Non-enzymatic electrochemical sensing platform based on metal complex immobilized carbon nanotubes for glucose determination

Abstract

This work demonstrates the preparation of an electrochemical sensing platform (ESP) based on nickel salophen (abbreviated as Ni^II–S, where salophen is N,N′-bis(salicylidene)-1,2-phenylenediamine) immobilized multiwall carbon nanotubes (MWCNT) for electrochemical sensing of glucose in an alkaline medium. Ni^II–S is immobilized onto MWCNT by stirring MWCNT and Ni^II–S in DMF (MWCNT–Ni^II–S). The MWCNT–Ni^II–S is characterized by physicochemical and electrochemical techniques. Then, a glassy carbon (GC) electrode was modifies with the MWCNT–Ni^II–S composite (GC/MWCNT–Ni^II–S) and it exhibits efficient electrocatalytic activity towards glucose oxidation when compared to GC modified with Ni^II–S. Cyclic voltammetry and chronoamperometry techniques are performed to understand the reaction kinetics and to determine the kinetic parameters such as electron transfer coefficient, rate constant of electrode reaction and catalytic rate constant. At the GC/MWCNT–Ni^II–S ESP a linear calibration range for the glucose determination is observed from 500 nM to 20 mM with a limit of detection of 80 nM (S/N = 3) and sensitivity of 70 μA mM⁻¹. Further, the present ESP is successfully utilized for the detection of glucose in a human blood serum sample with a good recovery (96.4–104.1%).

---

1. Introduction

Electrochemical sensing platforms (ESPs) based on carbon nanotubes have recently been explored for biomolecule sensing due to their exceptional conductivity, thermal and chemical stability, high surface area and good mechanical strength.^1–5 It has already been demonstrated that multiwall carbon nanotubes (MWCNT) enhanced the electron transfer activity at electrodes for several reactions, like oxidation of glucose³ and alcohol.^6,7 On the other hand, Schiff base based transition metal complexes^8–12 are extensively employed for polymerization,^6,8 reduction of oxygen⁹ and ketone¹⁰ and oxidation of hydrazine,⁹ phenol,¹⁰ cyclohexene,¹² and alcohols.¹¹ Usually, transition metal Schiff base complexes are employed in homogeneous catalysis. However, they suffer from corrosion, contamination of reaction products and separation difficulties.¹² To overcome these difficulties, heterogeneous catalysis is preferred,⁶ which can be achieved either by dispersing the metal complexes on solid support or chemical binding of metal ions on a suitably functionalized support.⁶ Keeping this strategy in mind, in the present study, nickel–salophen (abbreviated as Ni^II–S, salophen = N,N′-disalicylidene-1,2-phenylenediamine) is immobilized on MWCNT to prepare heterogeneous catalyst for glucose oxidation. Due to its efficient catalytic properties and defined electrochemical characteristics,¹² Ni^II–S is chosen to immobilize on MWCNT. In addition, Ni^II–S is used in therapeutic applications due to their anti-bacterial and anti-tumor properties.¹³ Very recently, researchers turned their attention on materials such as graphene/nano-gold composites,¹⁴ nickel oxide–CNT hybrids³ and several other related materials^15,16 for glucose determination.

Determination of glucose has significant applications in food industry, metabolic disorders and in biofuel cell devices.^3,17–19 The clinical approaches of diabetes mellitus are being extensively studied since the number of diabetic patients are increasing globally day by day.^12,20 Enzyme based methods such as glucose oxidase–peroxidase (GOD–POD) or biosensors are commonly used for the diagnosis of diabetes in the clinical samples.^20,21 However, these methods depend on several factors such as temperature, pH, humidity and toxic chemicals which decrease their sensitivity for glucose determination.^3,21,22 Since the non-enzymatic sensors are free from these demerits they are more preferred than the enzyme based sensors for glucose.^17,18 In the past, the redox couple, Ni(II)/Ni(III) immobilized on various electrodes is utilized to increase the sensitivity of the electrochemical determination.^23,24 However, it should be noted that the use of Schiff base complex of nickel with MWCNTs is very limited.^6,25 Recently, Zhang and co-workers reported the successful application of porous NiO nanosheets grown on graphite disks for non-enzymatic glucose determination¹⁶ and Garcia-Garcia et al. reported porous NiO thin films prepared by reactive magnetron sputtering for effective electrochemical determination of glucose.¹⁵ In the present work, a new strategy is adopted in which Ni^II–S is immobilized onto MWCNT for the oxidation and subsequent determination of glucose in real samples.

2. Experimental

2.1. Chemicals and reagents

Multiwall carbon nanotubes (MWCNT) and glucose were purchased from Sigma-Aldrich and Sisco Research Laboratory Private Limited, India, respectively. Nickel(II)acetate tetrahydrate, 1,2-phenylenediamine, salicylaldehyde, chloroform, N,N-dimethylformamide (DMF) and nitric acid (HNO₃) were purchased from Sd Fine-Chem Limited, India. Chloroform (CHCl₃) was procured from Merck Specialties Private Limited, India. For synthesis and electrochemical studies triple distilled water was used.

2.2. Instrumentation

X-ray diffraction (XRD) patterns were collected using a classical powder diffractometer (1D 3000 Seifert, Germany). Electrochemical experiments were performed on a CH instruments electrochemical workstation (CHI-660C, USA) with a three electrode one compartment system comprising glassy carbon (GC) as working electrode, platinum wire as counter and saturated calomel electrode as reference electrodes. Transmission electron microscopy (TEM) images were taken in transmission electron microscope (make: Philips, model: CM200, resolution: 2.4 Å) operating at 20–200 kV. UV-vis absorption spectra were recorded using UV-vis spectrophotometer (UV 1700 Pharma Spec, Shimadzu). UV-vis absorption spectra of salophen, Ni^II–S, uniform dispersions of MWCNT and MWCNT–Ni^II–S were recorded in CHCl₃. Since MWCNT and MWCNT–Ni^II–S are not homogeneous in CHCl₃, they were ultrasonicated just before recording the UV-vis absorption spectra. FT-IR spectra were recorded using Perkin Elmer (USA). Scanning electron microscopy (SEM) images were collected using SEM-VEGA 3 TESCAN with EDAX (Brucker).

2.3. Synthesis of Ni^II–S complex

The ligand, salophen was prepared by condensation of 1,2-phenylenediamine with salicylaldehyde (1 : 2 mole ratio in ethanol) according to a literature procedure.⁹ The formed product was washed with water and recrystallized from ethanol. The metal complex, Ni^II–S was synthesized with the prepared Schiff base using the following procedure.²⁶ To a stirring solution of Ni(II)acetate tetrahydrate (2.0 mmol) in ethanol, equimolar solution of salophen was added. The reaction mixture was refluxed for 2 h and the pink solution turns to red indicating the formation of Ni^II–S. The product was filtered using Buchner funnel, washed with triple distilled water and kept in vacuum at room temperature for 12 h to dry. The product was recrystallized from ethanol and used for further studies.

2.4. Preparation of MWCNT and Ni^II–S (MWCNT–Ni^II–S) composite

MWCNTs were pretreated according to the reported procedure.²⁷ Typically, MWCNTs were mixed with concentrated nitric acid and heated to 70 °C. The reaction mixture was refluxed for 12 h, filtered, washed with water and dried in vacuum. MWCNT–Ni^II–S composite was prepared by mixing 2.0 mM solution (20 mL) of Ni^II–S in DMF with 0.1 g of pretreated MWCNT. The above mixture was ultrasonicated for 30 min and stirred for 12 h. The resulting material, MWCNT–Ni^II–S composite was filtered, washed, dried and used for electrochemical studies.

2.5. Preparation of the electrochemical sensing platforms

The surface of glassy carbon (GC) electrode was modified with MWCNT–Ni^II–S and used as a electrochemical sensing platform. Typically, 1.0 mg of MWCNT–Ni^II–S was homogeneously dispersed in 10 mL of DMF by ultrasonication. 5.0 μL of the above suspension was drop casted on the GC electrode (area = 0.07 cm²). The electrode was left for drying under the room temperature for ∼4 h to get the required electrochemical sensing platform (ESP) and used within 24 h for electrochemical studies. This ESP is abbreviated as GC/MWCNT–Ni^II–S. Similar procedure was also used to prepare GC/MWCNT electrode. To study the nature of species involved in electrocatalysis, MWCNT–Ni^II–S films were prepared on indium tin oxide (ITO) and it is denoted as ITO/MWCNT–Ni^II–S. The ITO/MWCNT–Ni^II–S was prepared by drop casting 71 μL MWCNT–Ni^II–S suspension onto 1.0 cm² ITO plate and dried at room temperature for 4 h. Similarly MWCNT was also prepared on the ITO plate. The GC/Ni^II–S electrode was prepared by drop casting Ni^II–S (20 μL, 1.0 mM in DMF) on the GC electrode and dried at room temperature for 4 h.

3. Results and discussion

3.1. FT-IR studies

The FT-IR spectrum of ligand, salophen exhibits its C–O stretching frequency at 1276 cm⁻¹ (Fig. S1†) and this band is shifted to 1285 cm⁻¹ for Ni^II–S due to the formation of C–O–Ni bond.²⁸ On the other hand, no significant change was observed in the frequencies of C–C and C=N bonds, indicating that these bonds are not directly involved in metal–ligand bond formation.^6,27,28 MWCNTs show bands for C–C, –OH and C=O at 1600, 3400 and 1719 cm⁻¹, respectively. The presence of a band for the stretching frequency of C=N at 1609 cm⁻¹ in MWCNT–Ni^II–S indicates the immobilization of Ni^II–S on MWCNT (Table S1, ESI†) which is very close to the reported values.^27,29,30 The weak interaction between –COOH group of MWCNT with Ni^II–S may shift the stretching frequency of [double bond splayed left]C=O from 1719 cm⁻¹ in MWCNT to 1710 cm⁻¹ in MWCNT–Ni^II–S. The immobilization of Ni^II–S on MWCNTs is further supported by UV-vis, powder XRD, SEM and TEM analyzes.

3.2. UV-vis absorption studies

UV-vis absorption spectra of salophen and Ni^II–S together with MWCNTs and MWCNT–Ni^II–S are shown in Fig. 1A. The ligand, salophen is deep yellowish in color and shows an absorption band centered at 333 nm (λ_max) while Ni^II–S is deep red in color and exhibits bands at 376 and 478 nm in CHCl₃, similar to the reported values,^13,28 and in line with the results of FT-IR studies. MWCNTs alone (0.01% in CHCl₃) do not display any characteristic bands in this region but the characteristics of MWCNT are well recognized from FT-IR data. The absorption spectrum of MWCNT–Ni^II–S (0.01% in CHCl₃) is similar to that of Ni^II–S (bands at 376 and 478 nm), representing no change in the structural features of Ni^II–S due to the immobilization on MWCNT.⁶ Thus, the presence of Ni^II–S on MWCNT is strongly supported by UV-vis studies, confirming the conclusions drawn from the FT-IR studies.

Fig. 1 (A) UV-vis absorption spectra of 1.0 mM salophen (a), 1.0 mM Ni^II–S (b), 0.01 wt%. MWCNT–Ni^II–S suspension (c) and MWCNT suspension (normalized to 0.0 absorbance at 600 nm) (d) in chloroform. Inset shows the photographs of the respective materials. (B) Powder XRD patterns of MWCNT (a) and MWCNT–Ni^II–S (b) in the 2θ range 10 to 80°.

3.3. Powder-XRD characterization

Fig. 1B shows the powder-XRD diffraction pattern of MWCNT and MWCNT–Ni^II–S. It shows an intense peak around 26° which is a characteristic peak for MWCNT.^31,32 XRD pattern of MWCNT–Ni^II–S is very much similar to that of MWCNT, which represents that there is no structural change occurs in MWCNT due to the immobilization of Ni^II–S. Since the presence of Ni^II–S in MWCNT is already confirmed by UV-vis and FT-IR studies, the absence of XRD reflections due to Ni^II–S indicates that Ni^II–S is anchored on the surface of the MWCNT in different orientations (i.e. probably in an irregular way). These results are very similar to the results of MWCNT bound with nickel salen (where salen = N,N′-bis(4-hydroxysalicylidene)ethylene-1,2-diamine) and salophen used for alcohol oxidation.⁶ Thus, the crystalline form of Ni^II–S might be no longer available on the surface of MWCNT–Ni^II–S, which is further verified by microscopic studies.

3.4. SEM, TEM, and EDAX characterization

The nanotubes are clearly visible in the SEM images of MWCNT (Fig. 2A). Dense net like tubular appearances of MWCNT is clearly observed from the SEM images of Ni^II–S in MWCNT (Fig. 2B) and presence of Ni^II–S on MWCNT is evidenced by the appearance of some nodular features in high resolution SEM images which are similar to the reported literatures.^33,34 Structural confirmation of MWCNT–Ni^II–S is further supported by TEM characterization. The tubular like relatively straight structures of nanotubes are clearly noticeable from Fig. 2A′. Further, immobilization of Ni^II–S on MWCNT can be seen by the appearance of dense black badges on MWCNT in Fig. 2B′. These results are very similar to the MWCNT bound nickel Schiff-base complexes reported by Rajarao et al.⁶ and similar works reported by other groups.^22,30 Powder XRD studies on MWCNT and MWCNT–Ni^II–S are supported by the TEM-SAED analysis exhibiting the similar diffraction patterns. SAED patterns of MWCNT and MWCNT–Ni^II–S show very similar patterns (Fig. S2†) indicating the similar crystalline nature of both materials as observed in powder XRD analysis. The elemental composition of the MWCNT and MWCNT–Ni^II–S were determined by EDAX measurements (Fig. 2A′′ and B′′ and Table S2†). The presence of Ni in MWCNT–Ni^II–S is confirmed by EDAX (Fig. 2B′′). The results obtained from FT-IR, UV-vis, powder XRD, SEM, TEM, and EDAX studies confirmed that Ni^II–S is well immobilized on MWCNT.

Fig. 2 SEM images (A and B), TEM images (A′ and B′) and EDAX (A′′ and B′′) of MWCNT (A, A′ and A′′) and MWCNT–Ni^II–S (B, B′ and B′′).

3.5. Electrochemical characterization

To evaluate the electrochemical characteristics of the GC/MWCNT–Ni^II–S ESP, cyclic voltammograms of 50 consecutive potential sweeps from 0.3 to 0.6 V in 0.1 M NaOH at a scan rate of 20 mV s⁻¹ are recorded and shown in Fig. 3A. The presence of an anodic peak at +0.46 V and the corresponding cathodic peak at +0.38 V indicate the one electron redox chemistry of Ni^II–S. The anodic and cathodic peak currents gradually increases on each cycle^2,24,35 which can be attributed to the formation of a polymeric film through –O–Ni^II–O– (Scheme S1†) bonds^2,24,36 (vide infra). After 50 cycles (from 51 to 100^th cycles), slight increase in current (inset (i) of Fig. 3A) was observed which suggests that the formation of such polymer probably reached to a saturation. After 100^th cycle, the currents are constant indicating no further growth of the polymer. Further, electrochemical studies were carried out using the ESP pretreated for 100 cycles. The I_pa/I_pc ratio is found to be 0.90 at a scan rate 20 mV s⁻¹ revealing that the system is close to the reversible system.³⁷ E_1/2 and ΔE_p values are found to be 0.42 and 0.07 V (at a scan rate of 20 mV s⁻¹), respectively (inset (ii) of Fig. 3A). The anodic and cathodic peak currents increase on increasing the scan rate (Fig. 3B) as expected.^2,24,36

Fig. 3 (A) CV curves (50 consecutive cycles) recorded at GC/MWCNT–Ni^II–S ESP in 0.1 M NaOH at a scan rate of 20 mV s⁻¹. Inset (i) represents the next 50 consecutive cycles (51^st cycle to 100^th cycle) and inset (ii) represents the 101^st cycle. (B) CV curves of GC/MWCNT–Ni^II–S ESP in 0.1 M NaOH at different scan rates from 5 to 200 mV s⁻¹. (C) Plot of anodic (E_pa) and cathodic peak potential (E_pc) vs. log ν for GC/MWCNT–Ni^II–S. (D) UV-vis absorbance spectra of ITO/MWCNT–NiOOH film (a), ITO/MWCNT–Ni^II–S film before (b) and after (c) electrochemical pretreatment (i.e. 100 continuous CV cycling in NaOH).

Rate constant (k_s) of electrode reaction can be calculated from the plot, E_p vs. log ν (Fig. 3C) using eqn (1)^38,39 where α is electron transfer coefficient, M_a and M_c are anodic and cathodic slopes, respectively. The value of α for GC/MWCNT–Ni^II–S and GC/Ni^II–S is calculated as 0.56 and 0.51, respectively.

α = M_a/(M_a − M_c) (1)

The k_s values for GC/Ni^II–S (5.2 s⁻¹) and GC/MWCNT–Ni^II–S (5.7 s⁻¹) are calculated using eqn (2) based on Laviron theory^38,39 where ν_a and ν_c are anodic and cathodic critical scan rates and rest of the symbols signifies their conventional meanings. The obtained higher k_s value at GC/MWCNT–Ni^II–S indicates that the MWCNT promotes the electron transfer between Ni^II–S and the electrode.⁴⁰ The k_s values obtained for GC/MWCNT–Ni^II–S are comparable to the values of other modified electrodes reported by Zhao et al.⁴⁰ and Selvaraju et al.³⁹

3.6. Nature of the Ni^II species on MWCNT before and after electrochemical treatment

The nature of the Ni^II species involved in electrocatalysis is subjected to critical considerations for a long time.^36,41,42 In the present work, the exact species involved in electrocatalysis is carefully analyzed. The observed redox peaks show close similarities to α- or β-Ni(OH)₂ species. However, a careful analysis suggests the formation of polymeric film involving oxo bridges between the Ni. Though the Ni^II–S species did not undergo any structural or chemical change during the immobilization onto MWCNT (vide supra), electrochemical pretreatment characteristics of the MWCNT–Ni^II–S material in 0.1 M NaOH (Fig. 3A) suggests a possible change in the nature of the species. To understand this process, UV-vis spectra of ITO/MWCNT–Ni^II–S film before and after the electrochemical treatment is compared with ITO/MWCNT–NiOOH film (Fig. 3D). The ITO/MWCNT–NiOOH film is prepared by repeated cycling of ITO/MWCNT electrode in an aqueous mixture of NiNO₃, sodium tartrate, polyvinyl pyrrolidone and NaOH (each 2.5 mM).⁴³ The UV-vis absorption spectrum of ITO/MWCNT–NiOOH film shows intense band at 325 nm along with a shoulder band around 475 nm (Fig. 3D-a) while ITO/MWCNT–Ni^II–S film after the electrochemical pretreatment (i.e., after 100 continuous cycles) shows peaks at 303, 407 and 506 nm (Fig. 3D-b). These results clearly indicate that the Ni^II–S species in MWCNT–Ni^II–S material is not changed to NiOOH even after 100 continuous CV cycles. However, it should be noted here that Ni^II–S did not retain its original absorbance spectrum exhibited before the electrochemical pretreatment. Considerable shift in the peak positions are observed as compared to the ITO/MWCNT–Ni^II–S film before the electrochemical pretreatment (Fig. 3D-c). It may be due to the polymerization of Ni^II–S through the formation of –O–(Ni^II–O)_n– (Scheme S1†) during the electrochemical pretreatment.^36,44 According to Hahn et al. Ni(OH)₂ shows intense band around 530 nm after continuous potential cycling⁴⁵ in NaOH. Similarly NiO exhibits an absorption peak at 362 nm.⁴⁶ The above studies support the absence of NiOOH, Ni(OH)₂ or NiO species in the final material and confirm the presence of a polymeric network possessing –O–Ni^II–O– (Scheme S1†) units.^{36,44,47–50} It has been reported the formation of poly-Ni(OH)TAPc (where TAPc is tetraamino metallophthalocyanine) in aqueous 0.1 M NaOH solution through O–Ni–O oxo bridges.⁴⁹ In polymeric [Ni^II(teta)]²⁺ (where teta is C-meso-5,5,7,12,12,14-hexamethyl-1,4,8,11-tetra-azacyclotetradecane), the Ni is interconnected through the oxo bridges.^36,44,50

3.7. Effect of NaOH concentration on Ni^II–S oxidation

Effect of NaOH concentration (at a constant ionic strength of 0.5 M) on the Ni^II–S oxidation is shown in Fig. 4A. The ionic strength of the supporting electrolyte is maintained by adding NaCl to the solution. On increasing the NaOH concentration, the oxidation peak potential, E_pa (Fig. 4B) decreases drastically. However, a plot of log[NaOH] vs. E_pa (inset of Fig. 4B) gives a linear relationship. It signifies that GC/MWCNT–Ni^II–S can be used for glucose oxidation within the NaOH concentration range 0.5 M to 1.0 mM. However, in 5.0 mM NaOH, GC/MWCNT–Ni^II–S shows very weak peak for the oxidation of Ni^II–S. When the NaOH concentration is decreased to 1.0 mM (pH ≈ 11), the peak due to Ni^II–S to Ni^III–S is decreased drastically. A narrow calibration range and slightly less catalytic activity observed in 0.5 M NaOH (data not shown), made us to select 0.1 M NaOH for glucose oxidation and other electrochemical measurements because it shows high sensitivity and broad calibration range.

Fig. 4 (A) CV curves of GC/MWCNT–Ni^II–S ESP at constant ionic strength (0.5 M) with different NaOH concentrations (curves (a) to (f) represents 0.5, 0.1, 0.05, 0.01, 0.005 and 0.001 M NaOH concentrations, respectively) at a scan rate of 20 mV s⁻¹. Ionic strength of 0.5 M is maintained by the addition of required amount of NaCl. (B) Plot of NaOH concentration vs. anodic peak potential (E_pa) of Ni^II–S oxidation at GC/MWCNT–Ni^II–S. Inset displays the plot between log[NaOH] vs. E_pa.

3.8. Electrocatalytic oxidation of glucose

GC/MWCNT–Ni^II–S works as an efficient ESP for glucose oxidation and determination. To demonstrate the efficient electrocatalysis,^2,3,17,24,36 electrochemical response of GC/MWCNT–Ni^II–S on addition of 1 mM glucose to 0.1 M NaOH is shown in Fig. 5A. A significant increase in the anodic current at 0.47 V is observed (from 57 to 124 μA) at the GC/MWCNT–Ni^II–S ESP. This indicates the oxidation of glucose by the oxidized Ni species, Ni^III–S. The increase in catalytic current in the presence of 1 mM glucose at GC/Ni^II–S is less than that observed at GC/MWCNT–Ni^II–S ESP (Fig. 5B). This demonstrates the improvement in the electrochemical oxidation process due to the presence of MWCNT.²¹ Fig. S3† shows the calibration plot for the determination of glucose based on oxidation current observed at GC/MWCNT–Ni^II–S. There is a linear increase in anodic current from 1.0 μM to 15 mM and after that very less increase in oxidation current is observed probably due to the saturation of the catalytic sites. The mechanism for the electrocatalytic oxidation of glucose is expected to be same as reported in literature.^2,3 The observed pair of redox peaks in the absence of glucose (Fig. 5A-b) correspond to Ni(III)/Ni(II) redox process. As shown in Fig. 5A, the anodic peak increases noticeably in the presence of glucose and the cathodic peak decreases during the reverse scan indicating the involvement of Ni(II)/Ni(III) in the electrocatalytic oxidation of glucose. The peak potential of glucose oxidation coincides well with the oxidation of Ni(II) to Ni(III). During the reverse scan, a considerable decrease in the reduction of Ni(III) to Ni(II) is observed. These observations demonstrate that the Ni(III) species actively participates in the electrochemical oxidation of glucose. The glucose oxidation process involving Ni^III–S can be shown in eqn (3) and (4) and the complete process is shown schematically in Scheme 1.

2Ni^II–S → 2Ni^III–S + 2e⁻ (electrode oxidation) (3)

2Ni^III–S + glucose → gluconic acid + 2Ni^II–S (4)

Fig. 5 (A) CV curves of GC/MWCNT (a) and GC/MWCNT–Ni^II–S (b) in 0.1 M NaOH at a scan rate of 20 mV s⁻¹. Curves (a′) and (b′) represents the CV response in presence of 1.0 mM glucose at GC/MWCNT and GC/MWCNT–Ni^II–S respectively, in 0.1 M NaOH at a scan rate of 20 mV s⁻¹. (B) Represents CV curves of GC/Ni^II–S without glucose (a) and with 1.0 mM glucose (b) in 0.1 M NaOH at a scan rate of 20 mV s⁻¹. (C) Amperometry i–t curve recorded for GC/MWCNT–Ni^II–S ESP with incessant addition of glucose from 500 nM to 20 mM at an applied potential 0.52 V in 0.1 M NaOH. Inset shows the enlarged view of amperometric graph from 500 nM to 100 μM of glucose addition. Arrows indicate the total amount of glucose added. (D) Calibration plot for glucose determination by amperometry from 500 nM to 20 000 μM. Inset shows enlarged view of calibration plot from 500 nM to 1000 μM.

Scheme 1 Schematic representation for preparation of modified electrode and electrochemical glucose oxidation.

3.9. Kinetic parameters

Diffusion coefficient (D) of glucose is calculated from chronoamperometry response of GC/MWCNT–Ni^II–S in presence of different concentrations of glucose (Fig. S4†). Slope of I_p vs. t^−1/2 (obtained from the inset (i) of Fig. S4†) is used for the calculation of D by applying Cottrell equation (eqn (5)), where I_p is anodic current in the presence of 0.4 mM glucose, A is the geometric surface area of the GC electrode 0.07 cm², C₀ is concentration of glucose in mol cm⁻³ and t is the elapsed time in s.

I = nAD^1/2FC₀t^−1/2π^−1/2 (5)

The D value is found to be 6.86 × 10⁻⁶ cm² s⁻¹ (n = 11.7 (ref. 24)) which is close to the reported value.²⁴ Catalytic rate constant (k_cat) is calculated from the slope of I_cat/I_L vs. t^1/2 plots (obtained from inset (ii) of Fig. S4†) using eqn (6).^2,34,35

where I_L and I_cat are the currents obtained from chronoamperometry in the absence and presence of different concentrations of glucose. The k_cat value for glucose oxidation at GC/MWCNT–Ni^II–S and GC/Ni^II–S is found to be 9.2 × 10⁵ and 1.42 × 10⁵ M⁻¹ s⁻¹, respectively. A comparison of observed kinetic parameters at GC/MWCNT–Ni^II–S and GC/Ni^II–S are given in Table 1, which indicates that the GC/MWCNT–Ni^II–S has better catalytic activity than GC/Ni^II–S.

Table 1 Comparison of certain electrochemical parameters of GC/Ni^II–S and GC/MWCNT–Ni^II–S in 0.1 M NaOH

Parameters	GC/Ni^II–S	GC/MWCNT–Ni^II–S
ΔEp (mV)	60	70
E1/2 (V)	0.42	0.42
Ipa/Ipc ratio	1.2	0.90
Electron transfer rate constant (ks) (s^−1)	5.2	5.7
kcat (s^−1)	1.45 × 10^5	9.2 × 10^5

---

3.10. Amperometric determination of glucose at GC/MWCNT–Ni^II–S

On the basis of the results obtained from CV, sensitive determination of glucose was carried out by amperometry in 0.1 M NaOH at an applied potential of 0.52 V. The oxidation current at the GC/MWCNT–Ni^II–S ESP is continuously recorded with an applied potential 0.52 V under stirring conditions (Fig. 5C).^3,18,51 Glucose is added to the supporting electrolyte solution at regular intervals. Each addition of glucose increases the oxidation current immediately (within a second). The increase in current is used for the construction of calibration plot for the determination of glucose. Inset of Fig. 5C shows the enlarged view of amperometric curves from 500 nM to 1000 μM of glucose.

Incessant addition of glucose in amperometry shows a linear increase from 500 nM to 20 mM (Fig. 5D). Inset of Fig. 5D represents the enlarged view of calibration plot from 500 nM to 1000 μM of glucose. The limit of detection (based on 3S of the blank, where S is the standard deviation) and sensitivity are found to be 80 nM and 70 μM mM⁻¹, respectively in the range of 500 nM to 20 mM additions of glucose. The observed detection limit is comparable or lower than the recently reported systems for glucose determination including Ni(II)–quercetin complex modified MWCNT with ionic liquid paste electrode (100 nM)² and GC electrode modified with nickel oxide–CNT (56 nM).³ Similarly the observed linear calibration range of the present system is larger than those reported for the above electrodes.^2,3 A detailed comparative assessment of GC/MWCNT–Ni^II–S with other reported sensors are summarized in Table 2 (ref. 2, 3, 14, 18, 21, 24, 48, 52 and 53) which shows the superiority of the GC/MWCNT–Ni^II–S ESP.

Table 2 Comparison of the analytical parameters of GC/MWCNT–Ni^II–S ESP with previously reported glucose sensors^a

Electrode	Linear range	Detection limit	Sensitivity	Applied potential (V)	Reference
a NiO–SCNTs/GCE = NiO single wall carbon nanotubes hybrid nanobelts modified GC electrode, CS = chitosan, NPs = nanoparticles, Ni–PMA/Au–PtNPs/NFs/CNT–GC = Ni(II)–pyromellitic acid (PMA) film immobilized on the surface of bimetallic Au–Pt inorganic–organic hybrid nanocomposite carbon nanotube on GC electrode, GP = graphene modified, AuNPs = gold nanoparticles, Ni(II)–Qu–MCNT–IL–PE = electrodepositing Ni(II)–quercetin complex on the surface of multiwall carbon nanotube ionic liquid paste electrode, GOD = glucose oxidase, PG/OPPyNF/CoPcTS = overoxidized polypyrrole nanofiber onto pencil graphite electrode and modified with cobalt(II)phthalocyanine tetrasulfonate, NCGC = Ni–curcumin modified GC electrode.
Ni(II)–Qu–MCNT–IL–PE	5.0 μM to 2.8 mM	1.0 μM	—	0.6	2
NiO–SCNTs/GCE	0.5–1300 μM	0.056 μM	2980 μA mM^−1 cm^−2	0.45	3
Au/GOD/GCE	0.2–2 and 2–20 mM	17 μM	56.93 and 13.48 μA mmol^−1 L cm^−2	0.39	14
GP/NiO	5 μM to 2.8 mM	1.0 μM	1571 μA mM^−1 cm^−2	0.4	18
Ni–PMA/Au–PtNPs/NFs/CNT–GC	0.1–90 μM	0.055 μM	—	0.35	21
NCGC	1.0 μM to 10 mM	0.1 μM	—	0.7	24
CS/Ni(OH)2-NPs/GC	0.5–10 mM		72 μA mM^−1 cm^−2	0.61	48
AuNPs/GC	0.1 mM to 25 mM	0.05 mM	87.5 μA cm^−2 mM^−1	0.3	52
PG/OPPyNF/CoPcTS	0.25–20 mM	0.1 mM	5.695 μA mM^−1	0.4	53
GC/MWCNT–Ni^II–S	0.5–20 000 μM	80 nM	70 μA mM^−1	0.52	This work

---

3.11. Interference study

Interference study is performed to understand the selectivity of the GC/MWCNT–Ni^II–S ESP towards the oxidation of glucose. Common interfering compounds are tested at GC/MWCNT–Ni^II–S in the presence of 200 μM glucose by amperometry in 0.1 M NaOH at an applied potential of 0.52 V (Fig. 6A). Same amount (i.e. 200 μM) of suspected interfering compounds such as uric acid, cysteine, H₂O₂, fructose, sucrose, ascorbic acid and dopamine are also tested. The interferences observed in presence of different interfering compounds for glucose determination at GC/MWCNT–Ni^II–S in terms of percentage of signal change are summarized in Table S3.† It is observed that uric acid and cysteine gives negligible/no interference (0–5%) although H₂O₂, fructose, sucrose, and ascorbic acid show moderate interference (5–10%). However, dopamine interferes significantly (16.2%) at this concentration level (200 μM). It should be noted that the average amount of glucose present in normal adult human blood sample is around 70–100 mg dL⁻¹ (or 0.7–1.0 mg mL⁻¹),⁵⁴ while the average amount of dopamine is 10–20 ng mL⁻¹ only.⁵⁵ Since the normal dopamine level in human blood is very much less than the glucose level, the maximum possible interference due to dopamine will be around 0.01 to 0.02% only (assuming that dopamine has equal sensitivity as compare to glucose at GC/MWCNT–Ni^II–S). Therefore, it is clear that high interference of dopamine cannot alter the glucose sensing in human blood serum sample. Further addition of glucose in the presence of above compounds provides the oxidation signal without any interference, indicating the high selectivity of glucose at GC/MWCNT–Ni^II–S ESP.^{2,18,46,56,57} The interference of paracetamol (1.0 mM) for the determination of 1.0 mM glucose is tested using CV at a scan rate of 20 mV s⁻¹ (Fig. S5†). It is found that paracetamol alters the oxidation current of glucose by 6.4%.

Fig. 6 (A) Interference study on GC/MWCNT–Ni^II–S ESP in 0.1 M NaOH at an applied potential 0.52 V. Arrow ‘a’ represents addition of 200 μM glucose. Arrows (b) to (i) represents the addition of same amount (i.e. 200 μM) of uric acid, cysteine, H₂O₂, fructose, sucrose, ascorbic acid, and dopamine, respectively. (B) CV curves of GC/MWCNT–Ni^II–S ESP for the oxidation of 1.0 mM glucose in 0.1 M NaOH on the first day (a), 10^th day (b) and 15^th day (c).

3.12. Real sample analysis

Human blood samples are centrifuged (1300 rpm for 30 min) to obtain clear and transparent blood serum samples that contain no fibrinogen or blood cells. Such blood serum samples were obtained from Sir Sundar Lal Hospital, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India. These blood serum samples were stored at 4 °C in a refrigerator when not in use. The serum samples (5.0 or 10.0 μL) were directly added to the 0.1 M NaOH to determine the glucose level using GC/MWCNT–Ni^II–S ESP. CV responses are recorded three times for each sample and the average current is used to calculate the concentration of glucose in the serum samples based on the previously constructed calibration plot. Results obtained from CV experiments for three human blood serum samples are summarized in Table 3. Amount of glucose determined from clinical analysis (glucose oxidase–peroxidase (GOD–POD) method)²⁰ and the results obtained from the present method (by CV) are very close to each other with an average relative error for three samples, 3.8%. Average recovery percentage and relative standard deviation (for 9 experiments) acquired from standard addition method is 99.8% and 2.2% respectively. Thus, it is demonstrated that the proposed ESP, GC/MWCNT–Ni^II–S can be conveniently used for the determination of glucose in clinical samples.

Table 3 Analysis of human blood serum sample for the determination of glucose level

Blood sample	Glucose concentration (mM)		RSD	Relative error (%)	Glucose concentration (mM)			% recovery^c	RSD
	Amount present	Found			Standard added	Total	Found
a Obtained from clinical analysis (GOD–POD method).b Obtained from GC/MWCNT–Ni^II–S through CV (average of three experiments).c % recovery = {[glucose](found)/[glucose](present)} × 100.
1	7.5^a	7.8^b	3.2	4.0	7.5	15.0	14.5	96.4	0.8
					15.0	22.5	22.7	100.6	3.1
					22.5	30.0	30.5	101.5	1.2
2	9.9^a	10.3^b	2.1	4.1	10.0	19.9	20.5	102.6	1.5
					20.0	29.9	29.9	99.9	1.8
					30.0	39.9	39.4	98.6	3.5
3	5.1^a	5.3^b	4.1	3.3	5.0	10.1	9.8	97.1	2.1
					10.0	15.1	14.9	98.2	1.7
					15.0	20.1	20.9	104.1	4.2

---

3.13. Stability, reproducibility and sensitivity of the ESP

There is very little decrease in anodic current of 2 and 7% for glucose (1.0 mM) determination observed after storing the ESP in air dry conditions at room temperature for 10 and 15 days, respectively (Fig. 6B). It indicates the good stability of GC/MWCNT–Ni^II–S for glucose sensing. A freshly prepared GC/MWCNT–Ni^II–S ESP was tested for glucose determination by CV, where four different measurements were recorded with an interval of one min. Similar experiments were repeated with two other freshly prepared GC/MWCNT–Ni^II–S ESPs also. The results show a very little change in oxidation current with a relative standard deviation (RSD) of 3.54% for the total of 12 experiments (4 experiments from each ESP). Comparison of the present GC/MWCNT–Ni^II–S ESP with other works, such as GC electrode modified with nickel-oxide–CNTs,³ GC electrode modified with chitosan membrane embedded Ni(OH)₂ nanoparticles¹⁷ and other studies^14,31 indicate that GC/MWCNT–Ni^II–S ESP has remarkable stability and reproducibility.

4. Conclusions

In the present work, Ni^II–S is immobilized into MWCNT to prepare a composite and it was modified on GC electrode for the determination of glucose. The UV-vis, FT-IR, SEM and TEM studies confirmed the successful preparation of MWCNT–Ni^II–S. The GC/MWCNT–Ni^II–S ESP exhibits efficient electrocatalytic activity for the oxidation of glucose at 0.47 V in 0.1 M NaOH with reliable stability, high sensitivity and reproducibility. It displays wide calibration range from 1.0 μM to 15 mM and 500 nM to 20 mM by CV and amperometry techniques, respectively. The oxidation of glucose at GC/MWCNT–Ni^II–S is dependent on the concentration of NaOH. GC/MWCNT–Ni^II–S has better catalytic activity when compared to GC/Ni^II–S. Thus, the GC/MWCNT–Ni^II–S ESP can be conveniently used as a non-enzymatic sensor for the determination of glucose in clinical samples. Further, the GC/MWCNT–Ni^II–S is highly sensitive towards glucose with good reproducibility. Therefore, it can be used as a promising non-enzymatic sensor for glucose determination in biological samples.

Acknowledgements

Generous fund from UGC (F 42-271/2013 (SR)) and CSIR (01/(2708)/13/EMR-II), New Delhi, India is acknowledged. We are grateful to Prof. O. N. Srivastava, Banaras Hindu University for XRD analysis and sophisticated analytical instrument facility, Mumbai, India for TEM characterizations.

References

1. M. Trojanowicz, Trends Anal. Chem., 2006, 25, 480–489.
2. L. Zheng, J.-q. Zhang and J.-f. Song, Electrochim. Acta, 2009, 54, 4559–4565.
3. W. Yi, D. Yang, H. Chen, P. Liu, J. Tan and H. Li, J. Solid State Electrochem., 2014, 18, 899–908.
4. P. Wei and H. Tanabe, Carbon, 2011, 49, 4877–4889.
5. S. Kundu, T. C. Nagaiah, X. Chen, W. Xia, M. Bron, W. Schuhmann and M. Muhler, Carbon, 2012, 50, 4534–4542.
6. R. Rajarao, T. Kim and B. R. Bhat, J. Coord. Chem., 2012, 65, 2671–2682.
7. Y.-T. Kim, M. A. Matin and Y.-U. Kwon, Carbon, 2014, 66, 691–698.
8. B. M. Trost, Acc. Chem. Res., 2002, 35, 695–705.
9. M. Pal and V. Ganesan, Catal. Sci. Technol., 2012, 2, 2383–2388.
10. K. C. Gupta and A. K. Sutar, J. Mol. Catal. A: Chem., 2008, 280, 173–185.
11. T. S. Reger and K. D. Janda, J. Am. Chem. Soc., 2000, 122, 6929–6934.
12. K. E. Toghill and R. G. Compton, Int. J. Electrochem. Sci., 2010, 5, 1246–1301.
13. M. Asadi, K. A. Jamshid and A. H. Kyanfar, Transition Met. Chem., 2007, 32, 822–827.
14. X. Wang and X. Zhang, Electrochim. Acta, 2013, 112, 774–782.
15. F. Garcia-Garcia, P. Salazar, F. Yubero and A. González-Elipe, Electrochim. Acta, 2016, 201, 38–44.
16. H. Liu, X. Wu, B. Yang, Z. Li, L. Lei and X. Zhang, Electrochim. Acta, 2015, 174, 745–752.
17. A. Ciszewski and I. Stepniak, Electrochim. Acta, 2013, 111, 185–191.
18. S.-J. Li, N. Xia, X.-L. Lv, M.-M. Zhao, B.-Q. Yuan and H. Pang, Sens. Actuators, B, 2014, 190, 809–817.
19. G. Wang, X. He, L. Wang, A. Gu, Y. Huang, B. Fang, B. Geng and X. Zhang, Microchim. Acta, 2013, 180, 161–186.
20. J. Wang, Chem. Rev., 2008, 108, 814–825.
21. M. B. Gholivand and A. Azadbakht, Electrochim. Acta, 2012, 76, 300–311.
22. B. He, L. Hong, J. Lu, J. Hu, Y. Yang, J. Yuan and L. Niu, Electrochim. Acta, 2013, 91, 353–360.
23. F. Qu, Y. Zhang, A. Rasooly and M. Yang, Anal. Chem., 2014, 86, 973–976.
24. M. Y. Elahi, H. Heli, S. Bathaie and M. Mousavi, J. Solid State Electrochem., 2007, 11, 273–282.
25. M. Bazarganipour and M. Salavati-Niasari, Chem. Eng. J., 2016, 286, 259–265.
26. J. B. Raoof, R. Ojani and Z. Mohammadpour, Anal. Bioanal. Electrochem., 2010, 2, 24.
27. G. Gao and C. D. Vecitis, Environ. Sci. Technol., 2011, 45, 9726–9734.
28. M. M. Abd-Elzaher, Synth. React. Inorg. Met.-Org. Chem., 2000, 30, 1805–1816.
29. D. Baskaran, J. W. Mays and M. S. Bratcher, Angew. Chem., Int. Ed., 2004, 43, 2138–2142.
30. A. Shanmugharaj, J. Bae, K. Y. Lee, W. H. Noh, S. H. Lee and S. H. Ryu, Compos. Sci. Technol., 2007, 67, 1813–1822.
31. J. Y. Lee, K. Liang, K. H. An and Y. H. Lee, Synth. Met., 2005, 150, 153–157.
32. K. Shen, Q. Zhang, Z.-H. Huang, J. Yang, G. Yang, W. Shen and F. Kang, Composites, Part A, 2014, 56, 44–50.
33. H. Murakami, M. Hirakawa, C. Tanaka and H. Yamakawa, Appl. Phys. Lett., 2000, 76, 1776–1778.
34. B.-J. Yoon, S.-H. Jeong, K.-H. Lee, H. S. Kim, C. G. Park and J. H. Han, Chem. Phys. Lett., 2004, 388, 170–174.
35. J. Taraszewska and G. Rosłonek, J. Electroanal. Chem., 1994, 364, 209–213.
36. U. P. Azad and V. Ganesan, Electroanalysis, 2010, 22, 575–583.
37. A. J. Bard, L. R. Faulkner, J. Leddy and C. G. Zoski, Electrochemical methods: fundamentals and applications, Wiley, New York, 1980.
38. E. Laviron, J. Electroanal. Chem., 1979, 101, 19–28.
39. T. Selvaraju, S. Sivagami, S. Thangavel and R. Ramaraj, Bull. Mater. Sci., 2008, 31, 487–494.
40. G.-C. Zhao, Z.-Z. Yin, L. Zhang and X.-W. Wei, Electrochem. Commun., 2005, 7, 256–260.
41. T. R. Cataldi, E. Desimoni, G. Ricciardi and F. Lelj, Electroanalysis, 1995, 7, 435–441.
42. T. R. Cataldi, D. Centonze and G. Ricciardi, Electroanalysis, 1995, 7, 312–318.
43. T. Ramesh and P. V. Kamath, J. Power Sources, 2006, 156, 655–661.
44. V. Ganesan and R. Ramaraj, J. Appl. Electrochem., 2001, 31, 585–590.
45. F. Hahn, D. Floner, B. Beden and C. Lamy, Electrochim. Acta, 1987, 32, 1631–1636.
46. F. Davar, Z. Fereshteh and M. Salavati-Niasari, J. Alloys Compd., 2009, 476, 797–801.
47. M. Jafarian, F. Gobal, S. Rayati and M. G. Mahjani, Electrocatalysis, 2011, 2, 163–171.
48. J. Bukowska, G. Rosłonek and J. Taraszewska, J. Electroanal. Chem., 1996, 403, 47–52.
49. V. Chauke, F. Matemadombo and T. Nyokong, J. Hazard. Mater., 2010, 178, 180–186.
50. U. P. Azad and V. Ganesan, Electrochim. Acta, 2011, 56, 5766–5770.
51. E. Sharifi, A. Salimi and E. Shams, Bioelectrochemistry, 2012, 86, 9–21.
52. G. Chang, H. Shu, K. Ji, M. Oyama, X. Liu and Y. He, Appl. Surf. Sci., 2014, 288, 524–529.
53. L. Özcan, Y. Şahin and H. Türk, Biosens. Bioelectron., 2008, 24, 512–517.
54. D. Gardner and D. Shoback, Greenspan's basic & clinical endocrinology, McGraw-Hill, New York, 9th edn, 2011.
55. A. D. Association, Clinical diabetes, 2015, 33, 97–111.
56. J. V. Kumar, R. Karthik, S.-M. Chen, K. Saravanakumar, M. Govindasamy and V. Muthuraj, RSC Adv., 2016, 6, 40399–40407.
57. W. Jiang, Z. Chen, X. Cheng, W. Wu, Y. Wu, L. Xu and F. Fu, RSC Adv., 2016, 6, 40100–40105.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra16064f

---