Highly sensitive non-enzymatic electrochemical glucose sensor based on dumbbell-shaped double-shelled hollow nanoporous CuO/ZnO microstructures

A high-performance non-enzymatic glucose sensor based on hybrid metal-oxides is proposed. Dumbbell-shaped double-shelled hollow nanoporous CuO/ZnO microstructures (CuO/ZnO-DSDSHNM) were prepared via the hydrothermal method using pluronic F-127 as a surfactant. This structure is studied by various physicochemical characterizations such as scanning electron microscopy, X-ray diffraction spectroscopy, inductively coupled plasma atomic emission spectroscopy, elemental mapping techniques, X-ray photoelectron spectroscopy, and transmission electron microscopy. This unique CuO/ZnO-DSDSHNM provides both a large surface area and an easy penetrable structure facilitating improved electrochemical reactivity toward glucose oxidation. The prepared CuO/ZnO-DSDSHNM was used over the glassy carbon electrode (GCE) as the active material for glucose detection and then coated by Nafion to provide the proposed Nafion/CuO/ZnO-DSDSHNM/GCE. The fabricated glucose sensor exhibits an extremely wide dynamic range from 500 nM to 100 mM, a sensitivity of 1536.80 µA mM−1 cm−2, a low limit of detection of 357.5 nM, and a short response time of 1.60 s. The proposed sensor also showed long-term stability, good reproducibility, favorable repeatability, excellent selectivity, and satisfactory applicability for glucose detection in human serum samples. The achieved high-performance glucose sensing based on Nafion/CuO/ZnO-DSDSHNM/GCE shows that both the material synthesis and the sensor fabrication methods have been promising and they can be used in future researches.


Scientific Reports
| (2021) 11:344 | https://doi.org/10.1038/s41598-020-79460-2 www.nature.com/scientificreports/ synthesized CuO/ZnO-DSDSHNM is shown in Fig. 2a,b. As the main morphology of the material particles, the dumbbell shape structures are clearly observed. In many cases, the size of the heads at each side of the dumbbell is not equal (asymmetrical dumbbell). The hollow nature of the synthesized structure is signified at the inset in Fig. 2a. The hollow nature of these particles could be directly related to the presence of carbon microspheres as a hard template. During the hydrothermal process, the carbon microspheres (Formed in situ from glucose precursor) served as the template for the fabrication of the final hollow multi-shelled nanostructure. The removal of the template during the calcination step under the air atmosphere is responsible for the formation of a hollow multi-shelled nanostructure by the strategy of "hard template synthesis". Using the oxygen-containing air as the calcination atmosphere causes the conversion of the template to CO and facilitates the removal of the carbon template. According to the SEM micrograph at the inset of Fig. 2b, the nanostructured details of each particle could be observed. In addition, the integration of nanoparticles to form the final micron-sized particles has rendered the nanoporous nature of the synthesized particles which is also clear in the SEM micrograph. The formation of the nanoporous surfaces could be attributed to the presence of pluronic F-127 as a surfactant in the synthesis procedure. The holes over some particles are remarked by arrows in Fig. 2a which implies the hollow nature of these particles. Further observation utilizing transmission electron microscopy (TEM) confirmed the results obtained from the analysis of the SEM images. The TEM images indicate the double-shelled structures of CuO/ZnO. As can be seen in Fig. 2c,d, the dumbbell-shaped CuO/ZnO structures are composed of two layers. The inner shell diameter is 170.74 nm and the outer shelled diameter is 393.63 nm (Fig. 2d). Figure 3a,b represent the SEM images of separate CuO and ZnO microspheres (CuO-MS and ZnO-MS), respectively. According to the SEM micrograph of the CuO microstructures, the formation of bean-like particles with an average diameter of 1.50 µm from small nanoparticulate blocks could be observed (Fig. 3a) 44 . In the case of ZnO microstructures ( Fig. 3b) it has been shown that they have an average size of 1.50 µm. Also, the nanoparticulate, as well as the hollow nature of the synthesized structures, has been observed. The X-ray diffraction spectroscopy (XRD) of CuO-MS, ZnO-MS, and CuO/ZnO-DSDSHNM are represented in Fig. 3C   www.nature.com/scientificreports/ The inductively coupled plasma atomic emission spectroscopy (ICP-AES) showed that the atomic ratio of Cu over Zn is 1.4 which shows a very close distribution. In addition, the elemental mapping of the CuO/ZnO-DSDSHNM is depicted in Fig. 4. The uniform distribution of Cu (red), Zn (green), and O (blue) over the entire sample could be deduced. It can be seen that the ICP results are in almost good agreement with the elemental mapping results. Conclusively, the homogeneous structure of particles as a composite of CuO and ZnO materials has been verified.
X-ray photoelectron spectroscopy (XPS) of the samples was studied to further investigate the chemical state of the composing elements and also the surface composition of the CuO/ZnO-DSDSHNM samples. The XPS    Fig. 5c, high-resolution Zn 2p spectra of CuO/ZnO-DSDSHNM showed two symmetric peaks. The presence of the Zn in the matrix (in + 2 oxidation state) can be concluded from the peaks at ∼1020.08 eV (Zn 2p 3/2 ) and ∼1043.28 eV (Zn 2p 1/2 ) 48 . Since the O 1 s peak is asymmetric (Fig. 5d), it can be concluded that the surface is comprised of two components of the oxygen. One of the components located at ∼528.88 eV is attributed to the O 2− ions bonded with Zn 2+ or Cu + /Cu 2+ ions, whereas the second one which is located at ∼530.08 eV belongs to the O 2− ions in oxygen-deficient regions. The obtained XPS results verify that the oxygen vacancy defects are present on the surface of the synthesized samples 49−51 .
Electrochemical properties. The GCE has been modified with the synthesized hollow structures and the Nafion/CuO/ZnO-DSDSHNM/GCE was used as a modified working electrode for measurements. The electrochemical CV tests for different working electrodes were performed in 0.50 M NaOH solution at the scan rate of 50 mVs −1 within the potential range of 0 to + 0.80 V. The CVs shown in Fig. 6a illustrate the responses of the bare and modified GCE in the absence and presence of 5 mM glucose. The CVs shown in Fig. 6a illustrates the responses of the bare and modified GCE in the absence and presence of 5 mM glucose. As shown in Fig. 6a, in the absence of glucose, no response can be observed for bare GCE and Nafion/CuO/ZnO-DSDSHNM/GCE. However, when glucose is added, a distinguishable response was revealed, which demonstrated the reasonable performance of CuO/ZnO-DSDSHNM toward glucose detection. The CVs shown in Fig. 6a illustrates the responses of the bare and modified GCE in the absence and presence of 5 mM glucose. As shown in Fig. 6a, in the absence of glucose, no response can be observed for bare GCE and Nafion/CuO/ZnO-DSDSHNM/GCE. However, when glucose is added, a distinguishable response was revealed, which demonstrated the reasonable performance of CuO/ZnO-DSDSHNM toward glucose detection. The CV responses for the CuO/ZnO-DSDSHNM modified GCE to different glucose concentrations have been presented in Fig. 6b. As can be seen, by increasing the glucose concentration from 0 to 25 mM the current has also been increased. As Fig. 6b demonstrates, there are remarkable differences between the CVs which shows that the fabricated sensor is very sensitive to the changes in the glucose concentration and the sensor is reacting very sensitively to the increase in the glucose concentration. In  Fig. 6c reveals that the CuO response is much larger than ZnO. Although ZnO has no significant response to the glucose in the non-enzymatic detection process, however, it has a strong contribution to the enhancement of electrochemical properties of the sensor when it combines with CuO and the proposed hybrid material (CuO/ZnO-DSDSHNM) is formed 6,28 . The integration of metal oxide semiconductors of p-type and n-type enhances the concentration of charge carriers due to better charge separation. In addition, due to the heterogeneous function, an increase in the stability and catalytic activity is achievable 26,27 . As a result of Fig. 6c, the integration of these two semiconductors provides a more favorable electrochemical activity and outstanding responses in comparison with the Nafion/CuO-MS/GCE electrode.  According to Fig. 6, the voltammetry peak related to the conversion of Cu(2 +) to Cu(3 +) was not observable at CuO/ZnO-DSDSHNM in alkaline electrolyte. Instead, also in the absence of glucose, a plateau around 0.5-0.7 V vs. Ag/AgCl was observable. The capacitive nature of this plateau has been approved previously 52 which could be attributed to the adsorption of hydroxyl ions in the electrolyte to the surface of CuO/ZnO electrode as a p-type semiconductor. The sharp increase of the current after this plateau also could be attributed to the oxidation of these adsorbed hydroxyl species to O 2 in the high anodic potentials. Following the presence of glucose in the electrolyte, it seems that this electrochemical O 2 evolution is amplified and the enhancement of plateau is remarkable. Although the well-shaped peak of glucose oxidation over CuO/ZnO-DSDSHNM was not apparent, but the gradual increase of plateau current by successive addition of glucose concentration was encountered (Fig. 6b). Herein, for better distinguishing of the response to various concentrations of glucose, the chronoamperometric technique was chosen instead of cyclic voltammetry. In addition to the above-mentioned electrochemical evidences, the presence of O vacancies over the electrode surface was previously approved by using XPS analysis (Fig. 5).
Due to the application of the unique semiconductive hybrid nanostructures of CuO/ZnO in the construction of CuO/ZnO-DSDSHNM electrode, the adsorption of OH − species from alkaline electrolyte occurred remarkably. In the absence of glucose, the further oxidation of these adsorbed species caused the sharp current increase at the extreme of the anodic sweep. However, in the presence of glucose, some interactions of the analyte with the electrode surface was expectable 52 . According to interactions of glucose with the electrode surface, the amplification of the oxidation process of OH − species occurred and in turn, the current has been increased. The increase of current was linearly related to the glucose concentration and this has been successfully used for sensing application.
According to Fig. 6c, the ZnO-MS modified electrode has no response to the glucose in the non-enzymatic detection process which has no specific property of redox reaction like CuO, however, the role of ZnO in the proposed sensor is to accelerate the electron transfer, which demonstrates the high performance of CuO/ZnO-DSDSHNM as evidenced in Fig. 6c.
To survey the kinetics of glucose oxidation on the surface of the Nafion/CuO/ZnO-DSDSHNM/GCE, the effect of the scan rates on the CV currents has been studied for 10 mM of glucose in 0.50 M NaOH solution. Figure 7 shows the CV responses of glucose oxidation at Nafion/CuO/ZnO-DSDSHNM/GCE at different potential scan rates ranging from 10 to 250 mVs −1 . As illustrated at the inset (a) of Fig. 7, the peak currents exhibited a linear relation versus the square root of the scan rate, demonstrating that the oxidation of glucose at the modified electrode is under the diffusion-controlled behavior. In contrast, as depicted in the inset (b) of Fig. 7, the plot of peak currents versus the scan rate is not favorably linear which indicates the absence of adsorptive behavior 30,31,44 . Sensor optimization. The working electrode surface should be modified with CuO/ZnO-DSDSHNM. To enhance the performance of the sensor, it is necessary to cast an optimum amount of CuO/ZnO-DSDSHNM on the GCE surface. Optimizing this parameter will increase the accuracy and sensitivity of the sensor. In order to determine the best concentration of the CuO/ZnO-DSDSHNM, the CV tests were performed at different concentrations of the sample. According to figure S1, the optimum value of this parameter has been selected 5 mg/ ml. After that, a coating of Nafion is casted on the surface of CuO/ZnO-DSDSHNM. This coating prevents the falling of nanoparticles from the GCE surface and prevents interfering species from penetrating. It also increases the reproducibility and repeatability of the sensor. Therefore, as another step for optimization, it is necessary to optimize the concentration of the Nafion. Different concentrations of Nafion were prepared and CV tests were  Figure S2 shows the results of the different Nafion concentrations and accordingly, it is obvious that the optimum concentration is 0.50 wt%. Because the GCE and the CuO/ ZnO-DSDSHNM will be oxidized in the absence of the analyte in NaOH solution, several consecutive cycles of the CV measurements have been performed. These cycles were repeated in the absence of glucose until they are overlapped. This way we are sure that by adding glucose, the response of the electrode is only due to the glucose oxidation. The measurements showed that after 10 cycles the proposed sensor behaves in exactly the same way as the last cycle. Thus, to stabilize the results, 10 cycles of CVs must be run before the test starts. After the electrode activation, the best voltage range for improved sensing abilities should be selected. This is one of the most important parameters that is very effective in the performance of the sensor. Based on figure S3, the best results for modified GCE were observed in the range of 0-0.80 V. Optimizing the NaOH concentration results in the maximum production of OH − , in which case the oxidation of glucose will be as efficient as possible. According to figure S4, the linear increasing trend of current versus glucose concentration was observed in 0.50 M NaOH solution which also resulted in a higher current. As another design parameter, the best voltage of the glucose oxidation for the amperometry test has to be determined which enables the reliable performance of the sensor. In order to do that, the current-time (i-t) curves of Nafion/CuO/ZnO-DSDSHNM/GCE with the addition of 5 mM glucose in 0.5 M NaOH solution were acquired at different oxidation voltages ( Figure S5). According to figure S5, the highest amount of glucose oxidation occurs at a voltage of 0.60 V.

Non-enzymatic glucose sensing based on CuO/ZnO-DSDSHNM. For further highlighting and
confirming the superior electro-catalytic activity of CuO/ZnO-DSDSHNM, amperometry test was carried out. Figure 8a demonstrates the typical i-t curve of CuO/ZnO-DSDSHNM with successive addition of different glucose concentrations at an optimum applied potential of + 0.60 V under the stirring condition in 0.5 M NaOH solution. By successive additions of various glucose concentrations from 500 nM to 100 mM the consecutive increase in current has been observed. In addition, the inset of Fig. 8a illustrates the magnified view of the low concentration ranges of glucose. Obviously, the first current steps appeared for 500 nM glucose, suggesting the high detection sensitivity of the sensor. The corresponding calibration curve can be described by a power plot ( Figure S6) which could be broken into three linear ranges with the following regression equation for the first linear range: I(μA) = (0.04828 ± 0.0025) C(μM) + (1.1289 ± 0.0129) (R 2 = 0.9917) for 0.5 µM-2 mM glucose (Fig. 8b). The sensitivity can be calculated as the ratio of the slope to the electrode area. The novel electrochemical glucose biosensor was designed, constructed, and optimized so that it has a wide dynamic range from 500 nM to 100 mM which included three linear ranges 500 nM-2 mM, 2 mM-20 mM, and 20 mM-100 mM, and an outstanding sensitivity of 1536.80 ± 79.60 µA mM −1 cm −2 . Upon the addition of 500 nM of glucose, the Nafion/ CuO/ZnO-DSDSHNM/GCE reached a steady-state with a very sharp response. It raised to 90% of the response within 1.60 s which is highly favorable comparing with similar sensors which are brought in Table 1. Also, the limit of detection is calculated as 357.5 nM which shows that the sensor has provided a very low detection limit. This indicates that the proposed sensor has a high capability for fast electrocatalytic detection of glucose.
The analytical performance of the as-prepared CuO/ZnO-DSDSHNM sensing system for glucose detection is superior to or comparable with many previously enzymatic and non-enzymatic glucose sensor 4,30,31,54 (Table 1).
Reproducibility, repeatability, selectivity, and stability tests. Reproducibility, repeatability, selectivity, and stability are critical parameters for the investigation of the performance of sensing devices. Repeatability means the absence of incompatibility between consecutive measurements by the same electrode, whilst reproducibility is the vicinity of the results acquired by means of a number of the same modified electrodes with the same measurement procedure. The CV responses of three different electrodes with successive addition of 5 mM glucose were recorded at the optimal condition. The relative standard deviation (RSD) of the response currents for different CuO/ZnO-DSDSHNM electrodes was only 1.3%, suggesting good reproducibility and precision (Fig. 9a). The repeatability was examined by three times of measurements by a single modified GCE www.nature.com/scientificreports/ in 1 day. The RSD of the repeatability was calculated as 2.23% as shown in Fig. 9b, which shows the excellent repeatability of the proposed sensor. The selectivity of CuO/ZnO-DSDSHNM for glucose sensing was evaluated by the addition of some potential interferences when conducting a typical amperometric detection. As shown in Fig. 9c, an obvious steady-state current appeared upon addition of 5 mM glucose at 0.6 V. In addition, no current change was observed with the additions of 0. The stability of the proposed electrochemical sensor was evaluated by storing the Nafion/CuO/ZnO-DSDSHNM modified GCE electrodes in air and recording the current response of 5 mM glucose. Figure 9d showed that the response current of 5 mM glucose could retain about 92.88% of its initial value after storing Nafion/CuO/ZnO-DSDSHNM modified GCEs in the air for 15 days, indicating excellent storage stability.
Glucose detection in real human blood. Analysis of non-diabetic human blood serum samples was performed to evaluate the commercial viability of the developed non-enzymatic Nafion/CuO/ZnO-DSDSHNM/ GCE glucose sensor. Additionally, these real blood samples were analyzed using the DIRUI CS-800 Auto Chemistry Analyzer device for validating the results. The amperometric analysis was performed by injection of 30 µl www.nature.com/scientificreports/ of the sample into the 30 ml of 0.50 M NaOH at the potential of + 0.60 V (versus Ag/AgCl). The analysis was performed via the standard addition method. The same aliquots of serum samples were spiked by the known concentrations of glucose and 30 µl of each sample was injected into the 30 ml of NaOH solution. The electrochemical analysis of each sample was done by using Nafion/CuO/ZnO-DSDSHNM/GCE three times. Anywhere which was necessary, the volume correction was performed. The Nafion/CuO/ZnO-DSDSHNM/GCE glucose sensor electrode provided an acceptable recovery ranging from 95.84 to 104.83% for the human blood serum samples as shown in Table 2.

Conclusion
A new double-shelled hollow nanoporous CuO/ZnO microstructure was successfully fabricated by a hydrothermal method, designed for non-enzymatic electrochemical detection of glucose. As a result, an efficient and rapid electrochemical glucose biosensor was fabricated and optimized. The results revealed that the sensor optimizations performed have been successful and the proposed sensor showed a wide dynamic detection range of 500 nM-100 mM, remarkable sensitivity of 1536.8 ± 79.6 µA mM −1 cm −2 , and very low LOD of 357.5 nM. The Nafion/CuO/ZnO-DSDSHNM modified GCE exhibited excellent reproducibility and repeatability, favorable stability, and satisfactory selectivity toward glucose sensing in a 0.5 M NaOH solution. The cost-effective synthesis procedure and the outstanding catalytic performance suggest that the CuO/ZnO-DSDSHNM material may be promising for non-enzymatic glucose sensing in practical analysis applications.

Materials. Zinc Nitrate (Zn(NO 3 ) 2 ), Copper Nitrate (Cu(NO 3 ) 2 ), Sodium Hydroxide (NaOH), Aluminum
Oxide (Al 2 O 3 ), AA, DA, lactose, sucrose, NaCl, Nafion, glucose, and pluronic F-127 were purchased from Sigma-Aldrich, Merck and DuPont and human blood serum was taken from the hospital. In this study, all the materials were of analytical grade and used as received without any further purification. All solutions were prepared by DI water (3 μS/cm) generated from the Deltino water purifying system.
Apparatus. XRD was taken by BRUKER D8 Advanced. SEM and elemental mapping data were obtained by using TESCAN Mira3. The TEM images were taken with digital charge-coupled device JEOL-JEM-2010 equipped with a CCD camera. XPS experiments were measured using an ESCAlab 250 Analytical XPL Spectrometer with a monochromatic Al Ka source. ICP-AES test was performed by ICP-MS ELANDRC-E Perkin-Elmer.The Electrochemical measurements were performed by using IVIUM Vertex potentiostat/galvanostat electrochemical analyzer. All experiments were conducted using a three-electrode system with a glassy carbonbased working electrode (2 mm diameter), an Ag/AgCl reference electrode (3 M KCl), and a platinum wire for the counter electrode (1 cm length and 1 mm diameter). The CVs and choronoamperometric measurements were performed with an electrochemical analyzer in a 0.50 M NaOH solution.

Synthesis of materials.
Three types of hollow structured materials have been synthesized. As the main material, a unique dumbbell-shaped double-shelled hollow nanoporous CuO/ZnO microstructures was synthesized. For comparison of electrocatalytic activity, the CuO-MS and ZnO-MS also were synthesized via a similar method. The procedure for synthesizing the above-mentioned materials are as follows: First, 3 g of glucose, 1 g of F-127 as a surfactant were dissolved in 150 ml deionized water and the solution was sonicated for 10 min. Then, 2 ml of 1 M of Zn(NO 3 ) 2 solution and 2 ml of 1 M of Cu(NO 3 ) 2 solution were added to the mixture. The mixture of glucose, surfactant and both metal salts was placed into the autoclave and the hydrothermal process was directed for 24 h at 180 ֯ C. The resulted product was washed three times with distilled water and dried in an oven at 80 ֯ C overnight. The final structure was achieved via the calcination of the product from the hydrothermal step at 550 ֯ C for 270 min.
The synthesis of CuO-MS and ZnO-MS was achieved via the same procedure with the same amount of reagents but by using the single salt of each metal at processes.
Electrode preparation (Nafion/CuO/ZnO-DSDSHNM/GCE). At first step, the GCE was polished by means of alumina slurry over a Pad, and washed three times with DI water and ethanol, respectively. To fabricate the Nafion/CuO/ZnO-DSDSHNM/GCE, the solution of 5 mg/ml of synthesized material in ethanol was prepared and ultrasonically homogenized for 20 min. After that, 2 μl of CuO/ZnO-DSDSHNM solution was casted on the surface of GCE until it dried at room temperature. Next, 2 μl of 0.5 wt% Nafion solution was casted on the catalyst layer at GCE and then dried at ambient conditions. Henceforth, the Nafion/CuO/ZnO-DSDSHNM/ GCE electrode was prepared for performing electrochemical measurements.