A SiO2 Hybrid Enzyme-Based Biosensor with Enhanced Electrochemical Stability for Accuracy Detection of Glucose

A novel enzyme-based biosensor for glucose detection is successfully developed using layer-by-layer assembly technology. The introduction of commercially available SiO2 was found to be a facile way to improve overall electrochemical stability. After 30 CV cycles, the proposed biosensor could retain 95% of its original current. The biosensor presents good detection stability and reproducibility with the detection concentration range of 1.96 × 10−9 to 7.24 × 10−7 M. This study demonstrated that the hybridization of cheap inorganic nanoparticles was a useful method in preparing high-performance biosensors with a much lower cost.


Introduction
Diabetes mellitus is the most prevalent chronic disease, which always comes with abnormally high blood glucose levels [1]. On-time monitoring of the glucose level in the blood and body fuids could present useful information in the diabetes mellitus treatment. But the complex chemical and biological environment in the body fuids and blood presents a huge challenge for accurate direct detection of glucose. Since the frst report of glucose biosensors by Clark and Lyons in 1962, enzyme-based biosensors have been widely studied and developed [2]. Electrochemical biosensors possess numerous advantages in the rapid and accurate detection. Also, the co-use of enzymatic reactions further improves the selectivity of the biosensor [3].
Glucose oxidase (GOx) is the most commonly used enzyme in glucose detection due to its high selectivity and rapid response. For GOx-based biosensor, the loading amount of GOx plays an important role in the detection accuracy and sensitivity. However, the good water solubility of GOx might lead to the leaching of GOx during electrochemical detection and further lead to poor performance in detection accuracy and reproducibility [4]. Tus, attempts have been made to achieve efective GOx immobilization by the introduction of an electrode matrix. Several materials have been explored for the efcient loading of the GOx, including carbon nanotubes (CNTs), polyaniline, graphene, and metal nanoparticles [5][6][7][8]. Among these, CNTs show great potential for enzyme loading due to their large surface area. However, despite the good conductivity of the CNTs themselves, the GOx immobilized electrode matrix still sufers from poor conductivity, which results in poor electrochemical properties [9]. Tus, materials, such as gold nanoparticles (AuNP) and the thionine (THi), are always doped into the electrode matrix for better conductivity [10].
Multicomponents electrode matrix presents great application potential, but the formation of a stable electrode matrix is still a challenge. Chemical crosslinking of diferent components in the electrode matrix could improve physical stability but raise costs and lower fabrication convenience. Also, the introduction of extra reaction reagents might alter the electrochemical properties [11]. Tus, a better way of constructing high-performance biosensors with high stability is still needed.
Here, a novel electrochemical glucose biosensor (SiO 2 -CNTs/THi/AuNPs/GOx) with good stability was fabricated through noncovalent interaction with layer-by-layer assembly technology. CNTs were used as supporting materials, and SiO 2 hybrid was applied to improve the stability of the multi component electrode matrix. Tionine and AuNP were also introduced for enhanced conductivity. Te role of the SiO 2 nanoparticles in the formation of the electrode matrix was studied, and enhanced stability was found by electrochemical analysis. Te biosensor demonstrated that the SiO 2 hybrid was a novel and convenient method to produce high-performance glucose biosensors.

Materials and Methods
2.1. Chemicals. SiO 2 was purchased from Degussa AG, Germany. GO was purchased from Solarbio, Beijing. Glucose was from Jinshan Chemical Test, Chengdu. CNTs dispersion in N-methyl-2-pyrrolidone (NMP) was obtained from Chengdu Organic Chemicals co., Ltd. Chinese academy of sciences, Chengdu. Chloroauric acid (HAuCl4) was purchased from Sinopharm Chemical Reagent co., Ltd, Shanghai. Tionine was from Shanghai Yuanye Biotechnology Co., Ltd, Shanghai. All other reagents and solvents were of analytical grade and commercially available and used without further purifcation. Ultra-pure water (18.25 MΩ) was used throughout.

Preparation of SiO 2 -CNTs/THi/AuNPs/GO Biosensors.
Firstly, pretreat the glassy carbon electrode (GCE). GCE (φ3 mm) was frst polished using Al 2 O 3 (φ0.3 μm) and then with Al 2 O 3 (φ0.05 μm) to be the mirror surface. Ten, the electrode was washed with the ethanol and ultra-pure water for 3 times (10 minutes each time) in the sonicator. Ten, a cyclic voltammetry (from −0.6 V to 1.0 V in 0.1 M·H 2 SO 4 solution) was performed to activate the GCE.
Secondly, prepare the SiO 2 -CNTs dispersion and AuNP dispersion. Te SiO 2 -CNTs dispersion in NMP was prepared by mixing SiO 2 with CNTs. Te process can be done as follows: the SiO 2 powder (0.3 mg) was added to a CNTs solution in NMP (1.5 wt%, 7 μL) and then the ultra-pure water (1 mL) was added. Te dispersion was further sonicated for 2 h to ensure the even distribution. Te gold nanoparticle (AuNP) dispersion in water was prepared by reducing HAuCl 4 with sodium citrate, as reported before [12]. A typical procedure can be done as follows: HAuCl 4 (0.0100 g) was frst dissolved in 100 mL ultra-pure water and then heated up to 100°C with vigorous shaking. Te solution was then refuxed for 15 min, and a solution of sodium citrate (1.00 wt% in water) was added till the dispersion turned gray (approximately 0.8 mL was consumed). Te solution was further refuxed till the color turned claret-red and then the dispersion was cooled to room temperature. Te fnal AuNPs dispersion (approximately 0.01 wt%) was stored at 4°C.
Tirdly, fabricate the SiO 2 -CNTs/THi/AuNPs/GO biosensor. Te proposed electrode was prepared by repeatedly doping multicomponent onto the GCE in order ( Figure 1). SiO 2 -CNTs dispersion (4.00 μL), thionine solution (4.00 μL, 0.01 M in water) and AuNP dispersion (4.00 μL) were separately assembled onto the GCE. After each modifcation, the electrode was stored at 4°C for 4 h to vaporize the solvent. Te modifcation of GO was accomplished with a saturated phosphate bufer solution (PBS) of GO with the same procedure.

Electrochemical Measurements.
Electrochemical measurements were performed on a CHI 760E electrochemical workstation (Shanghai Chenhua Instruments Limited, China) with a conventional three-electrode system that consisted of a saturated calomel electrode (SCE) as the reference electrode, a platinum wire as the counter electrode, and a bare or modifed GCE as the working electrode.
Te detection of glucose via cyclic voltammetry (CV) was performed in PBS (5.00 mL, 0.05 M, pH 7.00) or PBS containing K 3 [Fe(CN) 6 ] (5.00 mL, 0.05 M, pH 7.00, concentration of K 3 [Fe(CN) 6 ] was 0.02 M) with a scanning rate of 100 mV/s (from −0.6 V to 1.0 V) at 25 ± 0.5°C ( Figure 2). AC impedance was employed to characterize the modifcation process of the biosensor with a frequency range of 1 to 105 Hz and an amplitude of the AC potential of 5 mV.
For the detection of glucose, a concentration-current working curve was frst acquired by adding a predetermined amount of glucose solution into the electrochemical cell. Te concentration of glucose can be calculated from the current at oxidation peaks using the working curve.

Analysis and Characterization.
Te morphology of the samples was characterized with a Zeiss sigma 300 scanning electron microscope (SEM). Samples were frstly doped onto GCE as described above. Ten the samples for each step were gently removed from GCE, collected, and used for SEM tests directly.
Te FT-IR spectroscopy analysis was performed on a Nicolet Satellite infrared spectrometer in the range 400-4000 cm −1 with a resolution of 4 cm −1 using KBr pellet technique.

Biosensor Fabrication.
Te glucose-sensitive biosensor was formed by the composition of several components on GCE. Te proposed biosensor utilizes GO to achieve specifc glucose detection. Te immobilization capacity of enzymes plays an important role in detection accuracy and reproducibility. To maintain the fexibility of biosensor production and enhance the reliability of enzyme loading, commercially available SiO 2 nanoparticles were introduced into the layer-by-layer assemble process. Also, the multicomponent electrode matrix, containing thionine and gold nanoparticles, was applied to the GCE to fulfl the requirements of sensitive and accurate glucose detection.

Layer-By-Layer Preparation of the Biosensor.
Te biosensor was prepared by doping the diferent composites onto GCE, and the successful modifcation of each step was confrmed by SEM ( Figure 2) and FT-IR (Figure 3) technology. SiO 2 -CNTs complex was frst introduced onto GCE. As shown in Figure 2(a), no obvious self-aggregate of SiO 2 nanoparticles can be seen. On the contrary, SiO 2 nanoparticles were assembled onto CNTs. Tis phenomenon provides the possibility of SiO 2 acting as a crosslinker. Te follow-up modifcation of thionine and gold nanoparticles leads to no obvious morphology changes (Figures 2(a)-2(c)), thus FT-IR and EDS were used to confrm the existence of thionine and gold nanoparticles, as shown in Figures 3 and 4. Te appearance of a peak at 1610.4 cm −1 confrms the successful introduction of thionine molecular, while the yellow color of Au in EDS tests, as shown in Figure 4 confrms the existence of AuNPs in the fnal biosensor. After the addition of GO, large aggregates can be witnessed (Figure 2(d)). GO was introduced to achieve selective glucose detection.
Te addition of SiO 2 was essential in this experiment. Pre-experiments (data not shown) found that without the addition of SiO 2 , the modifcation of the GCE with the SiO 2 -CNTs/THi/AuNPs/GO composites was not physically stable enough, and the composite might fall of from the GCE during the electrochemical experiments. Also, the unstable formation of the electrode matrix might result in the loss of enzyme during the electrochemical test, which further leads to a poor electrochemical stability, as shown in Figure 5. Te strong interaction between CNTs and SiO 2 nanoparticles made the modifcation of GCE through the dripping method much more efective and greatly helped to glue the composite onto GCE. By mixing SiO 2 nanoparticles with the NMP solution of CNTs, the SiO 2 was frst absorbed onto the CNTs without obvious self-aggregation of SiO 2 (Figure 2(a)). Further introduction of diferent composites (Figures2(b), 4 (c), 4 (d)) happened on the SiO 2 nanoparticles, thus providing the "crosslink" efect of SiO 2 .

Electrochemical Characterization.
CV and AC impedance were also performed in PBS (0.05 M, pH � 7.00 with 0.02 M K 3 [Fe(CN) 6 ]) to study the efects of modifcation on the electrochemical properties of the electrode. As shown in Figure 6, only the redox peak of K 3 [Fe(CN) 6 ]) can be seen for SiO 2 -CNTs (Figure 6(a)) electrode with a high impedance value (Figure 6(b)). After introducing thionine, the strong redox peaks of thionine can be found at −0.4 V and 0.1 V (Figure 6(b)), which confrms the existence of thionine. Te introduction of thionine also decreased the impedance value ( Figure 6(b)), due to the excellent electron transfer efciency of thionine [13]. After the modifcation of gold     International Journal of Analytical Chemistry nanoparticles ( Figure 6(a)), the peak current decreased while the impedance value increased slightly. We suggested this phenomenon might be caused by the thiol-gold interaction formed between AuNP and -SH in thionine, which limits the electron transfer ability of thionine. Further modifcation of GOx leads to an obvious increase in impedance value ( Figure 6(b)) since the GOx exhibits poor electroconductivity.

Optimization of Test Condition.
Te accurate detection of glucose heavily relies on the activity of the GOx modifed on the electrode. Tus, to maximize the activity of the GOx and the sensitivity and accuracy of the detection, temperature and pH during the CV test were, respectively, optimized. As shown in Figure 7, the current at 0.075 V reaches the maximum value (0.75 × 10 -4 A) when the CV test is performed at 25°C, which indicates this temperature favors the activity of the GOx. Tis phenomenon was consistent with the literature report before. Lower temperatures inhibited the activity of GOx and led to a smaller current, while high temperatures caused irreversible damage to the enzyme. Tus, 25°C was selected as a favorable temperature for this biosensor in future experiments.
During the optimization of pH, diferent phenomena were found. According to Figure 8, the biosensor shows a similar sensitivity to glucose in a wide pH range from 4.50 to 7.00, indicating this biosensor might be suitable for a variety of samples from weak acids to neutrals. Te peak current decreased dramatically when pH higher than 7.00, suggesting that the enzyme might be unstable in base. Tus, 7.00 was selected as favorable pH for this biosensor in future experiments.

Stability of the Biosensor.
To confrm the efect of SiO 2 as a crosslinker, CV test of 30 CV cycles was conducted to study its infuence on stability ( Figure 5). For the biosensor without the SiO 2 hybrid, the current retention rate was about 60% after 30 CV cycles. While the addition of SiO 2 resulted in a 96% current retention after 30 cycles with a current loss International Journal of Analytical Chemistry of less than 5%. Te much-improved long-term stability might result from the enhanced physical stability of the electrode matrix, as suggested from the SEM images above, in which SiO 2 acts as a crosslinker. Figure 9 shows the relation between the concentration of glucose and the peak current in which the concentration of glucose is between 1.96 × 10 −9 and 7.24 × 10 −7 M. Te regression equation is y � −2.86 × 10 −5 log C-6.14 × 10 −5 (R 2 � 0.993), which presents a good linear relationship between concentration and peak current.

Working Curve and Determination of Glucose.
Te detection accuracy and precision of the proposed electrode were further analyzed through the detection of preweighed samples under the optimized condition. Te accuracy was described as the mean relative error. [14] For a sample with a glucose concentration of 1.92 × 10 −8 M, the biosensor suggests a concentration of 1.82 × 10 −8 M, with an average detection recovery rate of 105% and accuracy of 5% (n � 5). Tis afrms that the proposed electrode could present an accurate detection toward glucose.

Conclusions
In summary, a novel glucose biosensor was successfully fabricated using layer-by-layer assembly technology. SiO 2 was introduced into the electrode matrix as the physical crosslinker. Electrochemical analyses were carried out to confrm the enhanced stability. Te improved physical stability of the electrode matrix promoted GOx immobilization and further ensured accurate, sensitive, and selective glucose detection with a detection range from 1.96 × 10 −9 to 7.24 × 10 −7 M. Tis work demonstrated a novel way to fabricate an efective and efcient electrochemical biosensor through the hybridization of cheap inorganic components, which provides a new possible solution to bridging the gap between laboratory and practical applications of electrochemical analysis.

Data Availability
Te TXT data used to support the fndings of this study are included within the supplementary information fles.

Conflicts of Interest
Te authors declare that there are no conficts of interest.  International Journal of Analytical Chemistry 7