	0.170	20.20	20.84	320	341	0.0538	0.0580	7.70

3.2 Porosity

From all the data given in the Table 3, porosity (Pr) of the membrane was calculated by eq. (2). The porosities of imprinted membrane were 76.90% which was greater than that of non-imprinted membrane (35.14%) due to template molecules. These data proved that the imprinting sites were created after removing template.

Table 3 Main parameters of MIM and NIM.

NO.	w1/g	w2/g	H/mm	S/mm2	V/mL	Pr (%)
SA-MIM	0.0324	0.0498	0.146	282	0.0412	76.9
NMIN	0.0262	0.0353	0.161	276	0.0445	35.1
PVDF			0.062	298	0.0185

3.3 Polymer preparation and characterization

To confirm the presence of the co-monomers in the synthesized copolymers, the FTIR spectra of the polyvinilidene fluoride (PVDF) membranes and imprinted membrane were performed. As shown in Fig. 1(a) the absorption peaks at 687, 1200, 1270, 1460, 1530 and 1640 cm⁻¹ indicate the existence of PVDF. The peak at 2940 cm⁻¹ corresponds to dissymmetry stretching vibration of –CH₂– there are no apparent adsorption bands within 2960–2940 which shows that there is no –CH₂– in the PVDF polymer. Due to the –C=O stretching vibrations, a peak at 1720 cm⁻¹ existed in imprinted membrane does not present in the PVDF's spectrum (Fig. 1(b)). The features around 1140 cm⁻¹ indicate the stretching vibration of C–N, and the absorption peak of 1473 and 1623 cm⁻¹ are assigned to the carbon skeleton on phenyl ring of SA. It is inferred that the surface of the polyvinilidene fluoride (PVDF) membrane has been covered by the imprinted polymer layer, leading to the typical peaks of PVDF disappeared and the typical peaks of imprinted polymer appeared^19,20).

Fig. 1

FT-IR spectra of PVDF membrane (a) and imprinted membrane (b).

Figure 2 shows the SEM images of the surface morphologies of PVDF membrane, NIM and MIM. It can be inferred from Fig. 2(a) that the network PVDF membrane has obvious porous structure. Polymerization reaction makes a great modification on the PVDF membrane. Figure 2(b) indicates that the surface of membrane was coated with polymer layer. There is no pore could be observed on the surface of the NIM. Figure 2(c) and 2(d) show that there are a large number of pores can be found on the surface of MIM, and these pores are different from those of the blank membrane. The surface porous property has a great effect on the ultimate capability of the imprint composite membranes²¹⁾.

Fig. 2

SEM images of the membranes: (a) PVDF membrane (10000×), (b) the nonimprinted membrane (10000×), (c) the imprinted membrane (15000×) and (d) the imprinted membrane (30000×).

TGA²²⁾ plots of MIM and NIM were shown in Fig. 3. There was a slight weight loss of the two samples at temperatures between 50 and 350℃, due to the loss of adsorbed water, acetonitrile, methanol and other volatile residuals. In the case of MIM, the large weight loss region between 350 and 450℃ indicates a faster weight loss rate of MIM compared with that of NIM, which is because of the number of pores on the surface of MIM and the open structure of MIM. Thus, the thermal stability of MIM is poorer than that of NIM. After 480℃, both of the membranes are completely decomposed.

Fig. 3

Thermogravimetric curves of MIM and NIM.

3.4 Optimization of reaction conditions

To investigate the effect of reaction conditions on the binding capability of the molecularly imprinted membranes (MIMs), different amounts of SA, AM, EGDMA and AIBN were investigated, and the specific binding capacities were also analyzed.

Table 4 indicates that the quantity of the MIM increased as the elevated amount of SA, and the quantity of the MIM increased to the highest value of 6.11 when the concentration of SA came to 0.30 mmol. After that, the adsorption quantity decreased with further increase of SA. This phenomenon can be explained as follow: at the initial stage of polymerization, a little amount of SA was put in, and then reacted with AM, which results in the imprinted membranes had low binding capacity. With the increase of SA, the binding capacity began to improve, until the amount of SA reached 0.30 mmol. However, it was difficult to produce much more imprinted cavities when excessive SA was added, due to the amount of AINB and AM was constant. What's more, if excessive SA was used, SA would be imprinted on the same site which made the imprinted cavities incomplete, even resulting in the decrease of adsorption quantity.

Table 4 Effects of dosage of template molecule on adsorption performance of the imprinted membranes.

NO.	1	2	3	4	5	6
SA (mmol)	0.10	0.15	0.20	0.30	0.40	0.60
Quantity (µg/g)	2.46	5.08	5.74	6.11	5.83	4.05

Table 5 illustrates that absorbability increased until the content of functional monomer came to 1.2 mmol. After that, absorbability decreased, possibly because AIBN, SA and EGDMA had already completely used up and an excessive amount of AM could not react with AIBN, SA and EGDMA, which resulted in the adsorption quantity decrease.

Table 5 Effects of amount of functional monomer on adsorption performance of the imprinted membranes.

NO.	7	8	4	9	10	11
AM (mmol)	1.8	1.5	1.2	0.8	0.6	0.4
Quantity (µg/g)	2.84	3.49	6.11	4.33	4.05	3.21

Table 6 shows that the absorbability of the MIM increased to 7.52 µg/g while the amount of AIBN reached 12 mg. The binding capacity decreased when higher concentrations of AIBN were used and the amounts of SA, AM and EGDMA were unchanged.

Table 6 Effects of dosage of initiator reagent on adsorption performance of the imprinted membranes.

NO.	12	13	14	4	15	16
AIBN (mg)	5	8	12	15	18	22
Quantity (µg/g)	4.99	7.24	7.52	6.11	5.27	4.90

Table 7 indicates that while the amount of EGDMA comes to 6.0 mmol, the adsorption quantity of the MIM increased to 9.86 µg/g, at which the absorbability reached maximum and then dropped down even EGDMA dosage increased. This is possibly because as the increase of EGDMA, imprinted cavities increased gradually and became stable, moreover, it was not conductive to removal of template.

Table 7 Effects of dosage of crosslinker on adsorption performance of the imprinted membranes.

NO.	14	17	18	19	20	21
EGDMA (mmol)	1.2	3	4.8	6	7.2	7.8
Quantity (µg/g)	7.52	7.80	9.20	9.86	7.61	7.14

Overall, the maximum absorbability of the MIM was obtained when the reaction condition was optimized to 0.3 mmol SA, 12 mg AIBN, 6.0 mmol EGDMA and 1.2 mmol AM at 60℃ for 24 h.

3.5 Binding isotherms

The binding performance of the imprinted membrane was studied using static binding tests and the Scatchard analysis.   

\[\frac{Q}{C} = \frac{Q_{\max} - Q}{K_d}\](5)

Where K_d is the equilibrium dissociation constant, Q_max is the maximum adsorption capacity, C is the final equilibrium concentration of SA in solution, Q is the adsorption capacity of caffeine adsorbed on the membrane at equilibrium.

The binding isotherms plots of MIM and NIM in Fig. 4(a) show that MIM can adsorb much more SA than NIM. The adsorption data were plotted using the Scatchard equation and shown in Fig. 4(b), indicating that the binding sites in the MIM were heterogeneous. The data points are nonlinear between Q/C and Q. However, a straight line could be obtained for 0 ≤ Q ≤ 8.98 µg/g and 8.98 ≤ Q ≤ 11.45 µg/g. The linearized plots of the two straight lines were obtained. For 0 ≤ Q ≤ 8.98 µg/g, the equation of linear regression was y = −0.0123x + 0.155, the related coefficient was 0.998. For 8.98 ≤ Q ≤ 11.45 µg/g, the equation of linear regression was y = −0.00651x + 0.103, and the related coefficient was 0.998. First dissociation constant K_d1 of the high-affinity sites was 81.5 µg/L and the corresponding Q_max1 was 12.6 µg/g. The dissociation constant, K_d2 of the low-affinity sites was 153.6 µg/L, and Q_max2 was 15.8 µg/g. These results indicate that there were two binding sites in the MIM, and they showed different affinities for SA when the concentration was the range of 0～0.40 g/L. The reason for this is because the imprinted SA has two binding sites¹⁸⁾.

Fig. 4

Adsorption isotherms of MIM and NIM (a) and scatchard plots of MIM (b).

As shown in Table 8, the selective properties of SA imprinted membranes were evaluated by using solutions containing SA and aspirin. The concentration of each compound was 0.15 g/L. When the experiments were carried out with imprinted membranes, and imprinted membrane was able to distinguish SA from aspirin.

Table 8 Selective adsorption of the imprinted membrane.

	Q (µg/g)	α
SA	9.67	26.4
aspirin	0.366

4. Conclusions

PVDF membrane substrate covered with a SA imprinted layer was successfully prepared by polymerization of AM as a functional monomer and EGDMA as a crosslinker. The optimum polymerization condition for the preparation is SA:AM:EGDMA = 1:4:20 in acetonitrile at 60℃ for 24 h. The Scatchard model described the adsorption behavior of the imprinted complex membrane suitably, and suggests that two binding sites exist in the membrane, K_d1 = 81.5 µg/L, Q_max1 = 12.6 µg/g and K_d2 = 153.6 µg/L, Q_max2 = 15.8 µg/g, and selectivity factor was 26.4. In brief, the complex imprinted polymeric membranes can be easily synthetized with the method and effectively applied in separating SA, which could be used in alternating the traditional methods for the selective removal of SA and it is significant to devise an efficient method of testing the SA content.

Acknowledgements

This work was supported by the Innovation Experiment Program for university students of Fujian province (No. 201511312051), and School Research Program of Longyan University (LQ2013020).

REFERENCES

• 1)  G. Vlatakis, L.I. Andersson, R. Muller and K. Mosbach: Nature 361 (1993) 645–647. 10.1038/361645a0
• 2)  S.A. Piletsky, T.L. Panasyuk, E.V. Piletskaya, I.A. Nicholls and M. Ulbricht: J. Membr. Sci. 157 (1999) 263–278. 10.1016/S0376-7388(99)00007-1
• 3)  P.A. Lieberzeit and F.L. Dickert: Anal. Bioanal. Chem. 391 (2008) 1629–1639. 10.1007/s00216-008-1909-z
• 4)  A. Rachkov, S. McNiven, A. El'skaya, K. Yano and I. Karube: Anal. Chim. Acta 405 (2000) 23–29. 10.1016/S0003-2670(99)00743-6
• 5)  Q.L. Deng, Z.H. Lun, R.Y. Gao, L.H. Zhang, W.B. Zhang and Y.K. Zhang: Electrophoresis 27 (2006) 4351–4358. 10.1002/elps.200600180
• 6)  S.H. Son and J. Jegal: J. Appl. Polym. Sci. 104 (2007) 1866–1872. 10.1002/app.25845
• 7)  S.A. Zaidi and W.J. Cheong: J. Sep. Sci. 31 (2008) 2962–2970. 10.1002/jssc.200800160
• 8)  F. Liu, C.H. Du, B.K. Zhu and Y.Y. Xu: Polymer 48 (2007) 2910–2918. 10.1016/j.polymer.2007.03.033
• 9)  H.Y. Liu, K.H. Row and G.L. Yan: Chromatographia 61 (2005) 429–432. 10.1365/s10337-005-0531-x
• 10)  S.A. Piletsky, H. Matuschewski, U. Schedler, A. Wilpert, E.V. Piletska, T.A. Thiele and M. Ulbricht: Macromolecules 33 (2000) 3092–3098. 10.1021/ma991087f
• 11)  T.A. Sergeyeva, S.A. Piletsky, A.A. Brovko, L.A. Slinchenko, L.M. Sergeeva and A.V. El'skaya: Anal. Chim. Acta 392 (1999) 105–111. 10.1016/S0003-2670(99)00225-1
• 12)  T.A. Sergeyeva, H. Matuschewski, S.A. Piletsky, J. Bendig, U. Schedler and M. Ulbricht: J. Chromatogr. A 907 (2001) 89–99. 10.1016/S0021-9673(00)01053-0
• 13)  S.A. Piletsky, I.Y. Dubei, D.M. Fedoryak and V.P. Kukhar: Biopolymers Cell 6 (1990) 55–58. 10.7124/bc.00028D
• 14)  D. Silvestri, C. Borrelli, P. Giusti, C. Cristallini and G. Ciardelli: Anal. Chim. Acta 542 (2005) 3–13. 10.1016/j.aca.2004.12.005
• 15)  J.M. Hong, P.E. Anderson, J. Qian and C.R. Martin: Chem. Mater. 10 (1998) 1029–1033. 10.1021/cm970608f
• 16)  C.H. Zhang, F.Y. Hu, H.B. Zhang, H. Li and S.Z. Yao: Acta Chimi. Sin. 67 (2009) 2121–2126.
• 17)  L. Donato, M.C. Greco and E. Drioli: Desal&Wat. Treat. 30 (2011) 171–177. 10.5004/dwt.2011.1960
• 18)  M.Y. Hao, X.L. Hu, P. Guan and F. Yang: Polym. Bull. 6 (2007) 8–14.
• 19)  X.J. Shi, B.J. Gao and X.L. Zhao: Acta Polymerica Sinica. 8 (2013) 1047–1056. 10.3724/sp.j.1105.2013.12369
• 20)  J.H. Gao: J. Shandong Ag. Admini. College. 1 (2011) 162–168. 10.15948/j.cnki.37-1500/s.2011.01.028
• 21)  X.J. Wang, Z.L. Xu, N.C. Bing and Z.G. Yang: J. Appl. Polym. Sci. 109 (2008) 64–73. 10.1002/app.26805
• 22)  Y.H. Zhai, Y.W. Liu, X.J. Chang, X.F. Ruan and J.L. Liu: React. Funct. Polym. 68 (2008) 284–291. 10.1016/j.reactfunctpolym.2007.08.013