2.1.

CMOS Image Sensor and Instrumentation

A CMOS image sensor (Sponsored from Silicon File Technologies) is an electronic device that converts electrons into digital numbers. In the present experiment, a NOON0PC30L, 110,000 active single chip CMOS image sensor was used. The sensor contains $367\times 314$ multi pixel photon arrays and is connected to a 10-bit analog-to-digital converter (ADC). The photodiode absorbs photons from living room lights and they are converted as electrons into a digital number by the ADC. Moreover, FE-SEM (JEOL-ISM-7500F; JEOL, Tokyo, Japan) images and the DLS method (Malvern Instruments Korea, Seoul, Republic of Korea) were used to confirm the average particle size of ${\mathrm{AFSiO}}_2$ nanoparticles and their distribution. Electrophoretic light scattering spectrophotometer (ELS-800; Japan) was used to measure the zeta potential or overall surface charge of the ${\mathrm{AFSiO}}_2$ nanoparticles. Furthermore, a UV-spectral study (Carry Win Spectrophotometer) determined the absorbance of the ${\mathrm{AFSiO}}_2$ nanoparticles before and after assay reagent immobilization and oxidation. Three individual samples were used for each spectral analysis at 25°C. XRD analysis (Model D/Max-2200; Rigaku, Woodlands, Texas) was used to identify the crystallites of the synthesized ${\mathrm{AFSiO}}_2$ nanoparticles and the reacted sample. Functional groups were recorded by an FTIR spectroscopy (TENSOR 27 instrument, Bruker Optics, Alexandria, NSW, Australia) for ${\mathrm{AFSiO}}_2$ nanoparticles before and after enzyme immobilization and oxidation.

2.2.

Synthesis of ${\mathrm{AFSiO}}_2$ Nanoparticles

For ${\mathrm{AFSiO}}_2$ nanoparticle synthesis, 6.16 ml of ammonia (${\mathrm{NH}}_3$), $500\mu \mathrm{l}$ of 3-aminopropyl triethoxysilane (APTES), and 4.48 ml of distilled water (DI) were added to 6.95 ml of ethanol and then stirred for 1 h using a magnetic stirrer. Then 2.5 ml of TEOS was applied to the mixture drop wise using an injection pump (flow rate: $830\mu \mathrm{l}/\min$) and it was stirred for 24 h. The silica colloids were recovered by high speed centrifugation. Then the solutions were dialyzed with ethanol and deionized (DI) water to remove unreactive materials. The collected nanoparticles were mixed with assay reagent before being applied to the hydrophilic PET film.

2.3.

Preparation of Assay Reagent

Glucose oxidase (GOD) ($\mathrm{100,000}\mathrm{units}/\mathrm{g}$), peroxidase (POD) ($300\mathrm{units}/\mathrm{g}$), and 35 mM chromogen of 2, 2’-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) were purchased from Sigma Aldrich, and reagents were prepared at their respective concentrations using standard laboratory procedures. All solutions were prepared and stored in the refrigerator until use. The enzymes and chromogen were mixed at 2:2:1 volume ratios. The required amount ($15\mu \mathrm{l}$) of assay reagent was added onto ${\mathrm{AFSiO}}_2$ nanoparticles to induce the interaction between the assay reagent and ${\mathrm{AFSiO}}_2$ nanoparticles.

2.4.

Blood Collection from Mouse

Eight-week-old male C57BL/6J mice were purchased from Charles River Laboratories, Japan. The mice were housed under controlled temperature ($22\pm 2°\mathrm{C}$) and 12 h light/dark cycle with free access to water and a regular diet (PicoLab Rodent Diet 20, #5053, LabDiet). After 1 week acclimation period, mice were divided into two groups and fed either a regular diet or a 60% high fat diet (HFD) (D12492, Research Diet) for 4 weeks. At 13 weeks of age, to test various levels of blood glucose, glucose ($2\mathrm{g}/\mathrm{kg}$ body weight) was intraperitoneally administrated in 16 h overnight-fasting mice and blood was collected from the tail vein at 0, 15, 30, 60, 90, 120, 140, and 160 min ($n=3$). A lower concentration of glucose analysis was carried out with 20 h fasting mice and 5-h fasting mice. Glucose concentration was analyzed using $5\mu \mathrm{l}$ whole blood by a glucose oxidase method on a GM9 glucose analyzer II (Analox, London) and CMOS-based measurements. All the experiments were performed in compliance with the relevant laws and institutional guidelines and also in accordance with the institutional Center of Animal Care and Use Committee of Lee Gil Ya Cancer and Diabetes Institute, Gachon University. For the CMOS analysis,²⁸ mice having different glucose concentrations were used for glucose analysis. The whole blood collection schedule and concentrations of glucose are given in Table 1.

Table 1

Whole blood collection schedule from RC and high fat diet (HFD) mice before and after being fed IpGTT.

2.5.

Chip Fabrication for Glucose Analysis

A simple disposable chip was fabricated by using hydrophilic PET film. Sample pad and plasma separation membranes were used to filter the plasma from whole blood, and the absorption pad was used to absorb excess plasma or an oxidized sample. The $10\mathrm{mm}\times 4\mathrm{mm}$ size of the reaction area with a capacity of $15\mu \mathrm{l}$ was developed in the bottom layer of the PET film ($20\mathrm{mm}\times 8\mathrm{mm}$). Assay reagent immobilized on ${\mathrm{AFSiO}}_2$ nanoparticles was dispensed over the reaction area of the hydrophilic PET film, which was stored in the refrigerator until use. To fabricate the chip, a $10\mathrm{mm}\times 8\mathrm{mm}$ size of a hydrophilic PET film was fixed over the bottom layer of the PET film, as shown in Fig. 1. Both lateral edges of the top layer were strongly fixed on the bottom layer of the PET film by a double-sided adhesive layer, which increases the capillary action. This was followed by a $4\mathrm{mm}\times 8\mathrm{mm}$ absorption pad, which was fixed at the right end of the chip, and was made of cellulose fiber with a high absorption capacity. One edge of the plasma separation membrane ($6\mathrm{mm}\times 8\mathrm{mm}$) was fixed on the left side of the bottom layer while the other end overlapped the top layer of the PET film. Four edges of the separation layer were pasted on the PET film. The plasma separation membrane gently touched the bottom layer for easy and rapid sample flow to the reaction area. The $6\mathrm{mm}\times 8\mathrm{mm}$ sample pad was fixed and pasted at the edge of the bottom layer. A schematic illustration of the glucose chip is shown in Fig. 1, layer by layer. Previously, various types of glucose chips were prepared for whole blood glucose analysis that exhibited filtration defects, time consuming kinetic reaction, and less photon sensitivity. Finally, a chip was fabricated without any defects along with a high sensitivity to the respective analysis method.

Fig. 1

Fabrication of hydrophilic poly-ethylene terephthalate (PET) film chip for whole blood glucose analysis. Assay reagent immobilized amine functionalized silica (${\mathrm{AFSiO}}_2$) nanoparticles are coated on the bottom layer of the PET film and the reaction area is covered with a top layer for capillary action. Then the plasma separation membrane, sample pad, and absorption pad arer fixed to make a fully fabricated chip. Each membrane has been fixed by a double sided adhesive layer.

2.6.

Glucose Assay in Disposable Poly-Ethylene Terephthalate Chip

For the glucose assay, assay reagent immobilized ${\mathrm{AFSiO}}_2$ nanoparticles ($15\mu \mathrm{l}$) were dispensed over the reaction area before fixing the top layer of the PET film. A hydrophilic PET film ensured uniform immobilization over the entire reaction area. ${\mathrm{SiO}}_2$ absorbed the water content of the assay reagent, and the amine group kept the enzymes from leaching. The enzyme immobilized nanoparticles helped to develop the chip as a semi-transparent substrate, which eradicates the scattering effect and produces consistent photon transmission. The fully fabricated chip was kept in a refrigerator until use. Twenty-eight whole blood samples were collected from normal and HFD mice and applied to individual glucose chips. Cell debris and blood cells were captured by the sample pad and plasma separation membrane, respectively. The plasma separation membrane could filter about 90% of the plasma, which was more pure than that obtained by centrifugation. Filtered plasma flowed onto the bottom layer of the chip as well as over the reaction area due to capillary action. Plasma glucose reacted well with the assay reagent for 1 to 2 min. An excess amount of sample flowed over the chip and was absorbed by the absorption pad. Each concentration of glucose produced a different intensity of green color that could be observed by the naked eye. Each chip was quantitatively analyzed by the CMOS image sensor.

2.7.

Biochemical Analysis for Enzyme Activity and Stability to the Chip

The enzyme activity and stability of the chip was analyzed by a CMOS image sensor. To determine the specificity and sensitivity, different aqueous sugars were analyzed in the chip including galactose, sucrose, and lactose. Analyses were independently performed on the chip three times. To determine the thermal stability of the immobilized assay reagent, five different concentrations of the whole blood glucose such as 110, 163, 262, 416, and $494\mathrm{mg}/\mathrm{dL}$ were used. The assay reagent immobilized chip was kept in a hot air oven at 20, 30, 40, 50, 60, and 70°C for 1 h and then the glucose oxidation was carried out and analyzed by the CMOS image sensor. To determine the pH stability of the chip, assay reagent was dissolved in the phosphate buffer with pHs of 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0. For the analysis, five different concentrations of whole blood glucose samples were used as previously mentioned. Moreover, the glucose chip’s storage stability was analyzed with a 6 months incubated chip. Five chips were used to analyze the glucose oxidation with various concentrations of blood glucose, such as 178, 249, 308, 388, and $472\mathrm{mg}/\mathrm{dL}$. The CMOS image sensor measured the color intensity changes and proved the photon number variation.

3.1.

Surface Charge and Surface Morphological Analysis

Assay reagents immobilized on ${\mathrm{AFSiO}}_2$ nanoparticles as electrostatic interactions and were spread over the hydrophilic PET film easily and uniformly. The interactions between ${\mathrm{AFSiO}}_2$ nanoparticles and assay regents were confirmed by the zeta potential. The zeta potential has shown around $+27\mathrm{mV}$ for ${\mathrm{AFSiO}}_2$ nanoparticles since the presence of a positively charged amine group was found on the surface and the zeta potential decreased to $-20\mathrm{mV}$ for negatively charged assay reagent. FE-SEM analysis was carried out to find the fundamental characterization of surface morphology of the chip. Thereby, ${\mathrm{AFSiO}}_2$ nanoparticles, enzyme-immobilized nanoparticles, and oxidized nanoparticles were observed by SEM, as shown in Fig. 2. Figures 2(a)–2(c) show the uniform distribution of ${\mathrm{AFSiO}}_2$ nanoparticles that adsorbed as two layers on PET film. The particle size was around 150 nm in each film. The reaction area of the chip images is presented in Figs. 2(e)–2(g), which proves the differences between ${\mathrm{AFSiO}}_2$ nanoparticles and assay reagent immobilized nanoparticles. The enzyme immobilized film had a thin layer formation on ${\mathrm{AFSiO}}_2$ nanoparticles as well as a thick layer formation after glucose oxidation. The FE-SEM images confirmed the interactions between assay reagent-${\mathrm{AFSiO}}_2$ nanoparticles and proved their adhesion over the film. Moreover, Figs. 2(d) and 2(h) show the DLS analysis at 633 nm and a scattering angle of 90°C. The data confirm a particle size of $140\pm 21\mathrm{nm}$ for ${\mathrm{AFSiO}}_2$ nanoparticles, and the size increased to $180\pm 32\mathrm{nm}$ after enzyme immobilization on ${\mathrm{AFSiO}}_2$ nanoparticles. The DLS data proved the assay reagent immobilization on ${\mathrm{AFSiO}}_2$ nanoparticles by particle size increment.

Fig. 2

Field emission scanning electron microscopic (FE-SEM) images of ${\mathrm{AFSiO}}_2$ nanoparticles immobilized on PET film. (a) Uniform particle distribution of ${\mathrm{AFSiO}}_2$ nanoparticles on the PET film. (b) Two-layer formation on the PET film with uniform distribution. (c) Particle size was uniform and the ${\mathrm{AFSiO}}_2$ nanoparticles size was 150 nm. (d) DLS estimation of ${\mathrm{AFSiO}}_2$ nanoparticle: measurements were performed before mixing an enzyme in the ${\mathrm{AFSiO}}_2$ nanoparticles; the particle size is $140\pm 21\mathrm{nm}$. (e) Enzyme-immobilized ${\mathrm{AFSiO}}_2$ nanoparticles on PET film. (f) Uniform enzyme immobilization on ${\mathrm{AFSiO}}_2$ nanoparticles as a thin layer. (g) Oxidized glucose on ${\mathrm{AFSiO}}_2$ nanoparticles immobilized on a PET chip. (h) Dynamic light scattering (DLS) estimation of ${\mathrm{AFSiO}}_2$ nanoparticle-enzyme complexes: measurements were performed after mixing an enzyme in ${\mathrm{AFSiO}}_2$ nanoparticles and the size was $180\pm 32\mathrm{nm}$.

3.2.

Ultraviloet-Spectral Analysis

Figure 3(i) shows the UV-visible absorbance spectrum of ${\mathrm{AFSiO}}_2$ nanoparticles, assay reagent immobilized nanoparticles, and oxidized glucose. The analysis was carried out to find the level of interaction between the assay reagent, ${\mathrm{AFSiO}}_2$ nanoparticles, and oxidized glucose which displayed an absorbance in the region from 200 to 800 nm. The absorption curve for ${\mathrm{AFSiO}}_2$ nanoparticles showed a hump at 210 nm due to the localized surface plasmon of nanosized ${\mathrm{SiO}}_2$ particles. The assay reagent showed two peaks at 350 and 400 nm; also, the assay reagent immobilized ${\mathrm{AFSiO}}_2$ nanoparticles showed broader peaks at 400 nm and the absence of peaks at 350 nm. There was a peak shift to 405 nm, and additional peaks developed at 560 nm after glucose oxidation. This result confirms the effective interaction between the assay reagent and ${\mathrm{AFSiO}}_2$ nanoparticles; it also proves that the interactions of the assay reagent immobilized ${\mathrm{AFSiO}}_2$ nanoparticles were not influenced by the quality of glucose oxidation on the chip. Moreover, ${\mathrm{AFSiO}}_2$ nanoparticles have various biological applications such as optical sensing, bio adsorption, and biocompatibility.

Fig. 3

UV-spectral absorbance and Fourier transform-infrared spectroscopy (FTIR) spectra analysis before and after glucose oxidation; (i) The ultraviolet (UV) spectrum shows the ${\mathrm{AFSiO}}_2$ nanoparticles, assay reagent, assay reagent-immobilized nanoparticles, and oxidized glucose. Peak shift identified at 405 nm and additional peak developed at 560 nm after glucose oxidation. (ii) FTIR spectra analysis before and after the reaction: (a) hydrophilic PET film, (b) ${\mathrm{AFSiO}}_2$ nanoparticles, (c) assay reagent immobilized on ${\mathrm{AFSiO}}_2$ nanoparticles, (d) oxidized glucose, (e) chip stability after 20 days. The FTIR results prove the structural and functional properties of a glucose chip are stable after immobilization and oxidation.

3.3.

Fourier Transform-Infrared Spectroscopy Structural Analysis

Structural information and interactions were studied for bare hydrophilic PET film, an ${\mathrm{AFSiO}}_2$ nanoparticle coated film, the assay reagent immobilized ${\mathrm{AFSiO}}_2$ film, and oxidized glucose films by FTIR, which are shown in Fig. 3(ii). Generally, hydrophilic PET film peak absorption energy considerably increased at $\sim 500–1750{\mathrm{cm}}^{-1}$, including a benzene ring (721), primary alcohol (1085 and 1013), and hybrid sp2 $\mathrm{C}─\mathrm{O}$ [1233 and $\mathrm{C}═\mathrm{O}$ (1707)]. Absorbance signals at $1546{\mathrm{cm}}^{-1}$ were assigned to the asymmetrical stretching vibration of $\mathrm{COO}─$.^27,29 The 2-Methacryloyloxyethyl phosphorylcholine unit featured one phosphor atom, an organic acid $─\mathrm{COOH}$, and an inorganic acid $\mathrm{P}─\mathrm{O}$ on both sides of the molecules structures, directly improving the hydrophilicity.³⁰ The spectrum of the hydrophilic PET film showed almost the same characteristic bands as the present study. The peak at $1060{\mathrm{cm}}^{-1}$ indicates $\mathrm{Si}─\mathrm{O}─\mathrm{Si}$ asymmetric stretching vibrations, whereas $795{\mathrm{cm}}^{-1}$ was assigned to $\mathrm{Si}─\mathrm{O}─\mathrm{Si}$ symmetric stretching vibrations. The peak at $937{\mathrm{cm}}^{-1}$ was assigned to ν ($\equiv \mathrm{Si}─\mathrm{OH}$) of a free silanol group.^31–34 Moreover, a new sharp peak was identified at $2965{\mathrm{cm}}^{-1}$, which indicates the presence of ${\mathrm{AFSiO}}_2$ nanoparticles.³³ Figure 3(ii) c almost shows the same peak as the ${\mathrm{AFSiO}}_2$ nanoparticles, and new peaks at 1563 and $1613{\mathrm{cm}}^{-1}$ indicate assay reagent immobilized successfully on ${\mathrm{AFSiO}}_2$. New sharp and moderate peaks appeared at 1556 and $1654{\mathrm{cm}}^{–1}$, which seem to indicate oxidized plasma glucose and other substances in the plasma. The assay reagent immobilized chip was analyzed after 20 days to confirm the stability of the assay reagent. There was no alteration of the characteristic peaks in the respective image of Fig. 3(ii) e, which supports chip stability. The structural information and interactions on bare hydrophilic PET film, ${\mathrm{AFSiO}}_2$ nanoparticle coated film, enzyme immobilized chip, oxidized glucose chip, and chip stability results prove that there was no alteration in their characteristic peak after interactions on the PET film, which does not affect the glucose oxidation. All these characterization studies have proven the ${\mathrm{AFSiO}}_2$ nanoparticle chip suitable for CMOS image-based whole blood glucose sensing. The intensive characteristic absorption peaks were observed in the liquid stage, while a new sharp peak appeared in the liquid stage of the sample at $2985{\mathrm{cm}}^{-1}$, indicating the amine groups in the ${\mathrm{SiO}}_2$ nanoparticle suspension.³⁵ Moreover, sharp peaks appeared at 2960 and $2893{\mathrm{cm}}^{–1}$, which indicate assay reagent immobilized on ${\mathrm{AFSiO}}_2$ nanoparticles in molecular water (data not shown). Based on these functional groups, we were able to understand the structural vibrations of ${\mathrm{AFSiO}}_2$ nanoparticles, enzyme immobilized ${\mathrm{AFSiO}}_2$, and oxidized glucose; their functional groups were not affected when bonded on the chip.

3.4.

X-Ray Diffraction Analysis

Figure 4 shows the XRD patterns of the hydrophilic PET film before and after enzyme immobilized films. XRD patterns of the hydrophilic PET film show a broader peak in the range of $2\theta =23\mathrm{deg}$ and 25 deg, which can be allotted to the semi-crystalline structure of the PET film or an extrusion process in the production of the PET film.³⁶ A sharp peak occurs in the range of $2\theta =23\mathrm{deg}$ and 25 deg, indicating the microcrystalline structure of the ${\mathrm{AFSiO}}_2$ nanoparticles, and the intensity decreased gradually after assay reagent immobilization. The broader peaks identified at $2\theta =23\mathrm{deg}$ and 25 deg indicate semi-crystallites for enzyme-immobilized and oxidized glucose substrates, which show that the immobilized enzyme did not alter the structure of the ${\mathrm{AFSiO}}_2$ nanoparticles.³⁷ The results corroborated other related studies on enzyme immobilized ${\mathrm{AFSiO}}_2$ nanoparticles adsorbed to PET film and oxidized glucose.

Fig. 4

X-ray diffraction (XRD) analysis for ${\mathrm{AFSiO}}_2$ nanoparticles in hydrophilic PET film before and after immobilization.

3.5.

Photon Analysis by Smartphone Camera Module

Suitable chip fabrication is an important part to this study. The hydrophilic PET film surface was suitable for coupling with the ${\mathrm{AFSiO}}_2$ nanoparticle layer. Amine groups could bind the negatively charged enzymes via the electrostatic interactions that helped to develop the technically suitable chip for analysis by CMOS image sensor. Further, the characterization study proves the ${\mathrm{AFSiO}}_2$ nanoparticle-based PET chip is suitable for glucose analysis. The CMOS image sensor analyzed the chip before and after glucose oxidation. Each CMOS image sensor contains a single photodiode to absorb the photons and convert them into the electrical energy via the photo electric effect. The electrical energy is converted into a digital number, which is directly proportional to the number of photon hits on sensor surface. Thereby, the photon number reduces any object exposed on the sensor. This simple concept creates the impacts on the POC glucose monitoring. Before measurement, the calibration value was fixed in the CMOS image sensor by adjusting the exposure time and analog gain value instead of by maximizing the background light intensity. Then an assay reagent added chip was exposed on the sensor, and the value was fixed as a reference value (175 A.U). In the glucose reaction, plasma glucose reacted with GOX, producing gluconic acid and ${\mathrm{H}}_2{\mathrm{O}}_2$. Following this, ${\mathrm{H}}_2{\mathrm{O}}_2$ reacted with 2, 2’-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) in the presence of POD. Oxidized ABTS produced a green color product based on the glucose concentration through the enzyme kinetic reaction, which is shown in Fig. 5(a). Previously, our group carried out the glucose analysis from $3\mathrm{mg}/\mathrm{dL}$ to $1000\mathrm{mg}/\mathrm{dL}$ successfully using the aforementioned reaction that proved the lower and upper limits of glucose detection.¹⁷ Figure 5(b) shows the schematic of glucose oxidation on the PET chip and monitoring by the CMOS image sensor as a photon-detection technique. For the analysis, 28 PET film chips were analyzed, and each chip produced a different color intensity based on the glucose concentration. Variations in color intensity at different glucose concentrations are shown in Fig. 5(c). The lower the concentration of glucose, the paler the color after oxidation with ABTS, and the color intensity increased with a higher concentration of glucose. Similarly, different concentrations of glucose produce different photon numbers based on the color intensity. A lower color intensity allowed the passing of more photons, and a higher color intensity allowed the passing of fewer photons. Thereby, the photon value decreased when the glucose concentration increased. In other words, the photon number decreased when the color intensity increased. The measured digital output of the photon value is shown in Fig. 6(a) for each individual glucose concentration. The experiment was performed thrice, and the average value is given. The photon value was 161 A.U at a low glucose concentration of $110\mathrm{mg}/\mathrm{dL}$, and the value linearly decreased to 31 A.U. at the highest concentration of $586\mathrm{mg}/\mathrm{dL}$. The change of photon value was linearly correlated to the glucose concentration, and the value was standard and consistent for each experiment. Even small differences of glucose concentrations could change the photon number. The relative oxidation level shown in Fig. 6(b) displays the cumulative value for each concentration of glucose. The relative oxidation level was calculated by subtracting the oxidized glucose value from the chip value. In addition, the CMOS image sensor could measure the glucose levels in mouse whole blood with an accuracy comparable to that of an Accu-Chek sensor, a glucose monitor that is commonly used in diabetes testing. The linearity graph as shown as Fig. 6(c) demonstrates the concentration linearity of the GM9 glucose analyzer and Accu-Chek sensor, whereas Fig. 6(d) shows the linearity of the CMOS image sensor-based photon value to quantify the accuracy. Moreover, the correlation coefficient ($R^2$) value was 0.9675 as a CMOS-based digital output value and 0.9712 for the glucose concentration. The comparison study with other glucose sensors proved the CMOS image sensor-based glucose monitoring system could be applicable for clinical diagnoses to determine blood glucose levels.

Fig. 5

CMOS image sensor-based glucose analysis. (a) Glucose assay by enzyme kinetic method. (b) Schematic representation shows the blood sample reacted with the immobilized assay reagent and developed color, followed by the glucose oxidized chip being monitored by the CMOS image sensor as a photon-detection technique with a digital output value displayed on the smartphone. (c) Various concentrations of oxidized glucose (110, 163, 262, 291, 416, and $494\mathrm{mg}/\mathrm{dL}$; from left to right) the hydrophilic PET film chip shows different color intensities based on glucose concentration.

Fig. 6

CMOS image sensor digital output for different concentrations of glucose and comparative analysis with other glucose sensor: (a) CMOS image sensor-based photon number variation shows the reduction of photon number with increasing glucose concentration. (b) Relative oxidation level shows the digital value for each glucose concentration and dotted lines represent the linearity. (c) Linearity analysis for glucose concentration between GM9 glucose analyzer and Accu-Chek sensor. (d) CMOS image sensor digital output linearity; the linearity data confirm that the CMOS image sensor is almost equivalent to a commercial sensor.

3.6.

Enzyme Activity and Stability Measurements

Analysis of enzyme stability and activity was carried out on the PET chip with five different concentrations of mouse blood glucose. The CMOS image sensor monitored the glucose chip for nonspecific enzyme activity with other sugars. As a result, the assay reagent did not respond with other aqueous sugars and there was no change in color and photon count. The study of nonspecific enzyme activity proved that the glucose chip is very stable and there was no interaction with other nonspecific sugars (data not shown). Subsequently, the chip has been analyzed for thermal stability. The color was developed after 10 min of glucose oxidation due to less kinetic energy at a low temperature of 20°C. However, the assay reagent did not lose stability and the photon count was not influenced at 20°C. The kinetic energy increases when the chip is incubated at 30°C–50°C. Also, the enzyme responded until 70°C and the reaction rate decreased from 50°C because the chip lost the 50% of its stability at 60°C. The chip also decreased its reaction rate at 70°C and lost 90% of its activity. The enzyme denatured at high temperature which leads to a decrease in the reaction rate and an increase in the photon count for the respective concentration of glucose. Figure 7(a) shows the thermal stability of the glucose chip when using different concentrations of blood glucose and proves the optimum temperature is 30°C–40°C. The kinetic energy was also increased and reached the reaction rate within a few seconds; the color was stable for 30 min. The data was compared with the standard value obtained from the CMOS image sensor and the results showed the consistent digital output value at 20°C–40°C. The enzyme was stable when the chip incubated at 20°C, and it took some time to reach the reaction rate. Small changes in temperature above or below the optimum have not affected the enzyme stability; also, ${\mathrm{SiO}}_2$ nanoparticles do not uptake high temperature that provides enzyme stability for longer period. Additionally, the pH stability has been analyzed by the CMOS image sensor indicating the glucose oxidation was carried out with different pH ranges (pH 2.0 to 9.0). The reaction rate was increased between pH 4.0 and 6.0, and the enzyme activity was stable in the noted ranges, which produced a consistent photon count. Furthermore, 60% of the enzyme activity lost their function at pH 2.0–3.0 and above pH 8.0. Figure 7(b) shows the photon number increment when the pH was unfavorable for the glucose oxidation. pH stability analysis proved that the optimum pH range was between 4.0 and 6.0 for the glucose chip. The assay reagent started to lose its stability in the range of pH 2.0–3.0 and pH 7.0–9.0, which increased the digital output value. Small changes in pH did not affect the enzyme activity since the bonds can be reformed. In addition, the storage or shelf stability of the glucose chip has been analyzed after 6 months of incubation. Interestingly, the chip was stable until 6 months without any defects in enzyme activity, which was demonstrated the reproducibility in the photon number after a 6-month incubation. Figure 7(c) shows the photon values of different concentrations of whole blood glucose with the 6-month incubated chip. Normally, the enzymes have a possibility of denaturing during long-term storage due to the environmental conditions. However, the amine functionalized enzyme and the controlled storage conditions prevented enzyme denaturation.

Fig. 7

Assay reagent stability and activity analysis by CMOS image sensor. (a) Thermal stability assay shows the optimum temperature is 30°C to 40°C and the digital value increases with higher temperatures due to a lower reaction rate. (b) pH stability assay shows the optimum pH is 4.0–6.0 and the digital value increases with an unfavourable pH. (c) Six-month incubated chip monitored by CMOS image sensor and shows the consistent photon value based on the glucose concentration.