Doubly imprinted polymer nanofilm-modified electrochemical sensor for ultra-trace simultaneous analysis of glyphosate and glufosinate

Biosensors and Bioelectronics 59 (2014) 81–88

Contents lists available at ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Doubly imprinted polymer nanofilm-modified electrochemical sensor for ultra-trace simultaneous analysis of glyphosate and glufosinate Bhim Bali Prasad n, Darshika Jauhari, Mahavir Prasad Tiwari Analytical Division, Department of Chemistry, Faculty of Science, Banaras Hindu University, Varanasi 221005, India

art ic l e i nf o

a b s t r a c t

Article history: Received 19 December 2013 Received in revised form 20 February 2014 Accepted 3 March 2014 Available online 20 March 2014

A rapid, selective, and sensitive double-template imprinted polymer nanofilm-modified pencil graphite electrode was fabricated for the simultaneous analysis of phosphorus-containing amino acid-type herbicides (glyphosate and glufosinate) in soil and human serum samples. Since both herbicides respond overlapped oxidation peaks and only glyphosate is prone to nitrosation, n-nitroso glyphosate and glufosinate were used as templates for obtaining the well-resolved quantitative differential pulse anodic stripping voltammetric peaks on the proposed sensor. Toward sensor fabrication, a nano-structured polymer film was first grown directly on the electrode via initial immobilization of gold nanoparticles at its surface. This was followed by linking of monomeric (N-methacryloyl-L-cysteine) molecules through S–Au bonds. Subsequently, these molecules were subjected to free radical polymerization, in the presence of templates, cross linker, initiator, and multiwalled carbon nanotubes as pre-polymer mixture. The modified sensor observed wide linear ranges (3.98–176.23 ng mL
1 and 0.54–3.96 ng mL
1) of simultaneous analysis with detection limits as low as 0.35 and 0.19 ng mL
1 (S/N¼ 3) for glyphosate and glufosinate, respectively, in aqueous samples. The respective oxidation peak potentials of both analytes were found to be substantially apart by 265 mV. This enabled the simultaneous determination of one target in the presence of other, without any cross reactivity, interferences, and false-positives, in real samples. & 2014 Elsevier B.V. All rights reserved.

Keywords: Double template imprinted polymer film Glyphosate Glufosinate Gold nanoparticle Differential pulse anodic stripping voltammetry Simultaneous ultra-trace analysis

1. Introduction Occurrence and accumulation of herbicides, and their metabolites, in soil have deleterious effect on the environment. Particularly, organophosphates and organophosphonates constitute one family of the most commonly applied pesticides in agriculture. For example, glyphosate [N-(phosphonomethyl) glycine, GLY] and glufosinate [(DL-homoalanine-4-yl)-methylphosphinic acid, GLU] are known as non-selective and post-emergence contact herbicides. These herbicides possess amino acid-like structures (Stalikas and Konidari, 2001) and interfere with the formation of amino acids and other chemicals in plants (Cikalo et al., 1996; Kataoka et al., 1996). GLY and GLU have similar structures, yet they are different in their mode-of-action (Hoerlein, 1994). Because of moderate toxicity to animals and humans, these have extensively been used as herbicides worldwide. However, if ingested over a period of time, GLY and GLU may affect the central nervous system, resulting in respiratory, myocardial, and neuromuscular malfunctions, which can even lead to death (Fujii et al., 1996; Richard et al., 2005; Walsh et al., 2000). Therefore, many authorities have suggested the maximum residue

n

Corresponding author. Tel.: þ 91 9451954449; fax: þ 91 5422268127. E-mail address: [email protected] (B.B. Prasad).

http://dx.doi.org/10.1016/j.bios.2014.03.019 0956-5663/& 2014 Elsevier B.V. All rights reserved.

levels (MRLs) of these compounds in water and agricultural products. The MRLs of GLY and GLU in most crops listed by the United Nations Food and Agricultural Organization are 0.1–5.0 mg/kg and 0.05 mg/kg (FAO, 2013 (http://www.codexalimentarius.net/download/report/655/ al29_24e.pdf (2006))), respectively. Currently, GLY is in the list of the U.S. national primary drinking water contaminants with a maximum contaminant level of 0.7 mg L
1. The European Union limit of any pesticide in drinking water has been set to 0.1 μg L
1, irrespective of their toxicological effects (Stalikas and Pilidis, 2000). In view of the fact that real samples are sufficiently diluted to mitigate the matrix effect, there is an urgent need for the simultaneous monitoring of GLY and GLU in the dilute environmental and biological samples, at trace level. Most of the methods reported for GLY and GLU simultaneous determination necessarily involved sample derivatization. These include gas chromatography (Hori et al., 2003; Motojyuku et al., 2008; Tseng et al., 2004) liquid chromatography (Hanke et al., 2008; Hao et al., 2011), high performance liquid chromatography (Khrolenko and Wieczorek (2005); Moye et al., 1983), and capillary electrophoresis (Chang and Wei, 2005; Corbera et al., 2005; Goodwin et al., 2003; Jiang and Lucy, 2007; See et al., 2010) techniques. However, in order to avoid costly instrumentations, electrochemical sensors have alternatively been used for on-site cost-effective and rapid analysis of these pesticides (Khenifia et al.,

82

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

2009; Simoes et al., 2006; Songa et al., 2009a, 2009b; Wei et al., 2013). Despite the fact that the simultaneous analysis of GLY and GLU requires a chemical alteration in one of the pesticides and it is inevitable prior to the analysis stage via any method, electrochemical sensors reported hitherto still lacked selectivity in complex matrices. Therefore, the development of simple, selective, and sensitive electrochemical sensors that can afford analysis of real samples without any cross-reactivity and false positives is warranted. Nowadays, imprinting two or more targets (print molecules) in a single molecularly imprinted polymer (MIP) format is an upcoming technology (Dai et al., 2013; Guo and Guo, 2013; Jing et al., 2010; Matsui et al., 2006; Prasad et al., 2013; Sreenivasan, 2001; Suedee et al., 2008; Tiwari et al., 2011; Xin et al., 2013). This saves both time and labor as compared to traditional imprinting. Molecularly imprinted electrochemical sensors apparently combine the characteristics of electrochemical detection and molecular imprinting technology. This might offer high selectivity, simple operation, and low cost (Blanco-López et al., 2004; Li et al., 2011). Recently, nanomaterials have been used in a wide range of applications as a material foundation of chemosensors. Such sensors have exhibited various degrees of success in the improvement of detection sensitivity and selectivity (Guan et al., 2008). Especially, gold nanoparticles (GNPs) have potential applications in the construction of highly selective electrochemical sensors due to their advantages of enhanced diffusion, good stability and biocompatibility. Furthermore, GNPs might serve as “electron antennae” for channeling electron transport between the electrode and the electro-active species (Atta et al., 2012). As a fascinating branch, polymers, artificially decorated onto GNPs in physical or chemical manners, show much potential with higher value in advanced material science (Li et al., 2009). Most surprisingly, only MIP for GLY is reported (Jenkins et al., 2001; Zhao et al., 2011), whereas GLU selective MIP is yet to be developed. Furthermore, the double imprinting of these herbicides in a single polymer format is, hitherto, not attempted. The prime concern, while developing a double-templated MIP for the development of an electrochemical sensor, is the occurrence of an almost overlapped oxidation current involving identical electrontransfer route of both templates. As a remedial, we have resorted to exploit derivatization selectively that yielded N-nitroso derivative of GLY (NGLY); whereas GLU fortuitously remained underivatized in the herbicide mixture. The reason being GLU is a primary amine that does not favor nitrosation. On the other hand, GLY is a secondary amine that could be derivatized readily as nitrosoalkylamines—a more stable product (Zhao et al., 2007). In this work, the pencil graphite electrode (PGE) surface was first decorated with GNPs followed by the self assembly of cysteine containing monomer molecules. Subsequently, the modified surface was subjected to activator generated by electron transfer for atom-transfer radical polymerization (AGET-ATRP) of the prepolymerization mixture (containing both NGLY and GLU as templates, ethylene glycol dimethacrylate (EGDMA) as cross-linker and chloroform as initiator) in the presence of heterogeneously dispersed multi-walled carbon nanotubes (MWCNTs). Herein, the concerted effect of GNPs and CNTs dispersants in MIP layer might produce nanohybrids to exhibit an excellent electron-transfer capability for the oxidation of herbicides, with improved electrocatalytic activity (Rajabzade et al., 2012).

2. Experimental 2.1. Chemicals and reagents All chemicals were of analytical reagent grade, and used without further purification. Demineralized triple distilled water (conducting

range 0.06–0.07  10
6 S cm
1) was used throughout the experiment. Methacryloyl chloride (MC), chloroauric acid (HAuCl4  H2O)), 2-20 azobis(isobutyronitrile) (AIBN), and L-cysteine hydrochloride monohydrate (Cys), were purchased from Loba Chemie (Mumbai, India) and Spectrochem Pvt. Ltd. (Mumbai, India). Sodium nitrite and ammonium sulfamate, were purchased from S.D. Fine Chem. Pvt. Ltd. (Boisar, India). Solvents dimethylsulfoxide (DMSO), dimethylformamide (DMF), acetic acid, acetonitrile (ACN), and triethylamine (TEA), were purchased from Spectrochem Pvt. Ltd. (Mumbai, India). Cupric chloride (CuCl2) and 2,20 -bipyridyl (bpy) were purchased from BDH chemicals (England). EGDMA, GLY, GLU, MWCNTs, and all interferents of GLY and GLU, were provided by Aldrich (Steinheimer, Germany) and Fluka (Steinheimer, Germany). Standard acetate buffers were made using acetic acid and NaOH and their pH values were adjusted with the addition of a few drops of either 0.1 M HCl or 0.1 M NaOH. Stock solutions (500 μg mL
1) of GLY and GLU were prepared in water. All working standards were prepared by diluting stock solutions with water. Human blood serum sample was obtained from the Institute of Medical sciences, Banaras Hindu University (Varanasi, India) and stored in a refrigerator at  4 1C, before use. Any pretreatment (deproteinization, ultrafiltration, ultrcentrifugation, etc.) of blood serum sample was avoided since this might lead to inaccuracies in the final results. Nevertheless, 50-fold dilution of blood serum was necessary in order to mitigate the matrix effect. Soil sample was collected from a local agricultural land and suspended in water (1.0 g/30 mL), followed with the removal of solid residues, if any, by centrifugation and filtration. Fortuitously, the soil sample solution did not cast any matrix effect on analysis. Therefore, its dilution was not necessary. Pencil rods (2B), 0.5 mm in diameter and 5 cm in length, were purchased from Hi Par, Camlin Ltd. (Mumbai, India). 2.2. Apparatus All voltammetric measurements were carried out with a polarographic analyzer/stripping voltammeter [model 264A, EG & G Princeton Applied Research (PAR), USA] in conjunction with an electrode assembly [PAR model 303A] and a X–Y chart recorder (PAR model, RE 0089). Chronocoloumetry was performed with an electrochemical analyzer [CH instruments, USA, model 1200A]. All experiments were carried out using a three electrode cell assembly consisting of modified PGE, platinum wire, and Ag/AgCl (3.0 M KCl) as working, counter, and reference electrodes, respectively. For FT-IR (KBr) spectral analysis, a minimum of about 5.0 mg samples were scrapped out from the modified electrode surface. Subsequently, this was mixed with KBr pellets in a dye to form a disc and then subjected to spectral recording using Varian 3100 FTIR (USA). Morphological study of the nanoparticles was made using Tunneling Electron Microscopy (TEM) (Technai-12 FEI, Eindhoven, Netherland). Morphological images of bare and modified PGE surfaces were recorded using scanning electron microscope (SEM), JEOL, JSM model-840A (Netherland). Atomic force microscopy (AFM) using a NT-MDT Microscope (NT-MDT Co., Russia) was performed in the semi contact mode. All experiments were carried out at 2571 1C. The coating of film over the GNPs-PGE surface was made using an indigenous spin-coater SCU-2008C (Apex Instruments Co., India). 2.3. Synthesis of monomer [N-methacryloyl-L-cysteine (MAC)] The monomer, MAC, was prepared and characterized as described elsewhere (Utku et al., 2008). For this, Cys (5.0 g, 0.028 mol) and sodium nitrite (0.2 g, 0.028 mol) were dissolved in 30.0 mL potassium carbonate solution (5% v/v) and ice cooled to 0 1C. To this ice cooled solution, MC (4.0 mL, 0.04 mol) was added drop-wise with vigorous stirring for 2 h. Afterward, pH of the solution was adjusted to 7.0 and then extracted with ethyl acetate.

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

83

The aqueous phase was evaporated in a rotary evaporator at 100 1C. The crude residue was crystallized in a mixed solvent, ethanol and ethyl acetate, to obtain MAC. FT-IR (KBr, cm
1): 1731 (–CQO), 1634 (amide I), 1553 (amide II), and 2650 (S–H bending). 2.4. Synthesis of GNPs GNPs were prepared as cited in the literature (Xiao et al., 1999). In short, 2.5 mL of 1% tri-sodium citrate was added to 100.0 mL of boiling 0.01% HAuCl4 solution. The size of the nanoparticles was ascertained as 20.07 1.9 nm by TEM analysis. The prepared GNPs were stored in dark at 4 1C. 2.5. Derivatization of GLY NGLY solution (500 μg mL
1) was prepared by mixing 50.0 mL GLY solution (1.0 mg mL
1), 20.0 mL HCl (5.0 mol L
1), and 5.0 mL sodium nitrite (10.0 mg mL
1) into a 100 mL volumetric flask. After 15 min, 5.0 mL ammonium sulphamate (0.10 g mL
1) was added to consume excess nitrite, and the volume was made up with water (Teófilo et al., 2004). Same procedure was used for the derivatization of GLY in soil and human blood serum samples. 2.6. Sensor fabrication A pencil rod (2B, 0.5 mm) was first pretreated by dipping in 6 M HNO3 for 15 min, washing with water, and subsequently smoothening the surface by soft cotton. This was inserted into a Teflon tube, where the tip of the pencil rod at one end was gently rubbed with an emery paper (No. 400) to level the pencil surface along the tube orifice. This yielded a flat tip for the smooth coating of monomer and pre-polymerization mixture with the help of a spincoater. Electrical contact was obtained by soldering a metallic wire to the exposed reverse side of the pencil rod. The pencil lead (2B) was preferred to harder ones (2H, H and HB) because it evolves lowest charging current (Prasad et al., 2010). Furthermore, the use of PGE as compared to other noble metal (Au, Ag, Pt, Pd, etc.) electrodes is advantageous because of its cost, wider potential window, and less background current (Gao and Song, 2005). Double imprinted polymer modified GNPs-PGE, was developed following three successive steps (Scheme 1): 1. Attachment of GNPs to PGE surface. PGE was modified by dip coating for an optimized time of 12 h in the colloidal nanoparticles solution. The resulting modified electrode (GNPs-PGE) was dried under the flow of nitrogen, and used for further modification. 2. Assembling MAC–gold nanoparticles complexes. For attachment of monomeric precursor to GNPs-PGE surface, covalent couplings of MAC were carried out directly over the nanoparticle surface via sulfur–gold bond. For this, MAC (0.2 mmol, 2 mL DMSO) was spin coated (at 1000 rpm) on the GNPs-PGE and kept standing for an hour. 3. Surface imprinting. Schematic protocol for the fabrication of double-template imprinted polymer modified GNPs-PGE (hereafter referred as MIP modified GNPs-PGE) employing AGETATRP technique, is shown in Scheme 1. Accordingly, bpy (0.02 mmol) and CuCl2 (0.02 mmol) were dissolved in 2 mL DMSO to get a solution of Cu (II)-complex. Subsequently, this complex was mixed with templates (NGLY, 0.1 mmol in 1 mL DMSO and GLU 0.1 mmol in 1 mL DMSO), EGDMA (3.0 mmol, 570 mL), and MWCNTs (40 mg), in the presence of a reducing agent TEA (0.2 mmol, 280 mL). A contrast color change from light blue to brownish yellow indicated an in-situ reduction of Cu(II)complex to Cu(I) complex, which catalyzed the chain propagation, in the presence of an initiator (chloroform,

Scheme 1. Schematic representation of double-template imprinted polymer-modified GNPs-PGE fabrication along with suggested binding mechanism for simultaneous analysis of NGLY and GLU in their respective MIP cavities.

0.2 mmol, 16 mL). The whole mixture in a glass tube was purged with N2 gas for 10 min. This mixture was spin coated at 2500 rpm onto the surface of MAC modified GNPs-PGE and then incubated in an oven for 3 h at 70 1C. The similar procedure was adopted to prepare the non-imprinted polymer-grafted GNPs-PGE (NIP-modified GNPs-PGE), in the absence of templates. Finally, both template molecules were retrieved from the MIP-adduct film by treating with a mixture of ACN and TEA (4:1, v/v) for 30 min, under dynamic condition. This was confirmed by the absence of any voltammetric current due to the template(s). For comparative study, GLY (underivatized) and GLU were also imprinted individually in the similar manner as stated above and corresponding single imprinted MIPmodified GNPs-PGE sensors were stored at room temperature.

2.7. Electro-analytical measurement Voltammetric measurements on MIP modified GNPs-PGE was performed in a cell containing 10 mL acetate buffer (pH 5.5). After blank run, both test solutions (NGLY and GLU) were added (a) one by one, (b) simultaneously, and (c) one in excess of another, into the cell. After analyte(s) accumulation for 180 s (accumulation time, tacc) at
1.3 V (accumulation potential, Eacc vs. Ag/AgCl) and 15 s equilibration time, differential pulse anodic stripping voltammetry (DPASV) runs were recorded in the potential range varying from
1.3 to 0.0 V, at a scan rate 10 mV s
1, pulse amplitude 25 mV, and pulse width 50 ms. Cyclic voltammetry (CV) experiments were performed within the same potential window at

84

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

various scan rates (ν) (10–200 mV s
1) in anodic stripping mode. Since oxygen did not influence the oxidation of any of analytes, the deaeration of the cell content was not required. All DPASV runs for each concentration of test analyte(s) were quantified using the method of standard addition.

3. Results and discussion 3.1. Spectral characterizations The initial covalent bondings (Au–S links) of functionalized thiolated cysteine molecules with GNPs at the electrode surface provided a nano-platform suited to further modification with the recognition element. The Au–S linkages were characterized through FT-IR studies. Accordingly, MAC solution showed a small peak at 2650 cm
1 corresponding to the S–H stretching vibration (Fig. S1C). Given enough time as much as 60 min, this band completely disappeared on Au–S linking (Fig S1F). The Cys has a strong effect on UV–vis spectrum of GNPs (Mocanua et al., 2009). The colloidal gold reportedly shows λmax about 528 nm (Majzik et al., 2010). This peak was literally reduced on combination with MAC solution under a color change from red to blue and a new adsorption band was emerged at 700 nm (Fig. S2a and b). This could be attributed to the anisotropic optical property of GNPs aggregates (Mocanua et al., 2009). FT-IR (KBr) spectra (Fig. S1) of templates, monomer, MIP, and MIP-adduct are compared to support the suggested mechanism of analyte rebinding (Scheme 1) in aqueous medium. The major adsorption bands of NGLY [1250 (P–OH), 979 cm
1 (PQO), curve A], GLU [1517 (NH bending), 1634 cm
1(CQO), curve B] and MAC [1634 (amide I), 1731 cm
1 (CQO), curve C] are found to be shifted downward to 1162, 943, 1438, 1600, 1590, and 1654 cm
1 (curve D), respectively. This supports H-bondings accompanying electrostatic interactions between template(s) and zwitter ionic monomer molecules, as depicted in Scheme 1. Interestingly, all

bands corresponding to both templates disappeared and the monomer bands are reinstated at their original positions (curve E), when the templates were retrieved from the MIP-adduct. 3.2. Morphological characterizations Several factors such as surface roughness, porosity of film, and inclusion of defects affect the current response of the electrode. The SEM image of GNPs-PGE (Fig. 1C) shows somewhat nonuniform distribution of highly packed and aggregated GNPs on the electrode surface. In contrast, the MIP-adduct modified GNPs-PGE (Fig. 1D) has a relatively compact, smooth and rigid structure with non-visible nanoparticles. Interestingly, upon templates retrieval, the MIP-modified GNPs-PGE surface revealed many micro-pores of different depths and apertures with dispersed MWCNTs (Fig. 1E) in the nanostructure. Fig. 1F displays the side view of MIP-modified GNPs-PGE suggesting the film thickness of about 48.6 nm. TEM micrographs of pure GNPs samples, before and after MAC additions, are shown in Fig. 1A and B, respectively. Accordingly, GNPs revealed a relatively uniform size distribution in nanometer scale with an average diameter of a discrete particle as 20 71.9 nm (Fig. 1A). However, the self assembled MAC-GNPs conjugates demonstrated an agglomerated structure with an average diameter of 30 73.8 nm (Fig. 1B). Surface morphologies were further examined from AFM (threedimensional) images recorded under semi-contact mode for MIP-modified GNPs-PGE (Fig. S3A) and MIP-adduct-modified GNPs-PGE (Fig. S3B). Accordingly, an elevated MIP layer with the surface-height 67.5 nm, root mean square roughness (Rq) 4.8 nm, and arithmetic mean roughness (Ra) 3.5 nm is obtained. Whereas the surface-height, Rq, and Ra values for the MIP-adduct are obtained to be 81.7, 3.2 and 1.9, respectively. The lower Rq and Ra for adduct revealed somewhat compact and a smoother surface as compared to MIP film. The increase of surface-height on adduct formation vis-à-vis MIP reflects the template embedded structure. The space between the clearly visible cavities shows elevated

Fig. 1. TEM images: (A) GNPs, (B) MAC þ GNPs; SEM images: (C) GNPs-PGE, (D) MIP-adduct, (E) MIP, and (F) side view of MIP-modified GNPs-PGE.

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

bumps of polymer material (Fig. S3B) along with raised peaks (indicative of CNTs), after the template extraction. An average estimate of the thickness of MIP film can be calculated using the following equation (Beaulieu et al., 2002): zðx; yÞ ¼ sðx; yÞ þ t þ Δzðx; yÞ

ð1Þ

where z(x,y) is the surface-height (67.5 nm) of MIP-modified GNPs-PGE, s(x,y) is the surface-height (14.8 nm) as reported for bare PGE (Prasad and Pandey, 2013), t is the average thickness, and Δz(x,y) is the inherent roughness (Rq ¼4.8 nm) of MIP film. Accordingly, the mean thickness (t) of imprinted polymer film on the exposed tip of PGE could be obtained as 48.0 nm, a value very close to that obtained by SEM. 3.3. Polymer characteristics In this work, 3.0 mmol of cross-linker amount along with 0.1 mmol NGLY, 0.1 mmol GLU and 0.2 mmol MAC were found to be optimum for MIP synthesis that revealed the maximum development of DPASV current response. Any amount of cross-linker more than 3.0 mmol revealed a decrease in the current response apparently due to higher cross linkages and thereby the restricted access to analyte adsorption. Further, different template–template–monomer molar ratios (1:1:1, 1:1:2, 1:1:3 and 1:1:4) were attempted to explore the optimum stoichiometry of the molecular complex. The maximum DPASV response of both the analytes was obtained when template–template–monomer ratio of 1:1:2 was used for the polymerization at 70 1C for 3 h. [For details on stoichiometry of MIP-adduct and optimization of polymerization conditions, vide Supporting data Section S.1.] 3.4. Electrochemical behaviors of analytes For electrochemical studies, all operating conditions such as Eacc, tacc, and pH of the test solution were optimized on MIPmodified GNPs-PGE [for details, vide Supporting data Section S.2]. Simultaneous analysis of GLY and GLU is problematic owing to their overlapped oxidation peaks. Derivatization of one of these (particularly, GLY as NGLY), however, helped obtaining their two distinct quantitative oxidation peaks. The redox behavior of NGLY

85

and GLU on bare (Fig. S4A) and MIP-modified GNPs-PGE (Fig. S4B) was explored with the help of CV, in anodic stripping mode, at different scan rates in acetate buffer (pH 5.5). Both analytes, accumulated simultaneously at Eacc
1.30 V for tacc 180 s, were allowed to strip off anodically after 15 s equilibration. This resulted in two pairs of oxidation–reduction current, with potential separation (ΔEp ¼Epc–Epa) of 200 and 310 mV for NGLY and 150 and 310 mV for GLU, on bare and MIP-sensor, respectively (ν ¼100 mV s
1, NGLY 1.50 ng mL
1 and GLU 0.51 ng mL
1). The bare electrode revealed broad and restricted oxidation peaks, even at higher scan rates Z50 mV s
1, despite reasonably good ΔEp (Fig. S4A). On the other hand, the relatively sharp and sensitive peaks at scan rates Z10 mV s
1 on MIP-modified GNPs-PGE (Fig. S4B) can be attributed to the faster electron-transfer kinetics, under electro-catalytic action of the modified electrode. The gradual positive shifts for NGLY and GLU peaks with scan rate indicate the difficulty in stripping process within the short time. For comparison, the CV behavior of NGLY (2.00 ng mL
1) and GLU (1.24 ng mL
1) were also investigated individually at different scan rates on MIPmodified GNPs-PGE, under similar conditions (Fig. S5A, B). Accordingly, both NGLY and GLU peaks appeared at same positions as observed in binary mixture. This reflects that electron-transfer processes for both the analytes are unperturbed, irrespective of double imprinted MIP surface. The electrode processes in the present instance are suggested in Fig. 2. Accordingly, given Eacc as high as
1.3 V, NGLY molecules were instantaneously reduced in MIP cavities followed by their oxidative stripping (Alemu and Hlalele, 2007; Wheeler et al., 1988). On the other hand, GLU (without any reducible group) directly underwent anodic stripping from the MIP film (Hampson et al., 1970; Khenifia et al., 2009). Notably, the oxidative imine derivative of GLU could not be further oxidized to yield nor-nitrile (or hydrolyzed as nor-aldehyde) within the limited time span of DPASV. This is in contrast to what suggested earlier for the oxidation of primary amine by the other workers (Hampson et al., 1969). The anodic peak current (Ipa) for both the analytes increased linearly with square root of scan rate (ν1/2) in the linear range 10–200 mV s
1, (R2 ¼0.99). This indicated that electro-oxidations involved were purely diffusion controlled. Chronocoulometric experiments were performed to estimate the surface coverage (Г1) and diffusion coefficient (D) for NGLY and

Fig. 2. Electron-transfer mechanism involved in DPASV of NGLY and GLU.

86

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

GLU. From the slope of the Anson plot, ‘D’ values were calculated as 0.31  10
6 cm2 s
1 (NGLY) and 3.56  10
6 cm2 s
1 (GLU). This indicates an approximately 10-fold higher D value of GLU on MIP-modified GNPs-PGE. [For details on chronocoulometry, vide Supporting data Section S.3.] Electron-transfer coefficient (α) can be obtained from the slope [2.3RT/(1
α)nF] of a linear profile [Epa ¼f(log ν)]. In this work, the same α value (0.75) for both analytes corroborated their identical nature, in terms of electro-oxidation at the electrode surface. The electron transfer rate constant (k) can also be calculated on substitution of α and ΔEp values, in the following equation (ΔEp Z200 mV) (Laviron, 1979):   RT
αð1
αÞnF ΔEp =2:3RT log k ¼ α log ð1
αÞ þ ð1
αÞlog α
log nF ν ð2Þ where F is the Faraday constant, ν is the scan rate (V s
1), R is the gas constant, T is the temperature. Despite the fact that both analytes have shown almost similar k values (3.483  10
2 s
1 NGLY and 3.267  10
2 s
2 for GLU, respectively), it was variation in D values that caused well-defined peak resolution of both analytes in the binary mixture.

(Ip, μA) and concentration (ng mL
1). Likewise, NGLY was taken at a fixed large concentration (155.58 ng mL
1) and GLU was increased from 2.94 to 7.11 ng mL
1 (Fig. 3B (curves a–f)). This also followed a linear regression curve (Fig. 3B, inset). When both analyte concentrations were simultaneously varied in the mixed solution, these revealed symmetrical DPASV peaks at their respective oxidation potentials with MIP-modified GNPs-PGE (Fig. 3C (curves a–e)). The calibration equations (Table 1) for NGLY and GLU are found to be linear (Fig. 3C, inset) for a wide range of concentrations in all cases. It is worth noting that NGLY shows a wider detection range in comparison to GLU. However, the limit of detection (LOD) of GLU was slightly lower than NGLY. To evaluate double-template MIP applicability in complex real samples, the proposed sensor was validated with typical agricultural soil and human blood serum samples which were free from GLY and GLU contaminations. Therefore, these samples were fortified with randomly selected spiked levels of NGLY and GLU.

3.5. Langmuir binding isotherm Selective adsorption of both analyte(s) from bulk to their respective MIP cavities could be approximated on the basis of the Langmuir equation (Smiechowski et al., 2006):

θ¼

bc 1 þ bc

ð3Þ

where θ is the ratio of the surface coverage Γ1 at any concentration ‘C’ to its maximum surface coverage Γmax. Eq. (3) can be rearranged as C

Γ˚

¼

1 C max þ max Γ Bads Γ

ð4Þ

Accordingly, linear equations [C/Γ1 ¼(5.617 71.917)  10
7C þ (1.9177 0.178) (R2 ¼ 0.99) and C/Γ1 ¼(0.061 70.015)  10
9C þ (0.931 70.200) (R2 ¼0.99)] representing C/Г1 vs. C plots are obtained for NGLY and GLU, respectively. The intercepts (equivalent to slope/b) of above equations revealed adsorption coefficient (b) to be 2.93  109 and 6.54  107 L mol
1. The Gibbs free energy change (ΔG ¼
RT ln b) due to analyte adsorption could be calculated as
63.96 and
54.55 kJ mol
1 for NGLY and GLU, respectively. 3.6. DPASV determination of glyphosate and glufosinate The proposed MIP-modified GNPs-PGE sensor was used for the simultaneous analysis of GLY (in terms of NGLY) and GLU in aqueous as well as real samples. For this, DPASV technique was preferred to CV because of its higher measurement sensitivity (without any background current) in the given time scale (pulse amplitude 25 mV, pulse width of 50 ms). Typical DPASV curves on MIP-modified GNPs-PGE for individual additions of NGLY and GLU were recorded (Fig. S6. A) in order to demonstrate specific and quantitative analyte adsorption by the modified electrode. [For details on individual DPASV runs, vide Supporting data Section S.4.] Simultaneous determination of NGLY and GLU together in a mixture is the main goal of this work. This warrants studies of the influence of NGLY on the redox behavior of GLU and vice-versa, using the proposed sensor. In this context, GLU was kept at a fixed concentration (3.16 ng mL
1) and the NGLY concentration was varied from 20.44 to 123.61 ng mL
1 (Fig. 3A (a–f)). This responded a linear calibration plot (Fig. 3A, inset) between current

Fig. 3. (A) DPASV runs of NGLY at MIP-modified GNPs-PGE in the presence of 3.16 ng mL
1 GLU in acetate buffer (pH 5.5). [NGLY concentrations (from a to f): 20.49, 36.38, 63.29, 80.40, 92.21 and 123.61 ng mL
1.] (B) DPASV runs of GLU at MIP-modified GNPs-PGE in the presence of 155.58 ng mL
1 NGLY in acetate buffer (pH 5.5). [GLU concentrations (from a to f): 2.87, 4.52, 5.01, 5.49, 6.19 and 7.11 ng mL
1.] (C) DPASV runs obtained for the simultaneous analysis of NGLY and GLU at MIP-modified GNPs-PGE in acetate buffer (pH 5.5). NGLY contents (from a to e): 3.98, 15.64, 49.64, 126.70 and 176.23 ng mL
1. GLU contents (from a to e): 0.50, 1.70, 2.95, 5.55 and 15.58 ng mL
1. [Operating conditions: Eacc
1.3 V, tacc 180 s, supporting electrolyte acetate buffer (0.1 M, pH 5.5), pulse amplitude 25 mV, scan rate 10 mV s
1.]

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

87

Table 1 Analytical results of DPASV measurements on MIP-modified GNPs-PGE in aqueous and real samples. Linear range (ng mL
1)

Recoverya (%)

1.00–173.42

97.8–102.3

0.34

0.13

0.51–27.71

98.8–102.5

0.19

0.63

3.98–176.23

98.1–100.9

0.35

0.48

0.54–3.96

98.5–101.9

0.19

2.56

20.44–123.61

97.5–102.3

0.36

0.74

2.94–7.11

99.9–102.3

0.19

3.01

NGLY Ip ¼ (0.5077 0.001)C þ(0.023 7 0.026), n ¼7, R2 ¼ 0.99 GLU Ip ¼ (17.598 7 0.901)C þ(2.1927 2.919), n ¼7, R2 ¼ 0.99

8.28–59.21

98.6–102.8

0.35

0.48

0.54–3.96

98.4–102.1

0.19

5.14

NGLY Ip ¼ (0.4977 0.027)Cþ (0.001 70.097), n ¼3, R2 ¼ 0.99 GLU Ip ¼ (17.8147 1.525)Cþ (2.1717 5.116), n ¼3, R2 ¼ 0.99

6.12–101.63

98.1–110.2

0.35

5.49

1.06–3.10

99.6–101.3 00.21

4.56

0.50–31.99

97.8–101.9

0.0.19

0.49

0.50–25.21

97.9–102.1

0.19

0.53

Sample

Analyte/s

Aqueous

Ip ¼ (0.506 70.001)Cþ (0.054 70.0431), n ¼12, R2 ¼ 0.99 GLU Ip ¼ (17.5757 0.112)Cþ (3.105 71.604), n ¼13, R2 ¼0.99 NGLY and GLU NGLY Ip ¼ (0.4977 0.002)Cþ (0.3757 0.241), n ¼8, R2 ¼ 0.99 mixture GLU Ip ¼ (17.859 7 0.455)Cþ (1.652 7 3.608), n ¼8, R2 ¼ 0.99 Ip ¼ (0.479 70.004)Cþ (0.5717 0.300), NGLY (with n ¼7, R2 ¼ 0.99 3.16 ngmL
1 GLU, fixed) IP ¼ (17.6477 0.531)Cþ (5.042 7 2.848), GLU (with n ¼6, R2 ¼ 0.99 155.58 ngmL
1 NGLY, fixed)

Soil

Human serum

Aqueous (single template imprinted sensor)

Regression equation

NGLY

NGLY and GLU mixture

NGLY and GLU mixture

GLY GLU

Ip ¼ (8.0877 0.0405)Cþ (1.77570.779 ), n ¼15, R2 ¼0.99 Ip ¼ (16.257 70.256)Cþ (2.5337 0.956), n ¼15, R2 ¼0.99

LODb (ng mL
1)

RSD (%) (for three sets of LODs)

a

% Recovery ¼ (amount of analyte determined/amount of analyte taken)  100. LOD based on the minimal distinguishable signal for lower concentration of analyte; the LOD concentrations (ng mL
1) of both analytes in the soil can be converted to respective MRL values (mg kg
1) as 1.0 g soil was suspended in 30 mL water. b

Typical DPASV curves for soil and blood serum samples (fortified with a mixture of NGLY and GLU) are shown in Fig. S6B. The corresponding analytical results are shown in Table 1 which depict excellent analytical figure of merits (i.e, precision, selectivity, sensitivity, recovery, etc.). Notably, approximately 34-fold higher slopes of linear equations (Table 1) of GLU (whether measured individually or in binary mixture) were observed than those of NGLY in aqueous and real samples. This indicates relatively high sensitivity of the measurement of GLU with larger DPASV peak in aqueous and real samples (Fig. S6. B). The reason could be the higher diffusion kinetics of GLU through the MIP film, despite its limited binding sites as compared to that of NGLY (vide Supporting data Section S.3). The proposed sensor afforded an independent ingress and egress of both analytes in respective cavities, and thereby un-perturbed electro-oxidations, to obtain well-resolved DPASV peaks. 3.7. Single-template imprinted sensor vis-à-vis double-template imprinted sensor It would be better to comprehend the effectiveness of doubletemplate imprinted polymer by comparing analytical results obtained on single template imprinted MIP sensor vis-à-vis double-template imprinted MIP sensor (the procedure for imprinting and operating conditions for voltammetric measurements in both cases are identical). When GLY (underivatized) was singly imprinted, the corresponding sensor (single-template MIP-modified GNPs-PGE) appeared to be highly sensitive responding a linear regression equation with higher slope (8.087, Table 1) and lower LOD. Comparatively, the slope value (0.506, Table 1) realized with NGLY when used individually for rebinding on the dual-template imprinted sensor was drastically diminished. This may be attributed to the solvation effect in water. As a consequence, GLY molecule might succumb to intramolecular electrostatic attractions between –NH2 þ group and 4CQO, carboxylate and phosphate anions, at pH 5.5. This may result in a coil-shaped

conformation of GLY and consequently the higher diffusion of analyte. NGLY molecule, on the other hand, may restrict such interactions owing to the intramolecular repulsion especially between 4CQO and –NQO moieties. This favoured presumably a linear straightened structure with poor diffusivity (mass-transport) of NGLY molecules toward their respective molecular cavities. Although sensitivity is drastically diminished due to the relatively slow diffusion of NGLY in aqueous environment, its detection with the proposed sensor was found to be unhindered in real samples. Interestingly, when sample containing GLU alone was respectively measured on both single and double imprinted MIP sensors, slopes (sensitivities) of corresponding linear calibration equations and LODs are quite matching with each other (Table 1). This reflects an unperturbed flux of electron and analyte diffusion in GLU cavities on either of the electrodes. The above situation remained unaltered even with simultaneous recapture of both analytes into their respective cavities on dual-template imprinted sensor. This corroborates an independent binding kinetics, without any mutual interferences between NGLY and GLU, in terms of the flux of electron and analyte diffusion in the simultaneous determination. Fortuitously, both NGLY and GLU in soil and blood serum samples approximated their behavior on double-template imprinted sensor close to the aqueous sample, in terms of slope values and detection sensitivity (Table 1). This indicates the feasibility of real sample analysis without any matrix effect, in this work. 3.8. Stability, reproducibility, and cross reactivity of MIP-modified GNPs-PGE sensor The sensor precision to yield quantitative recovery, electrode– electrode reproducibility, regeneration of electrodes for next uses, stability, and ruggedness were investigated to ensure the practicability of the proposed method for the simultaneous analysis. [For details, vide Supporting data Section S.5.] A comparative study of the proposed sensor with other known methods is presented in

88

B.B. Prasad et al. / Biosensors and Bioelectronics 59 (2014) 81–88

Table S1. This revealed higher detection sensitivity of the present method for both analytes, barring a few techniques like microchip electrophoresis, CE with contactless conductivity detection, and LC–electrospray tandem-MS which involved costly instrumentations. Moreover, when the proposed MIP and NIP sensors were subjected to interferences (normally encountered in soil (watersoluble compounds) and biological fluids either individually or concomitantly present in high concentration ratios with test analytes), no interferences or cross reactivity were observed (Fig. S7) [For details on cross-reactivity, vide Supporting data Section S6.] 4. Conclusion A simple, fast, selective, and sensitive double-template MIPbased electrochemical sensor was developed for the simultaneous detection of NGLY and GLU in aqueous, soil, and serum samples. Notwithstanding the fact that one of the analytes i.e. GLY has to be derivatized with a nitrosating agent as NGLY before dual imprinting to avoid overlapped current response in this work, the doubletemplate (NGLY and GLU) imprinted polymer-modified GNPs-PGE was found to be convenient, time-saving and cost-effective as compared to corresponding underivatized single-template (GLY or GLU) sensor. The proposed sensor showed almost independent electrodics identical to the single-template sensor, without any cross-perturbation, insofar as specific analyte recapture, mass, and electron transports are concerned. Thus, the sensitive and selective detection with the doubly imprinted polymer nanofilm-modified electrochemical sensor assures a reliable analysis of GLY and GLU in soil and human serum samples, without matrix complications. Acknowledgment Authors thank Council of Scientific and Industrial ResearchUniversity Grant Commission (CSIR-UGC), New Delhi (R./Dev./Ch./ (UGC-JRF-260)/S-01) for granting a senior research fellowship to one of us (DJ). Instrumental and financial helps were procured out of a recent Project (SR/S1/IC-30/2010) funded by the Department of Science and Technology, New Delhi. Appendix A. Supporting data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2014.03.019. References Alemu, H., Hlalele, L., 2007. Bull. Chem. Soc. Ethiop. 21, 1–12. Atta, N.F., Galal, A., El-Ads, E.H., 2012. Electrochim. Acta 69, 102–111. Beaulieu, L.Y., Rutenberg, A.D., Dahn, J.R., 2002. Microsc. Microanal. 8, 422–428. Blanco-López, M.C., Lobo-Castañón, M.J., Miranda-Ordieres, A.J., Tuñón-Blanco, P., 2004. Trends Anal. Chem. 23, 36–48. Chang, S.Y., Wei, M.Y., 2005. J. Chin. Chem. Soc. 52, 785–792. Cikalo, M.G., Goodall, D.M., Mathews, W., 1996. J. Chromatogr. A 745, 189–200. Corbera, M., Hidalgo, M., Salvadó, V., Wieczorek, P.P., 2005. Anal. Chim. Acta 540, 3–7. Dai, C., Zhang, J., Zhang, Y., Zhou, X., Duan, Y., Liu, S., 2013. Environ. Sci. Pollut. Res. 20, 5492–5501. Fujii, T., Ohata, T., Horinaka, M., 1996. Proc. Jpn. Acad. B 72, 7–10. FAO, 2013. Food and Agriculture Organization. Food Standards Programme. 〈http:// www.codexalimentarius.net/download/report/655/al29_24e.pdf〉 2006, (accessed 18.12.13).

Gao, W., Song, J., 2005. J. Electroanal. Chem. 576, 1–7. Goodwin, L., Startin, J.R., Keely, B.J., Goodall, D.M., 2003. J. Chromatogr. A 1004, 107–119. Guan, G., Liu, B., Wang, Z., Zhang, Z., 2008. Sensors 8, 8291–8320. Guo, Y., Guo, T., 2013. Chem. Commun. 49, 1073–1075. Hampson, N.A., Lee, J.B., MacDonald, K.I., Shaw, M.J., 1970. J. Chem. Soc. B), 1766–1769. Hampson, N.A., Lee, J.B., Morley, J.R., Scanlon, B., 1969. Can. J. Chem. 47, 3729–3736. Hanke, I., Singer, H., Hollender, J., 2008. Anal. Bioanal. Chem. 391, 2265–2276. Hao, C., Morse, D., Morra, F., Zhao, X., Yang, P., Nunn, B., 2011. J. Chromatogr. A 1218, 5638–5643. Hoerlein, G., 1994. Rev. Environ. Contam. Toxicol. 138, 73–145. Hori, Y., Manami Fujisawa, M., Shimada, K., Hirose, Y., 2003. J. Anal. Toxicol. 27, 162–166. Jenkins, A.L., Yin, R., Jensen, J.L., 2001. Analyst 126, 798–802. Jiang, J., Lucy, C.A., 2007. Talanta 72, 113–118. Jing, T., Wang, Y., Dai, Q., Xia, H., Niu, J., Hao, Q., Mei, S., Zhou, Y., 2010. Biosens. Bioelectron. 25, 2218. Kataoka, H., Ryu, S., Sakiyama, N., Makita, M., 1996. J. Chromatogr. A 726, 253–258. Khenifia, A., Derriche, Z., Forano, C., Prevot, V., Mousty, C., Scavetta, E., Ballarin, B., Guadagnini, L., Tonelli, D., 2009. Anal. Chim. Acta 654, 97–102. Khrolenko, M.V., Wieczorek, P.P., 2005. J. Chromatogr. A 1093, 111–117. Laviron, E.J., 1979. Electroanal. Chem. 101, 19–28. Li, D., He, Q., Li, J., 2009. Adv. Colloid Interface Sci. 149, 28–38. Li, J., Chen, Z., Li, Y., 2011. Anal. Chim. Acta 706, 255–260. Majzik, A., Fülöp, L., Csapó, E., Bogár, F., Martinek, T., Penke, B., Bíró, G., Dékány, I., 2010. Colloids Surf. B 81, 235–241. Matsui, J., Sodeyama, T., Tamakib, K., Sugimoto, N., 2006. Chem. Commun., 3217–3219 Mocanua, A., Cernicab, I., Tomoaiac, T., Bobosa, L.D., Horovitz, O., Tomoaia-Cotisel, M., 2009. Colloids Surf. A 338, 93–101. Motojyuku, M., Saito, T., Akieda, K., Otsuka, H., Yamamoto, I., Inokuchi, S., 2008. J. Chromatogr. B 875, 509–514. Moye, H.A., Miles, C.J., Scherer, S.J., 1983. J. Agric. Food Chem. 31, 69–72. Prasad, B.B., Madhuri, R., Tiwari, M.P., Sharma, P.S., 2010. Sens. Actuators B 146, 321–330. Prasad, B.B., Jauhari, D., Tiwari, M.P., 2013. Biosens. Bioelectron. 50, 19–27. Prasad, B.B., Pandey, I., 2013. Sens. Actuators. B 186, 407–416. Rajabzade, H., Daneshgar, P., Tazikeh, E., Mehrabian, R.Z., 2012. E-J. Chem. 9, 2540–2549. Richard, S., Moslemi, S., Sipahutar, S., Benachour, H., Seralini G.E., N., 2005. Environ. Health Perspect. 113, 716–720. See, H.H., Hauser, P.C., Ibrahim, W.W.A., Sanagi, M.M., 2010. Electrophoresis 31, 575–582. Simoes, F.R., Mattoso, L.H.C., Vaz, C.M.P., 2006. Sens. Lett. 4, 319–324. Smiechowski, M.F., Lvovich, V.F., Roy, S., Fleischman, A., Fissell, W.H., Riga, A.T., 2006. Biosens. Bioelectron. 22, 670–677. Songa, E.A., Somerset, V., Waryo, T., Baker, P.G.L., Iwuoha, E.I., 2009a. Pure Appl. Chem. 81, 123–139. Songa, E.A., Waryo, T., Jahed, N., Baker, P.G. L., Kgarebe, B.V., Iwuoha, E.I., 2009b. Electroanalysis 21, 671–674. Sreenivasan, K.J., 2001. Appl. Polym. Sci. 82, 889–893. Stalikas, C.D., Pilidis, G.A., 2000. J. Chromatogr. A 872, 215–225. Stalikas, C.D., Konidari, C.N., 2001. J. Chromatogr. A 907, 1–19. Suedee, R., Seechamnanturakit, V., Suksuwan, A., Canyuk, B., 2008. Int. J. Mol. Sci. 9, 2333–2356. Tiwari, M.P., Madhuri, R., Kumar, D., Jauhari, D., Prasad, B.B., 2011. Adv. Mater. Lett. 2 (4), 276–280. Teófilo, R.F., Reis, E.L., Reis, C., da Silva, G.A., Kubota, L.T., 2004. J. Braz. Chem. Soc. 15, 865–871. Tseng, S.H., Lo, Y.W., Chang, P.C., Chou, S.S., Chang, H.M., 2004. J. Agric. Food Chem. 52, 4057–4063. Utku, S., Yılmaz, E., Türkmen, D., Uzun, L., Garipcan, B., Say, R., Denizli, A., 2008. Hacet. J. Biol. Chem. 36, 291–304. Walsh, L.P., McCormick, C., Martin, C., Stocco, D.M., 2000. Environ. Health Perspect. 108, 769–776. Wei, X., Gao, X., Zhao, L., Peng, X., Zhou, L., Wang, J., Pu, Q., 2013. J. Chromatogr. A 1281, 148–154. Wheeler, J.F., Lunte, C.E., Heineman, W.R., Adams, L., Hess, E.V., 1988. Exp. Biol. Med. 188, 381–386. Xiao, Y., Ju, H.X., Chen, H.Y., 1999. Anal. Chim. Acta 39, 73–82. Xin, J., Qiao, X., Xu, Z., Zhou, J., 2013. Food Anal. Methods6, 274–281. Zhao, Y.L., Garrison, S.L., Gonzalez, C., Thweatt, W.D., Marquez, M., 2007. J. Phys. Chem. A 111, 2200–2205. Zhao, P., Yan, M., Zhang, C., Peng, R., Ma, D., Yu, J., 2011. Spectrochim. Acta Mol. Biomol. Spectrosc. 78, 1482–1486.