Manganese dioxide-core–shell hyperbranched chitosan (MnO₂–HBCs) nano-structured screen printed electrode for enzymatic glucose biosensors

Abstract

In this study, the synthesis, characterization and testing of new polymeric–metal oxide nanocomposites for enzymatic glucose biosensors were performed. Among various nano-composites, manganese dioxide-core–shell hyperbranched chitosan (MnO₂–HBCs) provided rapid and high efficiency direct electron transfer from the redox active centre of an immobilized enzyme and screen printed electrode. The assay optimization was achieved after testing the effects of several factors such as type of crosslinking agent, accumulation potential, toxicity of heavy metals and interferences on the bioactivity of GOx. Results demonstrated sensitivity of the proposed method to detect inhibition effects of metal ions and also the response of agents interfering with glucose measurement. A chronoamperometric calibration curve was obtained, and the oxidation current of the enzymatically produced H₂O₂ was linearly dependent on glucose concentration with a detection limit of 7 μg mL⁻¹. Thus, the clinical determination of glucose concentration was performed on blood samples and the results were correlated with a reference method. In conclusion, the current study suggests a new class of electrochemical biosensors and paves the way for further promising applications.

---

Introduction

Electron transfer (ET) in biological systems is a very important phenomenon for the areas of biophysical, biochemical and biomedical sciences. In particular, ET is a major regulating factor for bioelectrochemical systems (BESs) including electrochemical biosensors.¹ Basically, a high performance biosensor with high efficiency of ET is dependent on the platform's material that is used for immobilization of biomolecules.² For instance, in the case of enzyme-based biosensors, denaturation and loss of enzyme bioactivity is resulting from the use of unsuitable platforms for the adsorption of enzymes.^2,3 Thus, recent advances for immobilization strategies were employed to develop enzymatic biosensors.³ In this regards, the use of nanomaterials provided very efficient alternatives for constructing sensitive and selective biosensors due to the stable, active and well-oriented immobilized biomolecules.

Nanostructured materials were categorized as inorganic, organic or hybrid (organic–inorganic or metal–organic) nanostructures. The fabricated nanostructures-based electrochemical biosensors were designed for the detection of infectious diseases or for health care diagnosis.⁴ Since the type and structure of nanomaterials have high impact on the performance of BESs, modification of the electrode surface by using nanoparticles (NPs), nanorods, nanotubes or other nanostructures was conducted.^5,6

Metal nanostructures have been exploited for efficient and sensitive biosensors fabrication.⁷ For instance, gold,⁸ platinum^9,10 and silver^11,12 were used for glucose, cholesterol, and E. coli biosensing, respectively.

On the other hand, better electrocatalytic properties of the metal oxides and the ease of fabrication of their nanostructures make them extremely attractive materials for sensitive biosensing devices. Therefore, the immobilization of target biomolecules such as, cholesterol oxidase,^13–15 glucose oxidase,¹⁶ urease, cytochrome C, tyrosinase or horseradish peroxidase (HRP) on nanostructured metal oxides was successfully obtained.¹⁷

Alternatively, natural biocompatible and biodegradable polymers such as cellulose triacetate (CTA) and chitosan (Cs) were used as a convenient platform for enzyme loading. For instance, film formation of Cs with high mechanical strength, biocompatibility and good water permeability provides a good environment for protein or enzymes immobilization.^18,19 Much more attention has been paid for Cs due to the availability of amines and hydroxyl groups. Thus, Cs was used to support living organisms or enzymes for biosensor applications,^20,21 however, its lower conductivity and limited electrocatalytic activity made the direct electron transfer (DET) between the sensor surface and the active redox centers of the bio-recognition elements not possible.²² Therefore, in most of the previous enzymatic glucose biosensors, redox mediators were employed to promote the electron exchange.^23,24 For instance, to enable the DET and to enhance the conductivity of biocompatible polymers such as Cs, its hybrid nanocomposites with metal and/or metal oxide NPs were implemented.^25–27

Recently, branched polymers (BPs) were suggested to be promising and attractive materials for sensing applications.²¹ For example, BPs of new architectures and better surface fictionalization were proposed for constructing BESs.^26,28 Amongst the BPs, hyperbranched Cs (HBCs) was recently synthesized and used in one of our reported studies for ammonia sensing.²⁹

In the current study, HBCs–metal and metal oxides nanocomposites were developed, characterized and utilized for the direct biosensing of glucose. According to our hypothesis, synthesis of a tree-like structure of core–shell HBCs NPs with high surface area and carrying a huge number of positive charges around the shell by creating more amino groups, was suggested. Consequently, decoration of the HBCs-based NPs' shell with highly electrocatalytic active metal NPs (e.g. Ag NP or Au NP) or metal oxide NPs (e.g. MnO₂ or ZnO NPs) was planned to enhance the interaction between the electrode surface and the active site of the immobilized sensing molecules, e.g. glucose oxidase (GOx). Thus, the main goal of this study is the construction of a novel biosensor platform for efficient and enzymatic glucose biosensing using the synthesized HBCs-based nanocomposites on the screen printed-3 electrode system. Besides, a synergetic electrocatalytic performance of the developed nanocomposite, fast electron communication between the enzyme's active sites and the screen printed electrode is expected from the new approach.

Materials and methods

Materials

Chitosan (Cs) of low molecular weight was obtained from Aldrich (Germany). Methyl acrylate (MA), ethylene diamine (EDA), ammonia (33%) and zinc acetate dihydrate [Zn(CH₃COO)₂·2H₂O] were provided by Acros (Belgium). Screen printed electrode (SPE) was provided by Dropsens (Oviedo, Spain). Cellulose triacetate (CTA), caffeine, ascorbic acid (AA), paracetamol, cellulose acetate (CA), polyvinyl chloride (PVC), glucose and cadmium (Cd) were obtained from (Laboratory Rasayan Grade, LR & Extrapure). Copper (Cu), aluminum chloride (AlCl₃), ferrous chloride (FeCl₂), potassium chloride (KCl), hydrogen peroxide (H₂O₂) and manganese dioxide (MnO₂) were purchased from (BDH Co.). Phosphate buffer saline (PBS) was used as supporting electrolyte and was purchased from Bio-Basic Canada Inc. (Canada). Glucose oxidase enzyme (GOx) was purchased from SORACHIM (Paris, France). All other chemicals and solvents were of analytical grade and were used without further purification.

Methods

Preparation of Cs NPs. Cs NPs were prepared using nano-spray dryer (Buchi, B-90, Buchi Co., Switzerland). Briefly, a predetermined weight of Cs was dissolved in 1% (w/v) acetic acid to yield a solution with a final concentration of 0.2% (w/v). Afterwards, the Cs solution was spray dried at airflow of 135 L min⁻¹ and inlet temperature of 120 °C using 0.7 μm cap and pump no. 4.

Preparation of HBCs NPs. Dendritic polyamidoamine (PAMAM) branches were synthesized through a modified procedure to that described by Tomalia et al.³⁰ as illustrated in Scheme 1. Typically, a solution of EDA (1 g/20 mL methanol) was added dropwise with stirring to MA solution in methanol kept in an ice bath. Then, the reaction mixture was left at room temperature with stirring for 48 h, followed by removal of the excess reactants through evaporation. Afterwards, a solution of the product resulting from the previous step (1 g/5 mL methanol) was added slowly to EDA solution (0.6 g/10 mL methanol) with vigorous stirring in an ice bath, and then left at room temperature with stirring for 3 days. The previous reactions were repeated for another cycle using double of the amounts of EDA and MA to increase the extent of branching (second generation of branching). The synthesized PAMAM branches were then attached to the surface of Cs NPs via their addition (0.5 g) to 0.1 g of the Cs NPs dispersed into dist. water with stirring for 5 days. The resulting HBCs NPs were collected via centrifugation at 12 000 rpm for 10 min followed by washing with dist. water. The amidation of the terminal methyl ester moieties of the synthesized HBCs NPs was performed using alcoholic ammonia solution (10 mL NH₄OH in 30 mL methanol) with stirring at room temperature for 3 days.

Scheme 1 A schematic illustration of the synthesis of HBCs NP.

Modification of screen printed electrode (SPE) with bionanocomposites. A 2% of each polymer; CTA (dissolved in a mixture of 2 : 1 acetone/cyclohexanone), Cs (dissolved in 1% of acetic acid) or HBCs NPs (suspended in PBS) were sonicated with a suspension of different metal or metal oxide NPs (Ag NP, Au NP, ZnO NPs or MnO₂ NPs). Before the electrode modification, SPE was rinsed with isopropanol and dried. On the working electrode surface (located in the center of the strip of the SPE), a thin film of each bionanocomposite was formed by drop casting 20 μL of each suspension.

Electrochemical characterization of the modified SPE. All electrochemical measurements were performed using a computer controlled Gamry Potentiostat/Galvanostat/ZRA G750, which was connected to a three electrode system comprising of a SPE carbon working electrode, a platinum layer auxiliary electrode and an Ag/AgCl reference electrode. Prior to the measurements, the working electrode was electrochemically activated in 0.1 M KCl by 5-cyclic scans from −0.2 to 0.8 V with scan rate of 50 mV s⁻¹. Aliquots of the ferricyanide, FCN (1 to 3 mM) were introduced into the electrochemical cell containing 1.5 mL of KCl (10 mM). Cyclic voltammetric measurements were scanned from −0.2 to 0.8 V (vs. Ag/AgCl) with scan rate of 50 mV s⁻¹.

Testing of crosslinking agents for direct electron transfer. 20 μL of 1% w/v of different crosslinkers such as CA, PVC, Cs or HBCs was added as a second layer over the modified SPE. Eventually, 10 μL of GOx was placed as the outer layer. The surface of the nanocomposite sensor was activated in phosphate buffer (pH 6.8) by 5-cyclic scans from −0.2 to 0.8 V with a scan rate of 50 mV s⁻¹. Then, 50 μL of H₂O₂ with a final concentration of 0.5 to 2 mg mL⁻¹ was introduced into the electrochemical cell containing 1.5 mL of phosphate buffer (pH 7.4).

Chronoamperometric measurements. Chronoamperometric analyses at 0.7 V (vs. Ag/AgCl) using the freshly prepared biosensor was performed with stirring for the determination of accumulation potential effects, glucose calibration curve, inhibitors effects, interferences and blood sample analysis. All the experiments were carried out using standard glucose concentrations (Fig. 1).

Fig. 1 General screening chart of the current readout (conductivity) of the investigated polymers; Cs, HBCs and CTA before and after their physical decoration by four nano-modifiers; Ag NPs, Au NPs, ZnO NPs and MnO₂ NPs. The represented current values were the produced oxidation current of each modified electrode.

Results and discussion

Selection of the best sensor platform

The nano-structured surface of MnO₂–HBCs/SPE was suggested to construct a biocompatible interface with the redox proteins of GOx to enable direct electron transfer, and to maintain the biocatalytic activity of the enzyme. To select the best platform, different bionanocomposites based on natural polymers and their derivatives modified by metal and metal oxides NPs were developed and studied. In this regards, each of suggested natural polymers include Cs, HBCs, CA or CTA was physically decorated by metal or metal oxides nano-particles such as Ag NPs, Au NPs, ZnO NPs or MnO₂ NPs. The prepared composites were then used as individual modifiers for the screen printed electrode surface. From the cyclic voltammograms, it is obvious from that HBCs has the highest current readout activity. It is articulated also from the chart that HBCs–MnO₂ bionanocomposite demonstrated a significant difference in catalytic activity than that of the HBCs before any modification. Therefore, the HBCs–MnO₂ nanocomposite was selected in the current study as the optimum candidate for further investigation towards efficient electrochemical biosensing application (Scheme 2).

Scheme 2 Schematic illustration for electrochemical glucose sensing using the developed HBCs–MnO₂ modified SPE.

Morphological characterization of HBCs–MnO₂ nanocomposite

The morphology of the synthesized MnO₂–HBCs-nanocomposite was examined in comparison to HBCs and the MnO₂ NPs using a high resolution transmission electron microscopy (HRTEM) and field emission scanning electron microscopy (FESEM) as shown in Fig. 2. Fig. 2a confirms the spherical morphology of the prepared HBCs NPs with a size of 400–500 nm. HRTEM micrographs of MnO₂ NPs (Fig. 2b) demonstrated crystalline rod structures of variable size range. Both HRTEM and FESEM (Fig. 2c and d) demonstrated a successful formation of MnO₂–HBCs nanocomposite where the MnO₂ NPs were preferentially deposited on the surface of HBCs NPs.

Fig. 2 HRTEM images of (a) HBCs NP, (b) MnO₂ NP, (c) HBCs–MnO₂ nanocomposite, and (d) the FESEM image of the HBCs–MnO₂ nanocomposite.

Electrochemical of MnO₂–HBCs

Here, the electrochemical characteristics of the new core–shell MnO₂–HBCs NPs NPs, and ZnO–HBCs NPs were investigated using the FCN as a redox probe. Basically, the HBCs has lower catalytic activity, thus when metal oxides NPs were doped into its surface, the catalytic activity of the supported surface was greatly enhanced. As shown in Fig. 3, the peak currents (oxidation/reduction peaks of FCN) located at about 0.2 V (vs. Ag/AgCl) were higher than the unmodified electrode. However, the MnO₂–HBCs exhibited the highest electrochemical signals where 3-fold of current increase was attained by this HBCs–MnO₂ bionanocomposite. This could be attributed to the acceleration of the electron transfer with the strong electrocatalytic functions of the MnO₂ NPs which adsorbed on the top of the HBCs.

Fig. 3 Cyclic voltammograms of redox reaction of FCN (1 mM) on SPE modified with HBCs, HBCs–ZnO and HBCs–MnO₂.

HBCs self-crosslinking for GOx immobilization

To construct a high performance glucose biosensor with a sustainable bioactivity, an effective way for crosslinking was considered. Particularly, immobilization of GOx on the MnO₂–HBCs/SPEs was performed by different crosslinking agents such as PVC, CA, Cs or HBCs. In this regards, a thin film of each crosslinker was formed on the top of the MnO₂–HBCs/SPE. Consequently, an aliquot of the enzyme (GOx) was simply drop-casted. For evaluating the biosensor performance, direct electron transfer from the oxidation of the enzymatically produced H₂O₂ (eqn (1)) was measured. As a result, a strong effect on the enzyme stability and biocatalytic activity was found when the Cs or HBCs were used. The self-crosslinking using HBCs was exhibiting the highest electrochemical signal (Fig. 4). On the other hand, the formed thin layer of PVC on the HBCs–MnO₂/SPE acted as an insulating layer, and therefore a disconnection between the enzyme and the electrode surface was observed, as can be seen in Fig. 4d.

β-Glucose + O₂ (GOx) → β-gluconic acid + H₂O₂ (1a)

H₂O₂ → O₂ + 2H⁺ + 2e⁻ (1b)

Fig. 4 Effects of crosslinking agents (HBCs, Cs, CA or PVC) on the GOx immobilization and catalytic activity.

Effect of HBCs–MnO₂ nanocomposite on the direct electron transfer

Immobilized enzymes, as a whole macromolecule, usually lack the direct communication with the conductive solid surfaces such as electrodes. In this regards, the active centers of enzymes are surrounded by non-conductive protein shells and therefore, the direct electron transfer (DET) between electrodes and the active centers are blocked. Although they are not preferable, mediated electron transfers are widely used. To avoid the utilization of exogenous redox mediators, a DET is recommended. Therefore, the use of GOx/MnO₂–HBCs/SPE for the direct glucose biosensing was proposed in the present study. The electrochemical behavior was evaluated by CV in phosphate buffer containing different glucose concentrations and the direct oxidation of the enzymatically formed peroxides was measured. As a result, the oxidation peak current was well-correlated with the glucose concentration (Fig. 5a). The electrocatalytic properties of MnO₂–HBCs provided high efficiency direct electron transfer between the active centers of the immobilized GOx and the surface of the SPE.

Fig. 5 (a) Immobilization of GOx of different amounts on HBCs–MnO₂ NPs, and (b) effect of accumulation potential on the chronoamperometric measurements of peroxide as a bi-product of glucose oxidation by GOx.

As we may conclude from the current and our previous studies,³¹ the combination of metal oxides NPs with a polymer such as HBCs, as a sensor material, has a better performance than the integration of metal oxides NPs with carbon materials where the peak potential for the oxidation of hydrogen peroxide was lowered by at least 400 mV compared with that obtained in ZnO/GO/CPE.³¹

Chronoamperometric measurements

Effects of accumulation potential. To select the appropriate electrical potentials for chronoamperometric analysis, the catalytic activity of GOx was evaluated at different accumulation potentials, and at a single glucose concentration (1 mg mL⁻¹). As shown in Fig. 5b, the oxidation current of the enzymatically produced peroxide increased with increasing of the applied potentials. However, applied voltage values of more than 700 mV (vs. Ag/AgCl) led to a decrease in the oxidation current. Therefore, the potential of 700 mV was assigned for the next chronoamperometric studies.

Calibration curve. A freshly prepared biosensor was used for determining the calibration curve of glucose (Fig. 6a). The chronoamperometric analysis was carried out at constant potential (700 mV, vs. Ag/AgCl). Direct oxidation of the enzymatically produced hydrogen peroxide was then recorded at 700 mV vs. Ag/AgCl for the quantitative analysis. A standard addition of a certain concentration of glucose was carried out at fixed time intervals (30 s). A fast response towards each addition was noted, which reflects the rapid electron transfer as well as the electro-catalytic activity of the developed nano-structured electrode. After applying a fixed potential at 700 mV (vs. Ag/AgCl), the oxidation current was linearly dependent on the glucose concentration (from 28 μg mL⁻¹ to 93 μg mL⁻¹), then a steady state increase of the oxidation current was observed at the higher glucose concentrations as shown in Fig. 6a. The obtained linear regression coefficient was 0.995 and with a limit of detection (LOD) of 7 μg mL⁻¹.

Fig. 6 (a) Chronoamperometric calibration curve for glucose biosensing at 700 mV, and (b) effect of interference on the catalytic functions of the proposed glucose biosensor at a concentration of each substance of 1 mg mL⁻¹.

Interferences. From the selectivity point of view, the catalytic responses of the proposed biosensor were tested toward the oxidation of caffeine, ascorbic acid (AA) and paracetamol (Fig. 6b). Addition of two consequent glucose concentrations increased the oxidation current, whereas the addition of two different concentrations of caffeine or AA did not show any significant change in the amperometric signal. In a contrary situation, paracetamol exhibited an increase of the oxidation currents which revealed an interference interaction with the GOx function. Interference problem of the acetaminophen has been previously reported.³²

Toxicity testing. For each measurement, a freshly prepared biosensor strip was used for studying the toxic effect of different heavy metals (30 μM for each) on the GOx enzyme activity. The results showed that the enzyme active site was completely blocked by the action of copper (Cu²⁺), silver (Ag⁺) and ferrous (Fe²⁺) whereas, cadmium (Cd²⁺) and aluminum (Al³⁺) demonstrated a slightly toxic effect on the GOx enzyme activity. Thus, the proposed biosensor might be used for evaluating the enzyme activity under the environmental or chemical toxic effects.

Applying the developed biosensor for blood samples. The proposed biosensor was used for the quantitative analysis of glucose concentration in patient's blood samples. The detection has been performed using the developed biosensor in comparison to a reference method without any pretreatments for the blood samples. The glucose concentration results obtained from the analysis of two blood samples using the developed biosensor were found to be 285 mg mL⁻¹ and 290 mg mL⁻¹, respectively with a 5–7.3% difference from that obtained by the reference method.

Conclusion

Nanostructures-based electrochemical biosensors provide powerful platforms for constructing reliable and portable biosensing devices. Thus, in the current study, the use of SPE for entrapping of a catalytically active enzyme (GOx) on the new nanocomposite (MnO₂–HBCs NPs) was performed. The obtained rapid and high efficiency direct electron transfer from the redox active centre of immobilized enzyme and screen printed electrode is a promising avenue for fabricating further excellent electrochemical biosensors that could be useful for monitoring infectious disease, detecting cancer, analyzing disease biomarkers or monitoring the pharmacokinetics of drugs.

Acknowledgements

Dr Rabeay Hassan is grateful for the group leader of Biological Systems Analysis (Prof. Dr Ursula Bilitewski, Helmholtz Centre for Infection Research, HZI, Braunschweig, Germany) for presenting the potentiostat (Gamry Potentiostat/Galvanostat/ZRA G750).

References

1. S. D. Varfolomeev, I. N. Kurochkin and A. I. Yaropolov, Biosens. Bioelectron., 1996, 11, 863–871.
2. B. Prieto-Simin, M. Campas and J. L. Marty, Protein Pept. Lett., 2008, 15, 757–763.
3. A. Sassolas, L. J. Blum and B. D. Leca-Bouvier, Biotechnol. Adv., 2012, 30, 489–511.
4. R. Y. Hassan and U. Wollenberger, Anal. Bioanal. Chem., 2016, 408, 579–587.
5. S. H. Yeom, B. H. Kang, K. J. Kim and S. W. Kang, Front. Biosci., Landmark Ed., 2011, 16, 997–1023.
6. J. Ge, J. Lei and R. N. Zare, Nat. Nanotechnol., 2012, 7, 428–432.
7. F. Lyu, Y. Zhang, R. N. Zare, J. Ge and Z. Liu, Nano Lett., 2014, 14, 5761–5765.
8. P. Si, P. Kannan, L. Guo, H. Son and D. H. Kim, Biosens. Bioelectron., 2011, 26, 3845–3851.
9. R. S. Dey and C. R. Raj, J. Phys. Chem. C, 2010, 114, 21427–21433.
10. H. Zhou, H. Chen, S. Luo, J. Chen, W. Wei and Y. Kuang, Biosens. Bioelectron., 2005, 20, 1305–1311.
11. X. Ren, X. Meng, D. Chen, F. Tang and J. Jiao, Biosens. Bioelectron., 2005, 21, 433–437.
12. H. Sun, T. S. Choy, D. R. Zhu, W. C. Yam and Y. S. Fung, Biosens. Bioelectron., 2009, 24, 1405–1410.
13. A. Umar, M. M. Rahman, A. Al-Hajry and Y. B. Hahn, Talanta, 2009, 78, 284–289.
14. N. Batra, M. Tomar and V. Gupta, Biosens. Bioelectron., 2015, 67, 263–271.
15. S. Komathi, N. Muthuchamy, K. P. Lee and A. I. Gopalan, Biosens. Bioelectron., 2016, 84, 64–71.
16. K.-Y. Hwa and B. Subramani, Biosens. Bioelectron., 2014, 62, 127–133.
17. P. R. Solanki, A. Kaushik, V. V. Agrawal and B. D. Malhotra, NPG Asia Mater., 2011, 3, 17–24.
18. H. Zhao, X. Ji, B. Wang, N. Wang, X. Li, R. Ni and J. Ren, Biosens. Bioelectron., 2014, 65, 23–30.
19. W. Lian, S. Liu, J. Yu, X. Xing, J. Li, M. Cui and J. Huang, Biosens. Bioelectron., 2012, 38, 163–169.
20. C. Chen, L. Wang, Y. Tan, C. Qin, F. Xie, Y. Fu, Q. Xie, J. Chen and S. Yao, Biosens. Bioelectron., 2011, 26, 2311–2316.
21. M. Sedki, A. Hefnawy, R. Y. A. Hassan and I. M. El-Sherbiny, Int. J. Biol. Macromol., 2016, 93, 543–546.
22. B. Yin, R. Yuan, Y. Chai, S. Chen, S. Cao, Y. Xu and P. Fu, Biotechnol. Lett., 2008, 30, 317–322.
23. X. Wei, M. Zhang and W. Gorski, Anal. Chem., 2003, 75, 2060–2064.
24. O. Yilmaz, D. O. Demirkol, S. Gulcemal, A. Kilinc, S. Timur and B. Cetinkaya, Colloids Surf., B, 2012, 100, 62–68.
25. W. Zhang, X. Li, R. Zou, H. Wu, H. Shi, S. Yu and Y. Liu, Sci. Rep., 2015, 5, 11129.
26. H. D. Jirimali, R. K. Nagarale, J. M. Lee, D. Saravanakumar and W. Shin, ChemPhysChem, 2013, 14, 2232–2236.
27. X. Lu, Z. Wen and J. Li, Biomaterials, 2006, 27, 5740–5747.
28. S. De, S. Mohanty and S. K. Nayak, Bioprocess Biosyst. Eng., 2015, 38, 1671–1683.
29. I. M. El-Sherbiny, A. Hefnawy and E. Salih, Int. J. Biol. Macromol., 2016, 86, 782–788.
30. D. A. Tomalia, H. Baker, J. Dewald, M. Hall, G. Kallos, S. Martin, J. Roeck, J. Ryder and P. Smith, Polym. J., 1985, 17, 117–132.
31. E. Salih, M. Mekawy, R. Y. A. Hassan and I. M. El-Sherbiny, J. Nanostruct. Chem., 2016, 1–8.
32. E.-H. Yoo and S.-Y. Lee, Sensors, 2010, 10, 4558–4576.

---