Thermal energy harvesting
As shown in Fig. 2a, the proposed DC-ThEG was composed of three main components. First, a PI film was prepared as the substrate, which allows the device to possess good flexibility. Second, double-chain thermocouples made of Bi2Te2.7Se0.3 (n-type) and Sb2Te3 (p-type) were printed atop the PI film by using a screen-printing process, which serve as thermoelectric legs for thermoelectricity generation and electrodes for multi-functional sensing. Finally, a silk fibroin layer was employed to cover the gap between the two separate thermoelectric leg chains, working as a functional material to detect the existence of liquid-state water in the air and the temperature. Compared with single-chain thermoelectric legs, the structure of double-chain thermoelectric legs allows more functional sensing based on the capacitive effect to be realized (i.e., detecting the existence of liquid-state water in the air and the temperature) while ensuring the power density for thermoelectric generation due to the unchanged number of thermoelectric legs. Figure 2b shows a photograph of a fabricated DC-ThEG with dimensions of 60 mm × 35 mm × 0.2 mm in a bending state, and the demonstrated flexibility reveals the feasibility of the DC-ThEG being a wearable device. The partially enlarged image of the device shown in Fig. 2c shows that the thermocouples were well printed atop the PI substrate. In addition, in this work, the gap between any two adjacent thermoelectric legs, the width of all thermoelectric legs and the length of all thermoelectric legs of the fabricated device were 0.8 mm, 1.5 mm, and 22 mm, respectively. The resistance of two thermoelectric leg chains in series (10 pairs of thermocouples for the entire device) was 1.75 kΩ. The surface morphologies of the two thermoelectric materials of the fabricated device were investigated by using scanning electron microscopy (SEM), as shown in Fig. 2d, e. The compactness and good distribution of the surfaces reveal that the thermoelectric legs were well screen printed atop the PI substrate. In addition, the corresponding average thicknesses of the screen-printed n-type (Bi2Te2.7Se0.3) and p-type (Sb2Te3) thermoelectric legs were approximately 105 μm and 83 μm, respectively, as shown in Fig. S2 in the Supporting Information file.
To verify the feasibility of using the proposed DC-ThEG as a wearable power source for thermal energy harvesting, the comprehensive output performance of the fabricated device was systematically investigated by a series of tests and measurements. Figure 3a illustrates the schematic diagram of the working mechanism based on the Seebeck effect of the proposed DC-ThEG for thermal energy harvesting. When a ΔT is applied to the device, charge carriers (i.e., electrons in the n-type thermoelectric material of Bi2Te2.7Se0.3 and holes in the p-type thermoelectric material of Sb2Te3) diffuse from the hot side to the cold side, resulting in a potential being generated in the circuit, which can be measured by a digital multimeter. Moreover, a switch K1 was used to alternate the working states of the device (i.e., thermal energy harvesting and multi-functional sensing), and a switch K2 was used to alternate the testing states of the measurement platform (i.e., open-circuit state and load-connected state).
The temperature difference is an essential parameter to evaluate the performance of ThEGs, which is defined as the temperature at the hot side minus that at the cold side simultaneously measured by a thermometer. Figure 3b shows the practical temperature rising trends of hot side temperature T1 and cold side temperature T2 when a 90 °C heat source was applied to one side of the device. The hot side temperature T1 rose sharply from 25.2 °C (room temperature) to 86.5 °C, while the cold side temperature T2 rose slightly from 25.2 °C to 36.5 °C. As shown in Fig. 3c, when switch K1 was turned on and switch K2 was turned off, the open-circuit output voltage of the developed DC-ThEG was obtained, which increased linearly with increasing temperature difference ΔT applied to the device (i.e., the black curve). A high open-circuit output voltage of ~151 mV and an output power of 13 μW (i.e., the red curve) were achieved when the ΔT of the device was 50 °C. Figure 3d shows the output performance of the proposed DC-ThEG for external loads with various resistances from 0.1 kΩ to 30 kΩ at different temperature differences when switches K1 and K2 were turned on. We can observe the following two aspects: First, when the value of the load resistor was constant, the output power to the corresponding load resistor increased with increasing temperature difference of the device. Second, when the temperature difference of the device was constant, the output power to the resistors first showed a rising trend as the resistance of the external load resistors increased from 0.1 kΩ to 1.8 kΩ and then showed a declining trend as the resistance of the external load resistors increased from 1.8 kΩ to 30 kΩ. In other words, when the fabricated device worked as a thermal energy harvester, its matched load was ~1.8 kΩ. In this work, the maximum output power of the fabricated DC-ThEG to external resistors reached ~2.9 μW when the ΔT of the device was 50 °C and the load resistor had a matched resistance of 1.8 kΩ. In addition, taking the total area of the thermoelectric legs and junctions into account (i.e., 8.43 cm2), the corresponding maximum load output power density was 3.44 μW/m2. It is worth mentioning that due to the continuous direct current output property of ThEGs, the fabricated device has the potential to quickly charge capacitors to power wearable electronics, which is demonstrated in the section “The charging property of the DC-ThEG”.
Reliability and repeatability measurement
For wearable energy harvesting devices, remarkable mechanical reliability, and output repeatability are required. Therefore, a series of reliability and repeatability tests were carried out to systematically evaluate the wearability of the proposed DC-ThEG, as shown in Fig. 4. Figure 4a exhibits the resistance change ratios of a fabricated DC-ThEG in the bending state for various radii from 5 cm to 1.5 cm in both the long-axis (A–A′) and short axis (B–B′) directions. The change ratio of the resistance is defined as (R−R0)/R0, where R0 refers to the initial total resistance of the 10 pairs of thermocouples and R refers to the real-time resistance tested under special operation. The inset of Fig. 4a shows an illustration of the various bending radii. When the fabricated DC-ThEG was bent along the long-axis (A–A′) and short axis (B–B′) to a radius of 3 cm, the corresponding resistance change ratios of the 10 pairs of thermocouples were 1.69% and 1.38%, respectively. In other words, the fabricated device can be bent to a radius of 3 cm in both directions while the total resistance remains almost constant. Figure 4(b) shows the resistance change ratios and output performance of the fabricated DC-ThEG after enduring different bending cycles to a radius of 3 cm. After 1000 bending cycles to a radius of 3 cm in the long-axis direction (A–A′) and the short axis direction, the corresponding change ratios of the resistance of the device were 4.41% and 4.46%, respectively, while the output voltages of the device at ΔT = 50 °C remained almost consistent with the original value. The inset of Fig. 4b exhibits a bending illustration of the device marked with the bending directions. As shown in Fig. 4c, the resistance of the two chains slightly increased with increasing temperature, resulting from the enhancement of the phonon scattering of charge carriers reducing the mobility of charge carriers43,44. The resistance change was approximately linear, with a low rate of a less than 1% increase per 6 °C. In addition, the repeatability of the proposed DC-ThEG was studied by a 100-cycle heating experiment, as shown in Fig. 4d. After 100-cycle heating, both the total resistance of the 10 pairs of thermocouples (returned to room temperature) and the output voltage of the device (ΔT = 50 °C) remained highly consistent with the corresponding original values. In summary, the above analysis reveals the remarkable mechanical reliability and output repeatability of the developed DC-ThEG, which make it meet the requirements for being a reliable power source of wearable electronic devices.
To improve the integration and functionality of wearable ThEGs, we developed a novel structure of double-chain thermoelectric legs, as shown in Fig. 2a, which makes the ThEG feasible as a capacitance-based sensor while ensuring the generated electricity density. In this work, we used silk fibroin to cover the gap between the two thermoelectric leg chains to serve as the functional component for detecting the existence of liquid-state water in the air and the temperature.
In fact, two-state water (i.e., gas and liquid states) can coexist in the air. The gaseous-state water in the air is water vapor, while the liquid-state water in the air refers to suspended tiny droplets that have a balance between gravity and buoyancy, i.e., fog. Conventional humidity sensors can react to both liquid- and gaseous-state water in the air; therefore, it is very difficult for conventional humidity sensors to judge whether liquid-state water molecules exist in the air. Due to the differential absorption behaviors of silk fibroin for different-state water in the air (i.e., gas and liquid states), the proposed DC-ThEG was demonstrated to be a sensor for detecting the existence of liquid-state water in the air.
The working mechanism and measurement of the proposed DC-ThEG for the detection of liquid-state water in the air are exhibited in Fig. 5a–c. When the silk fibroin between the two thermoelectric leg chains absorbs water molecules, its dielectric constant increases, resulting in an increase in the capacitance of the DC-ThEG. In contrast, the dielectric constant of the silk fibroin declines when the silk fibroin desorbs water molecules, leading to a decrease in the capacitance of the DC-ThEG. The above working mechanism combined with the differential absorption behaviors of silk fibroin for different-state water in the air (i.e., gas and liquid states) allows the developed DC-ThEG to detect the existence of liquid-state water in the air. In the experiment, a test setup was built to verify this characteristic of the fabricated device, as shown in Fig. 5a. Two humidity controllers based on different working principles were separately used to supply gaseous- and liquid-state water molecules to simulate the corresponding air conditions, and an LCR meter was applied to trace the real-time capacitance change of the proposed DC-ThEG.
When switches K1 and K2 were turned off, the DC-ThEG worked in a capacitance-based sensing mode, as shown in the circuit diagram in the inset of Fig. 5b. The experimental results shown in Fig. 5b exactly validate the differential absorption behaviors of silk fibroin for the two-state water in the air. In the experiment, from 0 s to 10 s, two-state water were separately applied to the device to observe the capacitance changes in the corresponding absorption processes. At 10 s, the supply of water applied was stopped, and the device was removed to normal air conditions to investigate the desorption processes. When the two-state water were separately applied to the DC-ThEG, apparently different capacitance response behaviors of the fabricated device were observed. The capacitance response of the device to gaseous-state water in the air, shown as the red curve in Fig. 5b, remained almost constant in both the absorption and desorption processes. From the enlarged capacitance response waveform shown in Fig. 5c, only ~0.2 pF normal fluctuations were observed before and after the tenth second, revealing no reaction of the fabricated device to gaseous-state water in the air. In contrast, the fabricated DC-ThEG exhibited an intense capacitance increase in response to liquid-state water, as shown by the black curve in Fig. 5b. The capacitance of the DC-ThEG rapidly increased from the initial value of ~15.9 pF to ~100.1 pF within 10 s after the liquid-state water supplied by the humidity controller was applied to it. When the supply of liquid-state water was stopped and the device was removed to normal air conditions, the capacitance of the DC-ThEG rapidly dropped to 25 pF within 2 s due to the large water molecule concentration difference between the device and air environment and then recovered to the initial value ~8 s later. Therefore, the developed DC-ThEG was proven to possess the ability to detect the existence of liquid-state water in the air based on a combination of Fig. 5a–c and the above analysis. This feature also has good repeatability, which can be proven by combining Figs. S3, S4, and the corresponding analysis in the Supporting Information file.
In addition, a change in temperature usually causes a change in the dielectric constant of a dielectric material, i.e., an increase or a decrease, which may be linear or nonlinear. An increase in temperature will intensify the molecular motion of silk fibroin, leading to an increase in the dielectric constant of silk fibroin45. In this work, we also studied the impact of temperature on the dielectric constant of the silk fibroin we prepared by observing the change in the capacitance of the device with temperature. As a result, a linear relationship between the capacitance of the DC-ThEG and temperature was observed, as shown in Fig. 5d, which provides powerful evidence demonstrating the feasibility of the developed DC-ThEG functioning as a temperature sensor.
Interaction between thermal energy harvesting and functional sensing
In this work, thermal energy harvesting based on the thermoelectric effect (i.e., the Seebeck effect) and multi-functional sensing based on the capacitive effect were integrated in a single device; however, these characteristics might interact due to the differences between their working conditions, i.e., a temperature difference for power generation and moist air for functional sensing. Therefore, to systematically investigate the interaction between these characteristics, a series of experimental comparisons were carried out, as shown in Figs. 6 and 7. The effect of the silk fibroin cover on the thermoelectric output performance of the proposed DC-ThEG was studied by testing the same device before and after covering the gap between the two thermoelectric leg chains with silk fibroin, as shown in Fig. 6a. The open-circuit voltage of the device after covering the gap with silk fibroin film showed a slight decay only in cases of large temperature differences, such as ΔT > 25 °C. In other words, in the case of wearing, the silk fibroin cover had almost no effect on the thermoelectric performance of the device. Moreover, we placed the device in an atmospheric environment, a gaseous-state water molecule-filled environment and a liquid-state water molecule-filled environment and tested the corresponding output voltages to investigate the effect of the two-state water molecules on the output performance of the device. As shown in Fig. 6b, regardless of whether the fabricated DC-ThEG worked in the gaseous-state water molecule-filled environment or the liquid-state water molecule-filled environment, a negligible decay in the open-circuit voltage of the device was observed only in cases of large temperature differences, such as ΔT > 30 °C, indicating the stability of the developed DC-ThEG to moisture in thermal power generation, especially in the case of wearing.
In addition, the impact of temperature on functional sensing was studied by observing the capacitance response behaviors of the DC-ThEG for supplied liquid-state water molecules at different device temperatures, as shown in Fig. 7. From the capacitance response waveforms shown in Fig. 7a, the response of the proposed DC-ThEG to supplied liquid-state water molecules gradually weakened as the device temperature increased. When the temperature of the device reached 40 °C, the corresponding capacitance response only showed a slight increase from 20.4 pF to 21.6 pF within 10 s after the liquid-state water supplied by the humidity controller was applied to the device, and the capacitance of the device remained almost consistent when the device temperature reached 45 °C, which can be observed in the enlarged capacitance response waveforms shown in Fig. 7b, c, respectively. Therefore, the characteristic of the fabricated DC-ThEG of detecting the existence of liquid-state water in the air was susceptible to temperature. Fortunately, the surface temperature of human skin and ambient temperature are usually not higher than 40 °C; therefore, in wearing conditions, the developed DC-ThEG possesses a sensitive capacitive response to liquid-state water in the air, enabling it to serve as a detector of liquid-state water in the air.
It is worth mentioning that both thermal energy harvesting and temperature sensing are based on changes in temperature; therefore, there is no mutual influence between them. In other words, the thermal power generation and sensing functions of liquid-state water detection and temperature detection can coexist under each other’s working conditions in the case of wearing.
The charging property of the DC-ThEG
As a green energy technology, ThEGs are expected to be applied to convert industrial waste heat and human body heat into electrical energy; thus, they are considered a solution for the energy crisis and energy pollution. One of the main parameters for evaluating the performance of a ThEG is the charging capability. In this work, for two application scenarios of high-temperature environments and wearing conditions, we tested the charging property of the proposed DC-ThEG, as shown in Fig. 8. For high-temperature environments, we took ΔT = 50 °C as an example and tested the charging capability of the proposed DC-ThEG by charging a 2200 μF capacitor and twenty-two parallel 2200 μF capacitors. As shown in Fig. 8a, the charging times for the capacitor and twenty-two parallel capacitors to be charged from 0 to ~150 mV were 19.6 s and 369.0 s, respectively. It is worth mentioning that the twenty-two 2200 μF capacitors were connected to a series-parallel switching circuit and then to the DC-ThEG. The circuit diagram of the series-parallel switching circuit is shown in Fig. S5 in the Supporting Information file. After the twenty-two parallel capacitors were charged to 150 mV, we turned off the parallel switches and turned on the series switches; thus, a 3.3 V output was achieved, as shown in the inset of Fig. 8a. This 3.3 V output can power most commercial electronic devices, revealing the potential practicability of the proposed DC-ThEG. In addition, to further evaluate the charging capability of the DC-ThEG, several other capacitors with different capacitance values were selected to be charged with the DC-ThEG, i.e., 100 μF, 220 μF, 470 μF, and 1000 μF, as shown in Fig. S6 in the Supporting Information file. It took only 11.9 s for twenty-two parallel 100 μF capacitors to be charged from 0 to 150 mV, verifying the excellent charging capability of the proposed DC-ThEG. Furthermore, to indicate the real impact on the sustainable wearable energy supply, 4 DC-ThEGs were connected in series and worn on a human arm to convert human body heat into electricity to charge twenty-two parallel 1000 μF capacitors, as shown in Fig. 8b, c. It took 860.5 s to charge the twenty-two parallel 1000 μF capacitors to 55 mV, as shown in Fig. 8d, and a more than 1.2 V output was obtained by the series-parallel switching circuit, which can power some low-power-consumption electronic devices, such as the commercial calculator shown in Fig. 8e. The corresponding processes of charging the capacitors and powering the calculator are exhibited in Supplementary Video S1. It is worth mentioning that high silica cloth with a thickness of 1 mm was attached to one side of the devices, serving as a heat insulation layer to ensure a temperature difference. In summary, the above experimental results combined with the results shown in Fig. 5 indicate that the proposed DC-ThEG will have a wide range of applications in the foreseeable future.