Comparison of Insulin Determination on NiNPs/chitosan- MWCNTs and NiONPs/chitosan-MWCNTs Modified Pencil Graphite Electrode

The rising amount of patients suffering for diabetes mellitus increases the requirements for effective insulin sensors. Carbon materials are a suitable choice for the development of insulin sensors due to their electrochemical characteristics. Pencil graphite electrodes (PGE) represent the trade-off between price and excellent conductive properties. The modification of PGE by NiO and Ni nanoparticles fixed by chitosan results in surface area enlargement and improved electrocatalytic properties. This paper is focused on the comparison of different properties of Ni and NiO nanoparticles and their effect on redox reaction mechanism of insulin and detection characteristics. The electrode modified by Ni nanoparticles displays linear range of 1 μM 5 μM (R 0.80), limit of detection (LOD) of 4.34 μM and sensitivity of 0.12 μA/μM. On the other hand, the electrode modified by NiO nanoparticles displays enhanced electrochemical characteristics such as linear range of 0.05 μM 5 μM (R 0.99), limit of detection of 260 nM and sensitivity of 0.64 μA/μM. These properties make the NiO nanoparticles modified PGE the appropriate candidate for insulin determination.


Introduction
Insulin is a polypeptide hormone that consists of 51 amino acids divided into A and B chain linked by two disulphide bridges. A and B chains are composed of 21 and 30 residues, respectively. It is produced by βcells located in the pancreatic arrays called islets of Langerhans [1] and the normal fasting insulin level in blood reaches maximum value of 25 mIU/l (0.86 µM) [2]. Insulin ensures transport of blood glucose to the cells and controls glucose level in blood [3,4] The dysfunction of insulin hormone production causes the very common disease in the world called diabetes mellitus [5,6].
Therefore it is necessary to focus on development of cheap, fast and exact insulin sensor [7]. There is a variety of analytical methods used for insulin determination which can be classified into two main groups, immune and non-immune [8]. The most frequently used immune methods are radioimmunoassay (RIA) [9], enzyme immunoassay (EIA) [10] and luminescent immunoassay (LIA) [11]. The main disadvantages of these methods are long-lasting analysis and low sensitivity [12], which can be increased by derivatization of insulin with isotopes and florigenic labels [13]. Non-immune methods such as high performance liquid chromatography (HPLC) [14,15] and capillary electrophoresis (CE) [16] require expensive instrumentation, long lasting analysis and complex pre-treatment steps [17]. Among the various methods that have been used for insulin detection, electrochemical assays are still considered the best methods for insulin determination, because they could overcome the shortcomings of the other methods mentioned above [18]. The main advantages of electrochemical methods are low detection limit, wide linear range, low cost instruments and high sensitivity [19].
The most commonly used electrode material for insulin determination is carbon [20]. Different types of carbon electrodes such as glassy carbon electrodes [21], screen printed carbon electrodes [22], pencil graphite electrodes [23] and carbon paste electrodes [24] have been used for insulin determination. In general, PGE consists of graphite and clay. PGEs have several advantages in comparison to other carbon-based electrodes, including, very low cost, simple modification, widespread availability [25] and well defined surface area [11].
Despite carbon being considered the most suitable material for insulin determination, the bare unmodified carbon electrodes have some drawbacks which can be eliminated by modification with different nanostructures [26].
Therefore, various materials for electrode modification such as multiwalled carbon nanotubes (MWCNTs), metal nanoparticles, metal oxide nanoparticles or combination of these materials have been studied. MWCNTs are frequently used in electrochemical determination of insulin for electrode modifications due to their mechanical stability and fast electron transport. MWCNTs are also known as an excellent option to enlarge electrode surface area, create more active sites for insulin oxidation, increase voltammetric current and decrease overpotential [27].
It is also well known that large surface area of MWCNTs combined with metal nanoparticles or metal oxide nanoparticles can improve the performance of the final electrode material [28,29].

ELECTROANALYSIS
become an intensively studied material not only because of their low cost in comparison with Ag, Co, Pd or Pt nanoparticles but also because they possess some desirable characteristics such as high electrical conductivity and good electrocatalytic activity [34]. It is known that the chemical and physical properties of the NiONPs and NiNPs are quite different from bulk Ni and NiO crystals. Particularly, large surface area and excellent magnetic properties are the greatest advantages of synthetized nanoparticles [27]. In addition, it has been revealed that many nickel-based materials exhibit excellent electrocatalytic activity towards the oxidation of a wide range of small compounds, like insulin or glucose in alkaline media [35]. The utilisation of polymer membrane such as Nafion, chitosan, polyethylene glycol prevents the fast occupation of active sites at higher insulin concentration and occupation of active sites by chloride ions in phosphate buffered saline (under physiological condition) [13].
These polymers are also used because of their ability to fix nanoparticles on the electrode surface during electrochemical measurements [36].
In this paper, we have studied the electrocatalytic activity of the NiONPs/chitosan-MWCNTs/PGE and

ELECTROANALYSIS
All electrochemical experiments were performed by using AUTOLAB type PGSTAT302N (Metrohm, Switzerland) with a three-electrode setup using the bare or modified PGE (Herlitz, Germany) as the working electrode, Ag/AgCl (saturated KCl) as the reference electrode and platinum electrode as the counter electrode.
The structure and surface morphology of the electrodes were characterized by SEM (Jeol JSM 7000 N, Japan) with EDX analysis, TEM and STEM.

Preparation of the modified electrode
The surface of carbon PGE was polished with soft sandpaper (400)   and NiONPs/chitosan-MWCNTs/PGE, respectively.

Electrocatalytic oxidation of insulin
Electrochemical behaviour of various PGE surfaces towards insulin oxidation in 0.1 M NaOH and PBS was studied using cycling voltammetry method (CV) in a potential window ranging from -1 V to 2 V with a scan rate of 100 mV/s. As shown in these voltammograms, current response towards insulin oxidation on bare PGE (Fig. 3 b) is significantly lower in comparison to current response towards insulin oxidation on chitosan-MWCNTs/PGE (Fig.   3 c), NiONPs/chitosan-MWCNTs/PGE (Fig. 3 d) and NiNPs/chitosan-MWCNTs/PGE (Fig. 3 e). There is also no apparent peak found for the bare PGE (Fig. 3 b), confirming that the bare PGE has no significant electrocatalytic

Full Paper
ELECTROANALYSIS activity towards insulin in aqueous alkaline solutions. Two oxidation peaks were observed in cyclic voltammograms for NiONPs/chitosan-MWCNTs/PGE (d), and NiNPs/chitosan-MWCNTs/PGE (e) because of the two-step oxidation process on the modified electrode. Moreover, two reduction peaks were detected for both the electrodes, but the difference in the values of oxidation peaks potential indicates the different mechanism of oxidation process.
Firstly, the current response of NiONPs/chitosan-MWCNTs/PGE is significantly lower in comparison with this of the NiNPs/chitosan-MWCNTs/PGE. The mechanism of insulin oxidation on NiONPs/chitosan-MWCNTs/PGE can be shown as follows: (1) The oxidative current peaks at 1.2 V and 1.45 V (Fig. 3 d) The NiO(OH) species present during the electrochemical oxidation of insulin on both electrode surfaces, catalyse the oxidation of insulin and enhance current response when compared to bare electrode [29].The distance between oxidative and redox peaks in both cases indicates irreversible electrochemical oxidation of insulin.

Effect of scan rate and active surface area of working electrode
The effect of scan rate on the electrochemical behaviour was studied in 0.1 M NaOH and PBS for NiONPs/chitosan-MWCNTs/PGE (Fig. 4 A, C) and NiNPs/chitosan-MWCNTs/PGE (Fig. 4 B, D) in the presence of 2 μM insulin. In both cases the anodic peak current exhibited linear increase with the rising scan rate, with correlation coefficient R 2 0.99 and 0.95 for NiONPs/chitosan-MWCNTs/PGE and NiNPs/chitosan-MWCNTs/PGE, respectively. The linear regression equations are:

ELECTROANALYSIS
According these results, the electron transfer between the redox sites of both electrode surfaces is a typical surface-controlled process. The rate determining step is the charge transfer on electrode surface.
Enlargement of the electrode surface area due to the deposition of nanoparticles increases active surface area and number of active sites. The active surface area of electrode was measured by chronoamperometry and calculated according to the Cottrell equation: (8) where i is measured current, n is number of exchanged electrons,

Electrochemical impedance spectroscopy
EIS is important analytical tool which was used to study the catalytic procedure of insulin electrooxidation on different electrode surfaces. Fig. 6 shows The decrease in semicircles diameter with rising insulin concentration can be seen from the Nyquist plots of both NiNPs/Chitosan-MWCNTs/PGE (Fig. 7 A) and NiONPs/Chitosan-MWCNTs/PGE (Fig. 7 B)  with NiNPs/chitosan-MWCNTs/PGE.

Effect of different insulin concentration
On the basis of the voltammetric results described above, the NiONPs/chitosan-MWCNTs/PGE and NiNPs/chitosan-MWCNTs/PGE appears to be a suitable sensor for the sensitive determination of insulin. Fig. 8 shows  (Fig. 8, inset). The sensitivity of modified electrode emerged from the calibration plot with the correlation coefficient of 0.99 was 0.64 µA/µM. The limit of detection, evaluated at a signal to noise ratio of 3:2, was found to be 260 nM. The linear calibration range, sensitivity and detection limit for insulin determination at NiONPs/chitosan-MWCNTs/PGE are comparable with those obtained by common modified electrodes (Tab. 1).
Also correlation coefficient was similar like correlation coefficients obtained by common modified electrode mentioned in Tab. 1 [25,38,39].
Oxidation of insulin (0.05 µM -5 µM) in 0.1 M NaOH and PBS on the NiNPs/chitosan-MWCNTs/PGE surface was also studied. As shown in cyclic voltammograms of insulin on NiNPs/chitosan-MWCNTs/PGE surface (Fig. 9), the anodic peak current (1.1 V) was slightly shifted to lower values, compared to anodic peak current of insulin oxidation on NiONPs/chitosan-MWCNTs/PGE (1.45 V). The plot of current response versus insulin concentration was linear only over the concentration range of 1 μM -5 μM (Fig. 9, inset), which was lower concentration range in comparison with linear concentration range obtained for NiONPs/chitosan-MWCNTs/PGE (0.05 µM -5 µM) (Fig. 8). Moreover, the sensitivity (0.12 µA/µM) was lower, the correlation coefficient was insufficient (R 2 0.78) and linearity of calibration plot deteriorated. The limit of detection, evaluated at a signal to noise ratio of 3:2, was found to be 4.34 µM (Fig. 9, inset). Further this characteristic of NiNPs/chitosan-MWCNTs/PGE surface was worse as compared to characteristic obtained for

ELECTROANALYSIS
NiONPs/chitosanMWCNTs /PGE. According to these results, it can be concluded, that NiONPs particles possess better electrocatalytic properties and are most suitable material for insulin determination.
Tab. 1 provides the comparison of limits of detection, sensitivity and linear dynamic range towards insulin oxidation for different modified carbon electrodes. For all mentioned electrodes the amperometry method of insulin determination was used. As can be seen in Tab. 1 the prepared NiONPs/chitosan-MWCNTs/PGE possessed the widest dynamic range in comparison with other mentioned electrodes (0.05 µM -5 µM).
Even, insulin with concentration 2 µM was determined on five different NiONPs/chitosan-MWCNTs/PGEs and a reproducible current response with a great relative standard deviation (RSD) 4.25% was obtained (Fig. 11).