CdS quantum dots modified CuO inverse opal electrodes for ultrasensitive electrochemical and photoelectrochemical biosensor

The CuO inverse opal photonic crystals (IOPCs) were synthesized by the sol-gel method and modified with CdS quantum dots by successive ionic layer adsorption and reaction (SILAR). CdS QDs modified CuO IOPCs FTO electrodes of different SILAR cycles were fabricated and their electrochemical properties were studied by cyclic voltammetry (CV) and chronoamperometry (I–t). Structure and morphology of the samples were characterized by transmission electron microscopy (TEM), scanning electron microscopy (SEM), high-resolution TEM (HRTEM), Energy-dispersive X-ray analysis (EDX) and X-ray diffraction pattern (XRD). The result indicated that the structure of IOPCs and loading of CdS QDs could greatly improve the electrochemical properties. Three SILAR cycles of CdS QDs sensitization was the optimum condition for preparing electrodes, it exhibited a sensitivity of 4345 μA mM-1 cm-2 to glucose with a 0.15 μM detection limit (S/N= 3) and a linear range from 0.15 μM to 0.5 mM under a working potential of +0.7 V. It also showed strong stability, good reproducibility, excellent selectivity and fast amperometric response. This work provides a promising approach for realizing excellent photoelectrochemical nonenzymatic glucose biosensor of similar composite structure.


Results and Discussion
Material Characterization. The experimental procedure for the preparation of CdS QDs modified CuO IOPCs FTO electrodes consists of four simple steps, as illustrated in Fig. 1a: 1) self-assembly of close-packed polymethyl methacrylate (PMMA) spheres with multilayer structures onto FTO substrates, 2) infiltration of PMMA template with prepared precursor solution of CuO by sol-gel method, 3) removal of the original PMMA template by high temperature annealing to obtain CuO IOPCs, and 4) final sensitization of the CuO IOPCs with CdS QDs using SILAR route.
Characterization of the surface morphology and microstructure of the synthesized CuO IOPCs were performed by field emission scanning electron microscope (SEM) and transmission electron microscopy (TEM). The SEM images of the synthesized CuO IOPCs (Fig. 2a,b) show a long-range ordered hexagonal close-packed arrangement of the inverse opal structure in three-dimensional space. Two variants were observed, with diameters of 378 nm and 275 nm, and they were named as CuO IOPCs S1 and CuO IOPCs S2 respectively. As shown in Fig. 2c, the cross-sectional SEM images of CuO IOPCs also displayed a vertical ordering with face-centered-cubic (fcc) packing structure. In order to investigate further into the detailed structure and morphology of CuO IOPCs, high-magnification TEM images were obtained. The TEM image in Fig.2d illustrates that the wall thickness of the CuO IOPCs is about 45 nm, and shows that the structure consists of a large amount of small nanoparticles (NPs) and neighboring pores that are connected to each other by the thin wall. The small wall thickness of the CuO IOPCs structure is highly conductive to the unidirectional transmission of electrons. The TEM image in Fig. 2e shows that the inner and outer surfaces of the CuO IOPCs are well covered with small CdS quantum dots after 2 SILAR cycles. This coverage amount indicates that the Cd 2+ and S 2ions are able to easily diffuse into the pores of the IOPCs and form CdS nanocrystallites. Good crystalline structures were observed for both the CuO and CdS materials, from the high-resolution TEM (HRTEM) image (Fig. 2f). The thickness of the CdS layer was found to be around 5 nm. Interplanar distances of the fringes of the CuO and CdS QDs are 0.275 nm and 0.26 nm, which correspond to the d spacing of (110) planes of monoclinic CuO and the (102) planes of hexagonal CdS, respectively. To analyze the elemental components of the compound, Energy-dispersive X-ray analysis (EDX) was performed, and it was found that the compound consists of Cd and S, besides Cu and O (Fig. 3a), as expected. The lattice information and EDX analysis results confirm that the CdS QDs modified CuO IOPCs composites were successfully synthesized.
In order to assess the CdS QDs deposition and growth process on the CuO IOPCs, the TEM images and absorption spectra were monitored at different stages of SILAR route. Fig. 2g,h represent the TEM images of CdS QDs modified CuO IOPCs after 3 and 5 SILAR cycles, respectively. It can be seen from Fig. 2g that the structure of the CuO IOPCs remains intact and the CdS QDs cover the skeleton surface of the IOPCs after 3 SILAR cycles. With increasing number of SILAR cycles, the CdS QDs were deposited not only on the surface of the skeleton but also within the pores of the CuO IOPCs structure (Fig. 2h). The excess CdS QDs deposited in the pores could decrease the surface area and affect the contact of CuO with electrolyte. As seen from the absorption spectra (Fig. 3b), the absorption intensity increases with higher number of SILAR cycles. Also, the edge of the absorption band has a slight redshift from 457 nm to 482 nm ( Figure S1), which can be attributed to the increased size and decreased bandgap of QDs 43 .
The phase properties of the synthesized CuO IOPCs were investigated by X-ray powder diffraction (XRD) experiments. As shown in Fig. 3c, the XRD patterns of the CuO IOPCs matched well to the hexagonal CuO phase (JCPDS, card no: 48-1548), which suggests that the monoclinic CuO phase was formed during the process of heat treatment. No characteristic impurity peaks were observed, indicating a pure monoclinic CuO phase in the samples.

Electrochemical Properties of CuO IOPCs and CdS QDs modified CuO IOPCs FTO Electrode.
The key parameters of the electrochemical kinetics of CuO IOPCs were obtained from cyclic voltammetry (CV) experiments performed in 0.1 M NaOH solution at a scan rate of 100 mV s -1 . As seen from the CV curves in Fig. 3d, oxidation and reduction peaks were obtained at potentials of approximately 0.66 V and 0.7 V respectively, versus a saturated calomel electrode reference electrode. These peaks are believed to correspond to the Cu(II)/Cu(III) redox couple, which is the key component for non-enzymatic electrochemical glucose detection, according to previous studies 17 . The corresponding reaction processes are,  The effect of scan rate on the electrochemical response of CdS QDs/CuO IOPCs modified FTO electrode was investigated, and the results are shown in Fig. 3e. The measurement was performed in 0.1 M NaOH solution at different scan rates of 10-100 mV/s after 3 mM glucose injection. The redox peak currents show a linear response to the scan rate, indicating that the electron transfer process in the CuO redox reaction takes place through a surface-controlled mechanism 44 . Amperometric Detection of the Electrochemical and Photoelectrochemical biosensor. Current− time (I− t) curves were obtained to determine the amperometric sensing property of CuO IOPCs and CdS QDs modified CuO IOPCs FTO electrodes. All measurements were taken in 0.1 M NaOH solution at an applied potential of + 0.7 V. Figure 4a displays the typical amperometric responses of CuO IOPCs S1 photoelectrochemical sensor with successive addition of glucose, both with and without Xe-lamp illumination. The amperometric response indicates that the CuO IOPCs modified FTO electrodes have a fast electrochemical response of less than 5 s to the glucose. Figure 4b shows the corresponding calibration curves. Linear relationships were obtained between the current response and glucose concentration, both with and without Xe-lamp illumination. From the slope of the lines, sensitivity values were determined. The sensitivity was found to be 4065 μ A mM -1 cm -2 with Xe-lamp illumination, which was about 1.2 times higher than the sensitivity without Xe-lamp illumination (3417 μ A mM -1 cm -2 ). This difference indicates that Xe-lamp illumination improves the sensitivity of the CuO IOPCs electrodes.
The enhancement of sensitivity in the presence of Xe-lamp illumination can be attributed to the following two reasons. First, since the CuO electronic band gap is around 1.34 eV, corresponding to a wavelength of around 925 nm, the presence of light with wavelength less than 925 nm (higher energy) can cause the electrons to absorb photons and move from the valance band to the conduction band. Thus, when the electrodes are illuminated with Xe-lamp, more number of electrons are generated on the surface of the CuO IOPCs, which then participate in the redox reaction, as shown in equation 1. Second, the generation of a new electrical field inside the CuO nanoparticles due to the Xe-lamp illumination, inhibits the reverse direction recombination of electrons and holes, thus leading to improved sensitivity.
In order to demonstrate the advantages of the IOPCs structure, a reference CuO thin film FTO electrode was prepared by a spin-coating method from the same precursor solution of CuO as that used for preparing the Cuo IOPCs, and then annealed under the same condition. The electrochemical response calibration curves of the resulting reference thin film electrode are displayed in Figure S2. The linear response range of the reference thin film electrode was found to be up to 500 μ M, while the linear response range of the IOPCs electrode was up to 1000 μ M. This difference is likely due to the larger surface area of IOPCs, which allows more CuO nanoparticles to come in contact with the glucose solution before reaching saturation. The thickness of the CuO IOPCs and thin film electrodes were found to be 26 μ m and 70 μ m, respectively, as shown in Figure S3a and S3b. Moreover, the large void volume in the IOPCs structure (74%) makes the IOPCs lower in quality compared to the thin film electrode. However, the sensitivity of the reference thin film electrode was 1920 μ A mM -1 cm -2 , which is significantly lower than that of the IOPCs electrode (3417 μ A mM -1 cm -2 ) without Xe-lamp illumination. There are two reasons for this observed difference: lower surface area of the reference thin film electrode and large reverse direction recombination current due to long electronic transmission distance. When illuminated with Xe-lamp, the sensitivity of the reference thin film electrode increased to 3182 μ A mM -1 cm -2 , which was 1.66 times than the sensitivity of the reference thin film electrode in dark condition due to the same reason of IOPCs electrodes with Xe-lamp illumination.
As seen from Fig. 4c, the sensitivity of CuO IOPCs S2 was 2926 μ A mM -1 cm -2 without Xe-lamp illumination and 3559 μ A mM -1 cm -2 with Xe-lamp illumination. There is no significant difference between the sensitivity values of CuO IOPCs S1 and S2, thus indicating that the pore size of the IOPCs does not have much influence on the sensitivity, which is consistent with the results of our previous work 34 .
The I− t curves were also measured for CdS QDs modified CuO IOPCs obtained after 3 (Fig. 4d) and 5 (data is not shown) SILAR cycles, and their corresponding calibration curves are shown in Figs. 4e,f. Figs. 4g,h show the calibration curves of amperometric response of three different electrodes with and without Xe-lamp illumination: CuO IOPCs S1, and the two CdS QDs modified CuO IOPCs obtained from 3 and 5 SILAR cycles. The sensitivity values of these three samples were found to be 3417 μ A mM -1 cm -2 , 3944 μ A mM -1 cm -2 and 3300 μ A mM -1 cm -2 in the absence of Xe lamp illumination while the sensitivity values were increased to 4065 μ A mM -1 cm -2 , 4345 μ A mM -1 cm -2 and 3476 μ A mM -1 cm -2 with Xe-lamp illumination. These results indicate that whether or not the Xe-lamp illumination is used, the sensitivity is increased to the highest value after 3 SILAR cycles of CdS QDs sensitization. On the other hand after 5 SILAR cycles of CdS QDs sensitization, the sensitivity was decreased to lower values than that of the original CuO IOPCs electrode.
The increase in sensitivity after 3 SILAR cycles of CdS QDs sensitization can be attributed to two aspects. Firstly, many surface defects are created during the synthesis of the CuO IOPCs, and these defects could trap unpaired electrons and holes, leading to a decrease in the current. The CdS QDs within interfaces passivate the surface states of CuO IOPCs and block the CuO surface, thus suppressing the recombination of electrons and holes at the interfaces 45 . Secondly, as seen in Fig. 1b, the conduction band of CdS is considerably higher than that of CuO, so the CdS QDs act as a potential barrier to adjust the electric field and potential distribution at the interface, thus suppressing the dark current 46,47 . Moreover, when using Xe-lamp illumination, the CdS QDs can absorb light from the visible region of the spectrum and generate new electrons on the surface, improving the current response. Although the electrode after 5 SILAR cycles of CdS QDs sensitization has larger absorption range and higher absorption intensity, its sensitivity and linear range have a decrease than the electrode after 3 SILAR cycles of CdS QDs sensitization, which can be attributed to the reduction of effective surface area available to provide contact with the glucose molecules, due to the excess deposition of CdS QDs. Thus it appears that 3 SILAR cycles of CdS QDs sensitization is the optimum condition for preparing electrodes, providing the highest sensitivity value of 4345 μ A mM -1 cm -2 . In addition, the linear range was obtained over a concentration up to 0.5 mM with a detection limit of 0.15 μ M (with signal to noise ratio of 3).
A comparison of CdS QDs modified CuO IOPCs FTO electrodes with some of the other reported metal based non-enzymatic glucose biosensors is presented in Table 1. From the comparison, it can be seen that our new sensor exhibits a good integrative performance characterized by high sensitivity, low detection limit and considerable linear range. In particular, the sensitivity of this sensor is much higher than that of other glucose sensors.
Selectivity, reproducibility and stability. Anti-interference property is an important parameter for glucose biosensors. Since human blood contains many easily oxidized compounds like ascorbic acid (AA) and uric acid (UA), the influence of other oxidizable interferents must be taken into consideration. As shown in Fig. 5a, the addition of interferents, 0.1 mM of AA and 0.1 mM of UA, gives rise to only negligible current changes, while a significant current response is observed with the subsequent addition of 0.1 mM of glucose. The higher selectivity for glucose over the two interferents can be seen from the current ratios. The current ratio of AA to glucose was determined to be 5.68%, while the current ratio of UA to glucose was 8.33%.
In addition, carbohydrate interferents such as fructose and lactose also present in human blood, thus the selectivity to these carbohydrates were also evaluated. Although the concentration of interfering species in the human blood is 30 to 50 times less than that of glucose 48 , much higher concentrations of the interfering species (with glucose to interfering species ratio of 10) were used in this analysis to ensure the sensitivity. As is clearly seen in Fig. 5b, interfering carbohydrate have little effect compared to glucose. Thus the CdS QDs modified CuO IOPCs FTO electrodes were found to exhibit excellent selectivity in the interference experiments conducted.
To evaluate the reproducibility of the new electrodes, five CdS QDs modified CuO IOPCs FTO electrodes were prepared separately under the same conditions. The relative standard deviations (RSDs) were calculated to be no more than 5.8% (data is not shown), indicating a good reproducibility. The long-term stability of the CdS QDs modified CuO IOPCs FTO electrode was evaluated by placing it in 0.1 M NaOH solution at room temperature for a period of time (Fig. 6). It was found that the CdS QDs modified CuO IOPCs FTO electrode retained 92% of its original current response even after a storage period of 2 weeks, indicating good stability.
The combined results described above indicate that our novel CdS QDs modified CuO IOPCs FTO electrodes exhibit good performance as an electrochemical electrode, including high selectivity, reproducibility and stability.

Conclusions
In this work, we have synthesized new CuO IOPCs by the sol-gel method using PMMA as a template, and then modified them with CdS quantum dots using a SILAR approach. A series of CdS QDs modified CuO IOPCs FTO electrodes were fabricated using different numbers of SILAR cycles, all of which displayed many attractive analytical properties such as high sensitivity, good stability, good reproducibility, excellent selectivity, as well as quick response. The sensitivity of CuO IOPCs was found to be 4065 μ A mM -1 cm -2 , which is likely due to the large surface area and uniform porous distribution within the IOPCs structure. Further increase in sensitivity was attributed to the effect of surface passivation and the protection of the potential barrier of CdS QDs. It was found that 3 SILAR cycles of CdS QDs sensitization was the optimum condition for preparing electrodes, which gave high sensitivity of up to 4345 μ A mM -1 cm -2 . Overall, the novel CdS QDs modified CuO IOPCs FTO electrodes demonstrate good performance as non-enzymatic electrochemical and photoelectrochemical glucose biosensors.

Experimental Section
Materials. Cupric    methacrylate (MMA), ethanol, glucose and citric acid were purchased from Beijing Chemical Plant (Beijing, P. R. China). All chemicals were analytical grades and used without further purification.

Synthesis of CuO inverse opal photonic crystals.
CuO IOPCs were synthesized by the sol-gel method using the PMMA latex sphere as the colloidal crystal template. Firstly, the monodispersed PMMA spheres with controllable sizes were synthesized using our previously reported method 49 . Fluorine-doped tin oxide (FTO) glass (sheet resistance 30 Ω /cm 2 ; Qiseguang Glass Co. Ltd., Dalian) was washed ultrasonically in ace-tone, isopropyl alcohol, and deionized water mixed solution and dried, and then the FTO substrate was soaked in the mixed H 2 SO 4 /H 2 O 2 aqueous solution with the volume ratio of 3:1 for 1 hour and washed with distilled water to make its surface hydrophilic. Subsequently, the PMMA template was self-assembled through the vertical deposition process. The hydrophilic FTO substrate was submerged into the prepared PMMA colloidal suspension (5% solid content) and kept at 35 °C in an oven for 2 days. The PMMA colloidal spheres were slowly self-organized into highly ordered colloidal arrays on the FTO substrate, driven by surface tension of the liquid in the evaporating process. Following deposition, the photonic crystals were sintered for 40 min at 120 °C to enhance their physical strength. Cu (NO 3 ) 2 •3H 2 O (2.5 g) was dissolved in 10 mL ethanol solution as the copper ions precursor. Then, an appropriate amount of citric acid was added to the solution as chelating agent. The mixture eventually formed a bright blue and transparent solution after stirring for 1 hour. The prepared precursor solution was used to infiltrate into the interstices of the templates under the capillary force from one side of the PMMA templates. After sufficient infiltration, the templates were dried in air overnight. Then, the samples were heated to 500 °C at a rate of 1 °C/min and kept for 3 hours to remove the original PMMA template.
CdS QDs modified CuO IOPCs FTO electrode. The CuO IOPCs were modified with CdS QDs using a previously reported SILAR route 50,51 . In a typical procedure, the CuO IOPCs were immersed in a 20 mL ethanol solution containing 50 mM cadmium nitrate tetrahydrate (Cd (NO 3 ) 2 •4H 2 O) for 1 min, to allow Cd 2+ to adsorb onto the surface of CuO IOPCs. Then, the electrodes were immersed in ethanol solution and dried in a N 2 stream. The dried electrodes were dipped into a mixture solution of ethanol and water containing 50 mM Sodium sulfide nonahydrate (Na 2 S•9H 2 O) for 1 min, where the adsorbed Cd 2+ reacted with S 2to form the desired CdS. The electrodes were then rinsed in ethanol for 1 min to remove the excess ions and dried again in a N 2 stream. This procedure was repeated several times to obtain the desired thickness of CdS.
Apparatus. The morphology of the samples was inspected using a JEOL JSM-7500F field emission scanning electron microscope (SEM) (Japan) and a JEM-2010 transmission electron microscope (TEM) (Japan) under a working voltage of 200 kV. The high-resolution TEM (HRTEM) images and EDX patterns were measured on a JEOL-2100F high-resolution transmission electron microscope (Japan) with a working voltage of 200 kV. X-ray diffraction (XRD) patterns were conducted on a RigakuD/max 2550 X-ray diffractometer. Electrochemical measurements were performed on a model CHI630D electrochemical analyzer (ChenHua Instruments Co.Ltd., Shanghai, China). Ultraviolet-visible (UV-vis) absorption spectra were measured with a Shimadzu UV-3101PC scanning spectrophotometer. The electrochemical measurements were carried out using a three-electrode system, which employed a platinum wire as the counter electrode, a saturated calomel electrode (SCE) as the reference electrode, and CdS QDs modified CuO IOPCs FTO electrodes as the working electrode.