Figure 1a shows the experimental design of this work. Figure 1b schematically illustrates the material system and fabrication procedure of self-powered implantable skin-like glucometer [40, 41]. The device is assembled by positioning parallel ZnO nanowires (with uniform direction) onto the Ti interdigitated electrodes on a soft Kapton substrate. To ensure the outputting piezoelectric voltage, the distance between neighboring electrode couples is 20 μm, larger than the length of ZnO nanowire (~ 12 μm) [42]. Figure 1c shows the optical images of the device. The thin device is 0.4 × 1.3 cm2 in size and can be easily bended. The thickness of the Kapton film is 250 μm, and the overall thickness of the device is less than 260 μm. Compared with our previous work [14], the device structure and application are greatly improved in this study. The new substrate is more flexible and stretchable. The nanowires are horizontally aligned on the substrate instead of vertical arrangement, which can facilitate the biosensing process. The new device is very small and suitable for embedding into the organism and can be implanted into mouse body to detect glucose concentration in vivo.
The morphology and structure of ZnO nanowire arrays and self-powered implantable skin-like glucometer are shown in Fig. 2. Figure 2a shows the SEM image of as-grown ZnO nanowire arrays on Ti substrate on the side view, showing that ZnO nanowires are vertically aligned on the substrate. The average length of the nanowires is about 12 μm. Figure 2b, c shows the top-view SEM images of as-grown ZnO nanowire arrays, showing that ZnO nanowires have a uniform distribution on the substrate with the same growth direction and the average diameter of the nanowires is about 160 nm. Figure 2d shows the SEM image of horizontally aligned ZnO nanowire arrays (overwhelmed in one direction) before smearing on Kapton substrate [43]. The same growth direction of the overwhelmed nanowire arrays ensures that the crystallographic orientations of the nanowires in the following device are aligned along the same direction. Consequently, the polarities of the induced piezo-potential of individual nanowires are also in the same direction, leading to a macroscopic potential contributed constructively by all of ZnO nanowires. Figure 2e shows the high-resolution TEM image of one single ZnO nanowire. It can be observed that the lattice spacing is 0.52 nm, consistent with (001) crystal plane of wurtzite structural ZnO [44, 45]. The XRD pattern of ZnO nanowires is shown in Fig. 2f, and the sharp diffraction peaks indicate good crystalline quality. All the diffraction peaks can be indexed to Ti (JCPDS card no. 44-1294) and the hexagonal wurtzite structure of ZnO (JCPDS card no. 36-1451). The top-view SEM image of the device (before washing with acetone) and the enlarged view are shown in Fig. 2g, h, respectively. ZnO nanowires locate across Ti interdigitated electrodes. Figure 2i shows that after removing the photoresist with acetone one single ZnO nanowire bridges the two electrodes in one electrode couple.
Figure 3 shows the piezoelectric-biosensing performance of self-powered implantable skin-like glucometer. The device is completely immersed in the aqueous solution of glucose, as shown in Fig. 3a [46]. Under an applied deformation, the device (being bended) can transfer the mechanical energy to piezoelectric voltage that carries the information of glucose concentration. No other electric power is needed during this process, and the outputting piezoelectric voltages can act as both the power source and the biosensing signal. Figure 3b shows that the outputting piezoelectric voltage of the device decreases with increasing glucose concentration. As the concentrations of glucose are 0, 0.024, 0.045, 0.076, and 0.119 g L−1, the voltage of the device is about 0.49, 0.42, 0.32, 0.17, and 0.11 V, respectively. The enlarged views of the outputting piezoelectric voltage against 0.024, 0.045, and 0.076 g L−1 are shown in Fig. 3c. These results indicate that the device can monitor the glucose concentration [47, 48].
The response of the device can be calculated from Eq. 1 [49]:
$$R\% = \left| {\frac{{V_{0} - V_{i} }}{{V_{i} }}} \right| \times 100\%$$
(1)
where V0 and Vi are the outputting piezoelectric voltage of the device in pure water and glucose aqueous solution, respectively. Figure 3d shows that the response of the device against glucose concentrations of 0.024, 0.045, 0.076, and 0.119 g L−1 is 16, 53, 188, and 340, respectively. Obviously with the increasing concentration of glucose solution, the outputting piezoelectric voltage is reduced. On the contrary, the response increases with increasing glucose concentration. This device can sensitively detect glucose concentration, and no external electricity power is needed in the piezo-biosensing process. The outputting piezoelectric voltage is linearly dependent on glucose concentration (0.024–0.119 g L−1), as shown in Fig. S4. The limit of detection (LOD; the signal-to-noise ratio is 3:1) can be calculated to be about 0.019 g L−1 [50]. Figure 3e exhibits the piezoelectric voltage of ZnO nanowires with and without GOx modification in air with different force frequencies. It can be seen that the outputting piezoelectric voltage arises from the piezoelectric effect of ZnO nanowires.
When the self-powered implantable skin-like glucometer is practically used inside a human body, the various kinds of electrolytes and metabolites in body fluid may influence the biosensing accuracy [51]. Therefore, selectivity is an important parameter of the device to detect specific target analytes. Figure 4 shows the selectivity test of the device against three kinds of common ingredients in the body fluid (glucose, fructose, and urea) [52, 53]. The response of the device against fructose and urea is very small (Fig. 4a, b). Figure 4c, d shows that the change in outputting piezoelectric voltage is due to the reactions between glucose and oxidase, rather than the influence of water. The results suggest that the device has a good selectivity for detecting glucose. The influence caused by other electrolytes and metabolites can probably be neglected.
Figure S1 shows the sensing performance of self-powered implantable skin-like glucometer under different forces (14, 22, 31, and 40 N). The applied forces have the same frequency of 1.0 Hz. Under the force of 14 N, the outputting piezoelectric voltage of the device against 0, 0.024, 0.045, 0.076, and 0.119 g L−1 glucose solutions is 0.45, 0.39, 0.27, 0.18, and 0.09 V, respectively (Fig. S1a). Under the force of 22 N, the outputting piezoelectric voltage is 0.49, 0.42, 0.32, 0.17, and 0.11 V, respectively (Fig. S1b). Under the force of 31 N, the outputting piezoelectric voltage is 0.60, 0.50, 0.41, 0.29, and 0.16 V, respectively (Fig. S1c). Under the force of 40 N, the outputting piezoelectric voltage is 0.64, 0.56, 0.46, 0.34, and 0.22 V, respectively (Fig. S1d). As the applied force increases, the outputting piezoelectric voltage increases (Fig. S1e). Interestingly, as shown in Figure S1f, the response of the device under different applied forces is very close. This feature can facilitate the practical applications of our device.
For showing the flexibility of self-powered implantable skin-like glucometer, the biosensing performance at different bending angles is presented in Fig. S2 [54]. The bending angle is controlled by the moving distance of the stepper motor. Figure S2a shows the outputting piezoelectric voltage of the device against 0.045 g L−1 glucose solution at different bending angles. The relationship between outputting piezoelectric voltage and bending angle is shown in Fig. S2b. When the bending angle of the skin-like glucometer is 0°, 10°, 20°, and 30∘, the outputting piezoelectric voltage in 0.045 g L−1 glucose solution is 0.27, 0.32, 0.41, and 0.45 V, and the voltage in 0 g L−1 glucose solution is 0.45, 0.49, 0.58, and 0.63 V, respectively. As the bending angle increases, the response decreases (Fig. S2c). It may be attributed to the damage on GOx@ZnO nanowires as the bending angle increases.
The stability of self-powered implantable skin-like glucometer is shown in Fig. S2d. The glucose concentration is 0.045 g L−1; the applied force is 21 N; and the bending angle is kept at 20∘. After bending for 100, 200, 300, 400, 500, 600, 700, 800, and 900 times, the response is 47, 51, 52, 49, 49, 48, 48, 53, and 51, respectively. These results indicate that the device has high stability.
The working principle of self-powered implantable skin-like glucometer is schematically illustrated in Fig. 5. The working mechanism is based on the coupling between enzymatic reaction and piezo-screening effect of ZnO nanowires. In pure water, no reactions take place on the surface of GOx@ZnO nanowire, and the surface carrier density of the nanowire is low, as shown in Fig. 5a. Under applied deformation, GOx@ZnO nanowire can create piezoelectric potential with weak piezo-screening effect, and the outputting piezoelectric voltage is high (Fig. 5b). When the device is immersed in the glucose solution, the GOx attached on ZnO nanowire surface reacts with glucose, as shown in Fig. 5c. On the first step, gluconic acid and H2O2 are generated as the following reaction [55,56,57]:
$${\text{Glucose}} + {\text{H}}_{2} {\text{O}} + {\text{O}}_{2} \mathop{\longrightarrow}\limits^{\text{GOx}}{\text{Gluconic}}\;{\text{acid}} + {\text{H}}_{2} {\text{O}}_{2}$$
(2)
It has been reported that H2O2 can transfer electrons to ZnO nanowire and produce hydrogen ions as the following decomposition reaction [58]:
$${\text{H}}_{2} {\text{O}}_{2} \to 2{\text{H}}^{ + } + {\text{O}}_{2} + 2{\text{e}}^{ - }$$
(3)
In this process, electrons are absorbed on the surface of the nanowire, which can increase the surface carrier density [59]. And H + ions can also be absorbed on the surface of the nanowires, acting as extra carriers [14, 55]. Under applied deformation, the outputting piezoelectric voltage of the nanowire decreases due to the strong piezo-screening effect from the large amount of H+ and e− on the surface of the nanowire (Fig. 5d).
To further confirm the mechanism, the response of the device against H2O2 and CH3COOH (H+) solutions has been tested as shown in Fig. S3. As the concentration of H2O2 is 0, 0.024, 0.045, 0.076, and 0.119 g L−1, the outputting piezoelectric voltage of the device is about 0.46, 0.29, 0.19, 0.13, and 0.07 V (Fig. S3a), and the response is about 0, 55, 133, 258, and 505 (Fig. S3c), respectively. As the concentration of CH3COOH is 0, 0.024, 0.045, 0.076, and 0.119 g L−1, the outputting piezoelectric voltage of the device is about 0.601, 0.479, 0.435, 0.358, and 0.301 V (Fig. S3b), and the response is about 0, 25, 38, 67, and 99 (Fig. S3d), respectively. It can be seen that the outputting piezoelectric voltage of the device decreases with increasing concentration of H2O2 or H+. To easily understand all the above experimental results and compare them, Table S1 lists the information.
Figure 6 shows the practical application of self-powered implantable skin-like glucometer implanted in a mouse body for detecting blood glucose concentration without any external electricity power source [60,61,62]. The device is implanted into the mouse body by surgery process. After that, a constant force of 4 N controlled by a programming motor is applied on the device in the soft mouse abdomen (Fig. 6a). As the mouse is in anesthesia and cannot have movement, the motor is used to provide force to drive the device. The stepper motor keeps providing constant force (both the frequency and magnitude kept constant) on the device throughout the whole measurement. Thus, the applied strain of the device before and after glucose injection keeps the same. The response of the device is presented in Fig. 6b, c. It was shown that our devices are still able to respond sensitively to the changes in glucose concentration in the biological environment. Without the injection of glucose solution, the blood glucose concentration of mouse is 0.756 g L−1 (measured by a commercial blood glucometer), and the outputting piezoelectric voltage is around 0.16 V. After injecting 0.045 g L−1 glucose aqueous solution (5 mL) into the mouse abdomen, the outputting piezoelectric voltage decreases to 0.075 V. The commercial glucometer shows that the blood glucose concentration of mouse is 0.792 g L−1. These results can simply and roughly demonstrate that our device can work inside the mouse body and detect changes in blood glucose concentration.
Our device can harvest tiny mechanical energy of body movement, such as finger pressing or arm flexing. These motions can easily provide enough force (several tens of newton) for driving the device. It should be noted that the device needs to work during the moment of movement and cannot provide its own energy during sleep or rest. In our experiment, the piezoelectric voltage of the device inside the mouse body needs to be tested by external voltmeter. And the device needs to be driven by external applied deformation. Further work can be focused on integrating the signal processor, wireless signal emitter, and mechanical unit with the device, and the whole system can work independently inside the body.
It should also be pointed that the device is still invasive since a “surgery” process is conducted to implant this device into a mouse under the skin, which is against the future demand of noninvasive glucose monitoring. Thus, there are two development direction of the device in the future. One is that the device will detect body fluid outside the skin (tear, saliva, and so on) by improving performance. The working range of our device (0.024–0.119 g L−1) is below the concentration found in blood (0.36–5.4 g L−1). By considering the size of the device, its biocompatibility and problems with long-term usage of enzymatic sensors, we could probably fabricate a contact lenses glucose sensor (concentration in tears 0.018–0.108 g L−1) in the future study [17]. The other one is that the overall size of the sensor system will be integrated into the micro-nano-level, and it can be implanted into the human body through a simple method without surgery. It should also be noted that the Kapton film is relatively thick, and this Kapton film cannot be conformably implanted inside the body. The material system and the device structure have not reached the level for practical application. In the future work, we will print the device on much conformable substrates, such as PDMS.