Functional organic material for roxarsone and its derivatives recognition via molecular imprinting
Abstract
Roxarsone, one of feed add drugs containing arsenic, has been most widely used in poultry and swine industry. Roxarsone discharged into the environment has caused serious pollution problem. Herein, a reusable functional material for selective recognition and adsorption of roxarsone and its derivatives were designed and synthesized. The interaction mechanism is based on acid‐base interaction and surface molecular imprinting. Dual functionalized core‐shell structure with silica gel as the core was prepared to use as carrier for surface molecularly imprinted polymers. Surface molecularly imprinted polymers for roxarsone was successfully designed and synthesized using 3‐aminopropyltriethoxysilane and methyl acryloyloxypropyltriethoxy silane as functional monomers, Ethylene glycol dimethacrylate as crosslinker, Azobisisobutyronitrile as initiator, acetonitrile as solvent. Binding study showed that the recognition selectivity for roxarsone and its derivatives can be significantly improved (3.5‐4 folds) with molecular imprinting. Moreover, the prepared functional material for selective recognition and adsorption of Roxarsone was reusable for multiple times without significant decreasing their adsorption capacities.
1 INTRODUCTION
Roxarsone, an organoarsenic compound, is served as a feed additive most widely used in poultry and swine industry to promote animal growth and prevent infection.1 The United States Food and Drug Administration allowed roxarsone for chicken feed in 1964. Ministry of Agriculture of China approved use of the drug in 1996. Although the use of arsenic‐containing feed additive can promote the development of animal husbandry, a large number of livestock excreta containing organic arsenic compounds are discharged into the environment, causing serious pollution problems. In addition, roxarsone was determined to be the predominant water‐soluble as species and accounted for 70% to 90% of the total as in poultry manure.2 However, most research on roxarsone mainly concentrated on the mechanism of the animal metabolism, toxicity and the environment behavior.3 Although some adsorbents such as multiwall nanotubes,4 modified montmorillonite,5 and iron and aluminum oxides6 have been employed for roxarsone removal, most of them showed no or very limited adsorption capacities and nonselectivity for roxarsone. Therefore, it is particularly important and practically useful to design advanced functional materials with high adsorption capacity and high selectivity for roxarsone. To our best of knowledge, study on the specific recognition of roxarsone at molecular level and design functional materials for roxarsone adsorption and removal are rarely reported. Up to now, only 1 published paper reported removal of roxarsone analogs by molecular imprinted polymers.7 In their work, they chosen roxarsone and arsanilic acid as templates, traditional well known functional monomer MAA, 2‐vinyl pyridine, 4‐vinyl pyridine, and ethyleneglycoldimethacrylate (EGDMA) as crosslinker. The maximum adsorption capacity of the prepared molecular imprinted polymers for roxarsone is quite limited (0.93 mg/g).
Molecular imprinting technology is artificial recognition system, which mimics the nature's lock and key, enzyme and antibody mechanism. It holds great application potential for separation,8, 9 catalysis,10, 11 chemical or biological sensor,12-21 as well as biomedical materials.22-24 It is a process that template molecule and functional monomers form complexes by either covalent or noncovalent interactions. After polymerization and removing template molecule, highly cross‐linked polymer is obtained to produce specific binding sites complementary to the template molecule in terms of size, shape, and 3‐dimensional functionality. The significant advantages of molecularly imprinted polymers are its excellent mechanical strength and chemical stability, low preparation cost and ease of preparation (13, 18, 20; Zimmerman and others 2002). Surface molecular imprinting,25-31 one of molecular imprinting technology, is a method in which the designed functional groups are anchored on a support. After adding template and crosslinking agents, imprinting layers can be formed around the support yielding a core‐shell–type structure. If a polymerizable motif was engineered on the support material surface, it will facilitate the imprinting process by forming stable covalent bonds between functional monomer, crosslinker, and support thus improve the fidelity of the imprinting process. Furthermore, because of the relatively poor solubility of roxarsone in most organic solvents, the traditional method of preparing molecular‐imprinted polymer with nonpolar monomer and crosslinker system was not found to be satisfactory. Therefore, in subsequent studies, we turned to surface imprinting to address the issue.
The chemical structural of roxarsone is shown in Figure 1, a benzene ring containing a nitro and a hydroxyl group is connected with arsenic atoms. The strong electron withdrawing nitro group enhances acidity of the ortho‐phenol hydroxyl group. roxarsone itself is also an arsenic acid derivative, which means it is a moderately strong acid. As shown in Figure 1, the pKa values of roxarsone are 3.49, 5.74, and 9.13, respectively.32 The acidity of roxarsone implies that the acid‐base interaction is potential and promising to design functional materials for roxarsone at molecular level. The advantages of acid‐base interaction include the following: (1) easily accessible and (2) acid‐base interaction can be inverted by adding appropriate competing reagents, which means target molecule can be adsorbed and desorbed multiple cycles. It would be particularly useful with respect to the reuse of the designed functional materials.
In the work reported here, the amino group is chosen as basic functional group to achieve high adsorption capacity for roxarsone. Silica gel is chosen as supporting material to chemically anchor 3‐aminopropyltriethoxysilane (APS) to afford an amino‐rich functionalized surface. Furthermore, silica gel has a larger surface area, which can provide enough space for the reaction of siloxane reagent to afford quality‐functionalized surface with high density of amino groups. The advantages of our approach are affordable and synthetically viable. Binding studies showed that functional organic material prepared by surface molecular inprinting present high binding capacity and selectivity for roxarsone.
2 EXPERIMENTS
2.1 Materials
Methanol, acetonitrile, and toluene were all chromatography grade and were purchased from Sigma‐Aldrich and were used without further purification unless otherwise indicated; Silica gel and 3‐aminopropyltriethoxysilane (APS) were purchased from Alfa Aesar; azobisisobutyronitrile (AIBN) and roxarsone (ROX) were purchased from Adamas; ethyleneglycoldimethacrylate (EGDMA), 3‐(triethoxysilyl)propyl methacrylate (MPS), 4‐hydroxy benzene arsenate (OSA), and 4‐hydroxy‐phenylarsonic (ASA) were purchased from TCI; All reactions were carried out under a dry nitrogen atmosphere.
2.2 Instrumentation
Scanning electron microscopy images (SEM) were recorded on Zeiss Supra 55 (VP) (Carl Zeiss, Germany) microscope. Element analysis was carried out by Vario El cube element analyzer (Elementar, Germany). The FT‐IR spectra were obtained with a BIO‐RAD FTS165 Fourier transform infrared spectrometer with a 4 cm‐1 resolution in 4000 cm−1 to 400 cm−1 region by using KBr. The UV‐Vis spectra were recorded on a Shimadzu UV‐2600 spectrometer (Shimadzu, Japan). TGA studies were carried out under air atmosphere at heating rate of 10°C/min using a STA449F3 thermal analyzer (Netzsch, Germany).
2.3 Synthesis of APS‐SiO2 (1:2, w /w ratio)
Representative example to preparation APS‐SiO2 (1:2, w /w ratio) (See Scheme 1) is given as follows:The 3.0 g silica gel was dispersed in 20 mL of toluene under magnetic stirring and was added 0.15 mL deionized water, stirring for half an hour. Then 1.58 mL 3‐aminopropyl triethoxy silane (APS) was added drop wise and stayed for 16 hours at room temperature. Solvent was removed by distillation under vacuum. The product was dried under vacuum at 80°C for 3 hours to afford APS‐SiO2 as white powder. APS‐SiO2 was dispersed in 100 mL methanol and stirred. After it was settled for 5 minutes, the upper liquid layer was decanted and the solid part was filtered and washed with 20 mL of methanol 5 times. Then APS‐SiO2 was dried under vacuum at 60°C.
APS‐SiO2 (1:4, w /w ratio) was prepared in similar manner only by changing the amount of APS (0.79 mL) and with all other conditions remained the same.
2.4 Synthesis of MPS‐APS‐SiO2 (1:1:4, w/w/w Ratio)
Three grams silica gel was dispersed in 20 mL of toluene under stirring and was added 0.15 mL deionized water, stirring for 30 minutes, then 0.76 mL of 3‐(triethoxysilyl)propyl methacrylate (MPS) and 0.79 mL of 3‐aminopropyl triethoxy silane (APS) were added drop wise and stay for 16 hours at room temperature. After reaction, solvent was removed by vacuum distillation. The prepared MPS‐APS‐SiO2 (1:1:4, w /w ratio) was washed with methanol thoroughly and finally dried under vacuum at 80°C for 3 hours.
2.5 Synthesis of molecularly imprinted polymers (MIPs) and nonmolecularly imprinted polymers (NIPs)
Synthesis of molecularly imprinted polymers (MIP) (ROX) and nonmolecularly imprinted polymer (NIP): 0.27 g roxarsone (ROX) was dissolved in 100 mL methanol and was added 1.0 g MPS‐APS‐SiO2 (1:1:4, w /w/w ratio) under stirring for 60 minutes. The solid was filtered, washed with 60 mL acetonitrile 3 times, and dried under vacuum to get roxarsone loaded MPS‐APS‐SiO2 as yellow powder. The obtained powder was dispersed in 15 mL acetonitrile, and then was added 0.50 g EGDMA, 10 mg AIBN; the system was flushed with nitrogen stream for 10 minutes. Polymerization was done at 60°C under N2 for 24 hours. The prepared polymer was filtered. Template molecular was removed by Soxhlet extraction (methanol: triethylamine = 4:1, V:V) for 24 hours and followed by thorough washing with methanol 10 times (20 mL each time). Finally, the solid was dried under vacuum at 75°C for 12 hours to afford MIP (ROX).
All MIPs and NIPs were prepared in same manner except that NIPs were made without adding template molecule. Structure of template molecules is shown in Figure S2. The abbreviations are as follows: 3‐nitro‐4‐hydroxybenzoic acid (roxarsone, ROX); 4‐hydroxybenzoic acid (oxarsanilic acid, OSA) and 4‐aminobenzoic acid (arsanilic acid, ASA). MIP(ROX), MIP(OSA), MIP(ASA) was used for imprinted polymers accordingly.
3 RESULTS AND DISCUSSION
3.1 Preparation methods and characterizations
The typical synthesis of MIPs is depicted in Scheme 2. First, silica gel was chosen as support material because of its advantage of large surface area. Another reason is that silica gel surface contains a large number of hydroxyl groups, which can be easily modified to form dual functionized MPS‐APS‐SiO2 with APS and MPS. Then dual functionized surface was prepared by silane condensation chemistry. The modified surface feature not only the recognition amino group target roxarsone as well as a terminal double bond, which can be covalently linked to crosslinking agents during the molecular imprinting process. The acid‐base interaction between the amino and roxarsone acidic functional group tightly hold the template molecule on the surface of modified support. After that, cross‐linker and initiator were added to make crosslinked polymer layer. Finally, the template molecule was removed to obtain functional organic materials specific for roxarsone recognition.
In surface modification process with silanes, water is found to be critical. When encountering water, APS hydrolyzes into silicon alcohol quickly and condenses with hydroxyl groups on silica gel surface by dehydration. Water in the reaction actually plays an indispensible catalytic role. Too much water will induce severe mutual condensation of hydrolyzed silicon to form oligemers. In addition, the amino group of APS also plays a catalytic role by forming hydrogen bonding with silicon hydroxyl. Various techniques were used to evaluate the surface modification process.33, 34
3.1.1 Elemental analysis
Elemental analysis was conducted to measure composition of SiO2 and APS‐SiO2 (1:2, w /w ratio) to evaluate the efficacy of the modification process.
Results of elemental analysis were summarized in Table 1.The efficiency of APS grafted on SiO2 surface can be estimated by comparing the N content before and after APS modification. Theoretical nitrogen content was calculated as 4.35% in APS‐SiO2 assuming APS were all grafted on SiO2 surface. The measured nitrogen content was 4.02% in APS‐SiO2. Result implied that under the reaction conditions, the yield reached up to 92.4%. The possible side reactions could be APS condenses itself to form silane oligmers.
| Materials | Element content | ||
|---|---|---|---|
| N% | H% | C% | |
| A:SiO2 | 0.08 | 0.64 | 0.09 |
| B:APS‐SiO2 | 4.01 | 2.59 | 10.78 |
3.1.2 Infrared spectrum analysis
Infrared spectra of silica gel and APS‐SiO2 (1:4, w /w ratio) are shown in Figure 2 (A and B). Compared to IR spectra of silica gel, the stretching vibration peak of the silicon hydroxyl at 3450 cm‐1 was significantly decreased. Meanwhile, the bending vibration for Si‐O‐H at 1650 cm‐1 disappeared, indicating condensation occurred between hydroxyl groups and APS. Infrared spectra of MPS‐APS‐SiO2 (1:1:4, w /w/w ratio) and MIP(ROX) are shown in Figure 2 (C and D). Compared to IR spectra of silica gel in Figure 2A, new absorption peaks at 2960 cm‐1 and 1721 cm‐1 in spectrum of MPS‐APS‐SiO2 (1:1:4, w /w/w ratio) were attributed to the stretching vibration absorption of saturated C─H and carbonyl C═O in ester groups, respectively. These results suggested that MPS was successfully grafted onto the surface of SiO2. Compared with MPS‐APS‐SiO2 (1:1:4, w /w/w ratio), the spectra of the imprinted materials MIP(ROX) showed stretching vibration absorption of saturated C─H and carbonyl C═O of ester groups were strengthened, which was the result of polymerization process. The infrared data implied that the surface‐imprinted polymer MIP(ROX) was successfully prepared.
3.1.3 Scanning electron microscope
The morphology of materials was characterized by SEM. Figure 3 shows the SEM images of silica gel and APS‐SiO2 (1:4, w /w ratio). A distinct surface morphology change was observed which was consistent with the formation of surface organic layer. The SEM images of MPS‐APS‐SiO2 (1:1:4, w /w/w ratio) and MIP(ROX) were shown in Figure 3A. Results illustrated that the functionalized vinyl group indeed reacted with the monomers and cross‐linker during the polymerization.
3.1.4 Thermogravimetric analysis
Results of thermoravimetric analysis were showed in Figure 4. Weight loss of both MIP(ROX) and NIP were higher than that of silica gel and MPS‐APS‐SiO2 (1:1:4, w /w/w ratio). This is probably due to the loss of organic layer at elevated temperature. The results implied that the layers of imprinted polymer had been formed successfully during the polymerization process.
3.2 Binding study of roxarsone in solution by silica gel and APS‐SiO2 (1:4, w /w ratio)
Quantitative analysis of roxarsone recognition and adsorption in solution were carried out by high performance liquid chromatography and ultraviolet spectroscopy (UV).10, 11 High performance liquid chromatography works well for a complex system because of its capability of separating different components and then detect them. On the other hand, the ultraviolet adsorption characteristic aromatic benzene ring of roxarsone is a useful handle. Therefore, UV turns to be fast, easy, convenient, and inexpensive analytical method for roxarsone concentration measurement in solution.
Figure S1A is UV adsorption spectrum of 0.2 mM/L ROX in methanol. It shows characteristic adsorption peak at 226, 268, and 328 nm. The maximum adsorption peak at 226 nm was chosen for the quantitative analysis of roxarsone solution. There is a linear correlation between absorbance and concentration within 0.03 and 0.3 mM/L range. The standard curve was obtained and used to measure the concentration of roxarsone (Figure S1B).
To prepare functional material with high binding capacity for roxarsone, acid‐base interaction between amino‐carboxylic acid and amino‐phenol hydroxyl is chosen as critical design. Evaluation of binding property of APS‐SiO2 (1:4, w /w ratio) for roxarsone was conducted by taking a series of 10 mg APS‐SiO2 (1:4, w /w ratio) in 15 mL of 0.2 mM/L of roxarsone in methanol. After the equilibrium, the absorbance value of supernatant (filtered by 0.22 μm teflon filter) at 226 nm was recorded before and after adsorption through. Results (Figure 5) showed that it took 60 minutes to reach binding equilibrium between APS‐SiO2 (1:4, w /w ratio) and roxarsone. The binding capacity of APS‐SiO2 (1:4, w /w ratio) was measured as 68.79 mg/g, while for the control silica gel, it was only 2.51 mg/g.
The adsorption capacity of an adsorbent can be described by the equilibrium adsorption isotherm, which is characterized by sets of constants whose values express the surface properties and affinity of the adsorbent. The adsorption isotherm is important from both theoretical and practical points of view. The parameters obtained from the different models provide important information about the adsorption mechanism, the surface properties and the affinities of adsorbents. Among the isotherm equations available for analyzing experimental adsorption equilibrium data, the most widely used adsorption models for single‐solute systems are the Freundlich and Langmuir models.35, 36
The Freundlich equation is described as follows:
(1)The Langmuir equation is described as follows:
(2)Langmuir and Freundlich models were used to carry out nonlinear fitting of adsorption isotherms of APS‐SiO2 (1:4, w /w ratio). As showed in Figure 6, the R2 values of Langmuir model and Freundlich model are 0.9699 and 0.9966, respectively, (Table 2). So the adsorption isotherm of APS‐SiO2 (1:4, w /w ratio) is more consistent with the Freundlich model.
| Langmuir model | Freundlich model | |||||
|---|---|---|---|---|---|---|
| Materials | q m (mg/g) | K a (L/mg) | R 2 | n | K F (mg/g) | R 2 |
| APS‐SiO2 | 157.029 | 21.956 | 0.9699 | 2.9224 | 228.23 | 0.9966 |
3.3 Binding study of roxarsone in solution by MIP(ROX)
Ten milligrams MIP(ROX) and NIP was added 15 mL of 0.2 mM/L of ROX to methanol respectively. After the equilibrium, the absorbance value of supernatant (filtered by 0.22 μm Teflon filter) at 226 nm was recorded through which binding capacity and binding percentage were calculated. As shown in Figure 7, the adsorption amounts of ROX on MIP(ROX) increased quickly over time during the first 30 minutes and then began to slow down. After 60 minutes, the increase of adsorption percentage stopped. The measured binding capacity of MIP(ROX) was 49.36 mg/g, while for the control NIP, it was 38.29 mg/g. The calculated imprinting factor for ROX is 1.29. It is reasonable to assume that a large number of imprinted cavities exist on the surface of the imprinting material, so the template ROX is able to enter into the cavities and bind with the recognition sites. After the recognition sites on the surface of MIP(ROX) were filled up with the template ROX, the diffusion resistance would make it more difficult to access to the imprinted cavities at inner position, thus the rate of adsorption will drop gradually and eventually stop. The fast adsorption process indicated that the diffusion resistance for ROX is relatively smaller, and the imprinted sites in MIP(ROX) is easily accessible for ROX.
The binding ability of MIP(ROX) and NIP for ROX was studied with static solution binding experiments. Figure 8 showed the adsorption isotherms of ROX onto the MIP(ROX) and NIP surface. It was found that the binding amounts of ROX on MIP(ROX) increased dramatically with the concentration. The difference in adsorption amounts between MIP(ROX) and NIP was getting larger with the increasing concentration of ROX. The MIP(ROX) exhibited higher adsorption capacity for ROX than that of NIP at both low and high concentration. The results clearly showed that although MIP and NIP had almost the same chemical composition, MIP(ROX) adsorbed much more template than that of NIP could be attributed to the specific recognition sites generated in imprinting process.
Langmuir and Freundlich models were used to do nonlinear fitting of adsorption isotherms of MIP(ROX) and NIP at low concentration. As displayed in Table 3, the R2 values obtained by fitting to Langmuir model and Freundlich model are 0.9964 and 0.9803 for MIP(ROX) and 0.9911 and 0.9741 for NIP. The adsorption isotherms of MIP(ROX) and NIP at low concentration are more consistent with the Langmuir model. Considering MIP(ROX) has higher percentage of imprinted cavities well matched with template, the formation of these cavities enhance the affinity of MIP toward ROX molecules. For NIP, adsorption amount of ROX at equilibrium is much lower than that of MIP(ROX). It is reasonable that in the process of surface polymerization, some amino groups are embedded into the polymer shell and are no longer accessible for ROX. However, there is still abundant amount of amino groups that are accessible. It is understandable that nonspecific adsorption would occur both in surface molecular imprinting and traditional molecular imprinting because of the binding sites heterogeneity.
| Langmuir model | Freundlich model | |||||
|---|---|---|---|---|---|---|
| Materials | q m (mg/g) | K a (L/mg) | R 2 | n | K F (mg/g) | R 2 |
| MIP(ROX) | 100.04 | 12.685 | 0.9964 | 1.8484 | 191.78 | 0.9803 |
| NIP | 73.971 | 10.109 | 0.9911 | 1.9142 | 120.35 | 0.9741 |
3.3.1 Adsorption selectivity of APS‐SiO2 (1:4, w /w ratio) and MIP(ROX) for ROX
Two structural similar compounds: oxarsanilic acid(OSA) and arsanilic acid (ASA) (Figure S2) were chosen to investigate the binding selectivity. APS‐SiO2(1:4, w /w ratio) demonstrated decent binding capacity for ROX and its analogs (Figure 9), which implied effectiveness of the designed acid‐base working mechanism.
Although the designed APS‐SiO2 can achieve high binding capacity for ROX and its analogs, molecular imprinting strategy can be adopted to improve recognition selectivity. Along this line, MIPs using roxarsone analogs as templates were prepared along their corresponding NIPs.
Figure 10 shows the binding capacity and binding percentage for ROX, OSA, and ASA using MIP(ROX). After imprinting, although binding capacity decreased to some extent, recognition selectivity improved dramatically. The modified silica gel surfaces are wrapped around by crosslinked organic imprinting layer after polymerization. Highly specific and selective recognition sites are produced via imprinting process, however some recognition groups could be buried and be located in a closed region and become so called dead sites. This is probably the major reason for binding capacity decrease after imprinting.
Impressively, the recognition selectivity was significantly improved with molecular imprinting (Table 4). Roxarsone recognition selectivity of MIP(ROX) made with ROX was 3.5 to 4.0 folds higher than that of APS‐SiO2 (1:4, w /w ratio). Considering NIPs were made in exact same manner only without adding template molecule, these results support the mechanism that not only the designed acid‐base interaction is the key for roxarsone and its analogs recognition but also it is the molecular imprinting process that endows the improved selectivity and specificity.
| Binding Percentage(%) | Ratio | ||||
|---|---|---|---|---|---|
| Materials | ROX | OSA | ASA | ROX/OSA | ROX/ASA |
| APS‐SiO2 | 84.36 | 24.6 | 5.57 | 3.43 | 15.15 |
| MIP(ROX) | 60.34 | 4.35 | 1.13 | 13.87 | 53.40 |
| MIP(ROX)(selectivity improvement) | 4.04 | 3.52 | |||
The calculated imprinting factor for OSA and ASA are 1.83 and 1.71, respectively, (Figure 11). Results clearly demonstrated that through surface molecular imprinting strategy described here, the recognition selectivity could be significantly improved.
3.4 Reusability of MIP(ROX) for roxarsone recognition and adsorption
Results showed that designed acid‐base interaction holds MIP(ROX) and roxarsone tightly together. In addition, the interaction between amino group of MIP(ROX) and acidic group of ROX can be disrupted if a stronger base than primary amino group or a stronger acid than roxarsone was used as competing agents. For example, hydrochloric acid can replace roxarsone and protonate amino group. If a base wash was followed, amino groups can be restored for another adsorption cycle.
A more mild condition is to use a relatively stronger base as competing agents. The ROX absorbed on APS‐SiO2 can be desorbed by solvent Soxhlet extraction for 24 hours with stronger base such as triethylamine in methanol. Then the regenerated APS‐SiO2 can be used for roxarsone adsorption for multiple times.
The reusability study of MIP(ROX) for roxarsone recognition and adsorption were carried out, and results were summarized in Figure 12. In this study, 10 mg MIP(ROX) was added 15 mL of 0.2 mM/L of roxarsone in methanol. After the equilibrium, the absorbance value of supernatant (filtered by 0.22 μm Teflon filter) at 226 nm was recorded before and after adsorption through which binding capacity and binding percentage were calculated. The ROX absorbed on MIP(ROX) are desorbed by solvent Soxhlet extraction for 24 hours with triethylamine in methanol. Then the regenerated MIP(ROX) are used for roxarsone adsorption for multiple cycles. Results showed that after 7 repeated absorption‐desorption cycles, the binding capacity of MIP(ROX) was kept at about the same level without observable decrease.
4 CONCLUSIONS
A reusable functional organic material for selective recognition and adsorption of roxarsone and its derivatives were designed and synthesized. Molecular‐imprinted polymers with high adsorption capacity were designed and synthesized. The dual functionalized carrier was wrapped by organic imprinting layers. MIPs of roxarsone and its derivatives were successfully prepared using 3‐aminopropyltriethoxysilane and methyl acryloyloxypropyltriethoxy silane as functional monomers, EGDMA as crosslinker, AIBN as initiator, acetonitrile as solvent. Quantitive analysis of analytes was established using UV spectrophotometer. Binding study showed that synthesized functional materials are reusable. The recognition selectivity for roxarsone and its derivatives was significantly improved with molecular imprinting.
ACKNOWLEDGEMENTS
This work was financially supported by Natural Science Foundation of Xinjiang, China (2015211A046).
