A Thermally Chargeable Supercapacitor based on the g‐C3N4‐Doped PAMPS/PAA Hydrogel Solid Electrolyte and 2D MOF@Ti3C2Tx MXene Heterostructure Composite Electrode

With the development of individual wearable electronics, the requirements of self‐energy harvest devices from human skin or motion have increased. A thermal harvest device that receives thermal energy naturally existed in human skin is more attractive than a mechanical energy harvester that needs human motion or walking. Herein, a thermal‐chargeable supercapacitor (TCSC) is proposed, which can convert thermal energy into electrical energy and then store the energy only by occurring the temperature difference between the two ends of the TCSC. The all‐solid‐state g‐C3N4‐modified hydrogel electrolyte in the TCSC provides more free protons and energy for proton migration by the electrostatic interaction and hydrogen bond interaction between g‐C3N4 and the acid group. The 2D MOF@Ti3C2Tx MXene heterojunction electrodes with the advantages of large pore size, adjustable and abundant REDOX sites of MOFs, and high conductivity of Ti3C2Tx MXene also ensure the high performance of the TCSC. As a result, the assembled TCSC exhibits excellent ionic thermal‐voltage (55.68 mV), Seebeck coefficient (18.56 mV K−1), and energy exchange efficiency (3.4%) upon a temperature difference of 3 K, and successfully drives the pressure sensor work.


Introduction
Gathering low-grade heat (LGH) released by human skin is considered as a promising way to meet the requirement of selfpowered wearable electronic devices owing to the common, continuous, and natural thermal energy supply from human www.advmatinterfaces.de solid electrolyte. [14] However, protons come from the dissociation of the acidic substances, the dissociation is usually much less than that of metallic salts, resulting in poor conductivity of proton-type hydrogel solid electrolytes. Mixing the 3D porous network-structured polyacrylic acid (PAA) hydrogel with protonic acid [poly(2-acrylamide-2-methyl propane sulfonic acid)] PAMPS to prepare a composite hydrogel captures the virtues of both, in detail, sulfonic groups on PAMPS dissociate protons under hydration, and the short-chain structure formed by low polymerization restricts anionic migration. More importantly, the separation state between PAMPS and PAA skeleton, and the 3D porous network structure of the PAA skeleton also greatly reduces the traction on protons. Besides, the additive of g-C 3 N 4 with graphene structure in PAMPS/PAA hydrogel accelerates the ability of the acidic groups to dissociate protons, which also can serve as a platform for proton migration, whereby the dissociative protons migrate by forming hydrogen bonds with nitrogen atoms on the surface. [15] In addition, the large conjugate structure on g-C 3 N 4 also enhances the storage charge capacity of the hydrogel, the strong hydrophilic property also endows the hydrogel with more satisfactory water retention performance. [16][17] As for electrode materials, 2D Ti 3 C 2 T x MXene has been widely used in the field of electrochemical energy storage because of its highly reversible surface redox reactions, remarkable electrical conductivity, acidic and chemical stability. [18][19] However, Ti 3 C 2 T x MXene suffers from the weaker active sites on its surface, resulting in a low energy density. 2D metalorganic framework (MOF) composed of metal cations (or metal clusters) and organic connectors is especially suitable for highperformance electrode materials due to the extremely large specific surface area, easily adjustable pore size, and abundant pseudocapacitive redox sites. [20][21][22] Considering the electrical conductivity of Ti 3 C 2 T x and the redox activity of MOF, it is of great significance to create 2D heterojunctions with satisfactory performance in electrochemical energy storage.
Based on the above-mentioned hydrogel electrolytes and electrode materials, we proposed a TCSC consisting of 2D NH 2 -Ni/ Co-BDC MOF@Ti 3 C 2 T x MXene heterojunction electrodes and g-C 3 N 4 -doped PAMPS/PAA hydrogel electrolyte. The assembled TCSC shows excellent ionic thermal-voltage of 55.68 mV, Seebeck coefficient of 18.56 mV K −1 , and energy exchange efficiency of 3.4% at dT = 3 K. Moreover, an integrated pressuresensing system powered by TCSC is also shown to realize the real-time monitor of human activity, demonstrating new opportunities for the development of thermal-chargeable-integrated wearable electronics. Figure 1 shows the preparation process and structural/electrochemical properties characterization of the all-solid-state hydrogel electrolyte. g-C 3 N 4 was prepared by sintering dicyandiamide at high temperatures, as shown in Figure 1a. XRD pattern in Figure 1b refers to the standard card (JCPDS 87-1526), the diffraction peaks at 10.94° and 27.83° correspond to the crystal planes of (100) and (002), respectively. [23] Figure 1c depicts the FT-IR spectral, the sharp peak at 802 cm −1 comes from the absorption peak of the triazine ring, and the characteristic peak at 1100-1625 cm −1 is related to the stretching vibration peak of CN and CN. [24] The wide absorption peak at 3090 cm −1 can be attributed to the stretching vibration peak of primary amine and secondary amine at the defect site of g-C 3 N 4 and the intermolecular hydrogen bond. [17] Figure 1d reveals the preparation process of the g-C 3 N 4 -PAMPS/PAA hydrogel electrolyte. The digital photo demonstrates the surface morphology, flexibility, and thickness of the hydrogel. It can be seen that the hydrogel is translucent. It is worth noting that the prepared hydrogel can maintain a thickness of around 4 mm. Given the influence of g-C 3 N 4 doping on the mechanical and electrochemical properties of PAMPS/PAA substrate, we tested the tensile/compression stress-strain curves and electrochemical impedance spectroscopy (EIS) of the hydrogel with different doping amounts of g-C 3 N 4 . As shown in Figure 1e and Table S2 (Supporting Information), 0.16% g-C 3 N 4 -doped PAMPS/PAA hydrogel demonstrates a smallest elasticity modulus (0.02 MPa). On the one hand, the addition of g-C 3 N 4 promotes hydrogen bonding in the system and endows the hydrogel with more attractive flexibility. On the other hand, when more g-C 3 N 4 is doped, the elastic modulus of the hydrogel will also increase, which can be attributed to the electrostatic interaction existing between the positively charged imine group on the surface of g-C 3 N 4 and the acid ions on the polymer chain that is stronger than hydrogen bonds. In addition, as shown in Figure S2a (Supporting Information), according to the compression stress-strain curves of the hydrogels, the g-C 3 N 4 -0.16%@PAMPS/PAA hydrogel needs to resist less force in the process of being compressed to the set position, which also proves its outstanding flexibility. Figure  S2b (Supporting Information) shows that the hydrogel still has good shape retention ability under three cycles of compression. We assembled the symmetrical "sandwich" supercapacitors with a series of hydrogels and MOF-9.10%@Ti 3 C 2 T x MXene electrodes, and carried out GCD, CV, and EIS tests to demonstrate the electrochemical performance of g-C 3 N 4 -doped PAMPS/PAA hydrogel electrolytes. As shown in Figure 1f-g, Figure S3, and Table S3 (Supporting Information), we can find that the specific area capacitance and conductivity of the g-C 3 N 4 @PAMPS/PAA are higher than that of PAMPS/PAA substrate. The reason can be concluded as follows: First, the large-conjugated 2D structure of g-C 3 N 4 improves the capacity of charge transport and storage in the system. [25] Second, the amino and imine groups on g-C 3 N 4 can form acid-base pairs with the sulfonic and carboxylic acid groups on PAMPS and PAA to accelerate the dissociation of the acidic group. [26] Third, g-C 3 N 4 can serve as a platform for proton transport. The dissociated protons form hydrogen bonds with nitrogen atoms on the surface to facilitate proton transport. [27] However, as excessive g-C 3 N 4 is added to the hydrogel, due to the weakness of 2D materials in easy agglomeration, the reunited g-C 3 N 4 blocks part of the proton transport channel, resulting in the reduction of the area ratio capacitance and conductivity of the hydrogel. Based on this, the hydrogel with 0.16% g-C 3 N 4 -doping content of acid polymer has more outstanding mechanical and electrochemical properties. In addition, as shown in Figures S4 and S5 (Supporting Information), we tested the GCD curves of g-C 3 N 4 -0.16%@PAMPS/ PAA hydrogel at different current densities and CV curves of different sweep speeds, indicating that the hydrogel has good rate performance and Coulomb efficiency. Figure 2 shows the synthesis process and structural characterization of the MOF@Ti 3 C 2 T x MXene electrodes. Figure 2a provides a schematic diagram of the synthesis process of the www.advmatinterfaces.de Schiff reaction-induced self-assembly MOF@Ti 3 C 2 T x MXene composite. First, NH 2 -Ni/Co-BDC MOF nanosheets were synthesized by the coordination of bimetallic ions (Ni 2+ , Co 2+ ) with the carboxyl group of NH 2 -BDC ligand, in which the Ni and Co atoms were coordinated with six O atoms to form an octahedral structure. [28] As the ligand itself contains amino groups, there is a large amount of −NH 2 on the surface of MOF nanosheets. Ti 3 AlC 2 was etched by the mixed acid method (HF/ HCl), the surface of layered Ti 3 C 2 T x nanosheets was anchored by a large number of end groups (−F, =O, and −OH). [18] The prepared NH 2 -Ni/Co-BDC MOF was mixed and stirred with Ti 3 C 2 T x dispersion, and the covalent MOF@Ti 3 C 2 T x heterojunction composite was prepared by the Schiff reaction between the amino groups on MOF and the carbonyl groups on the surface of Ti 3 C 2 T x . The crystal structure information of MOF, Ti 3 C 2 T x , and MOF@Ti 3 C 2 T x MXene composite are shown in Figure 2b. The XRD patterns of MOF@Ti 3 C 2 T x MXene exhibit the characteristic peak located at 10.81° of NH 2 -Ni/Co-BDC MOF, corresponding to the crystal plane of (111), and the diffraction peak at 7.13° belongs to Ti 3 C 2 T x , corresponding to the crystal plane of (002). [29] Compared with pure Ti 3 C 2 T x , the characteristic peak of Ti 3 C 2 T x in MOF@Ti 3 C 2 T x MXene shifted from 7.44° to 7.13°, indicating that the insertion of a 2D sheet MOF increased the layer spacing of Ti 3 C 2 T x , which is conducive to the improvement of the electrochemical performance of the composite electrode. [31] The functional groups on the sample surface were characterized by FT-IR, as displayed in Figure 2c. According to the spectral, the peaks of MOF@Ti 3 C 2 T x MXene composite at 3424, 1622, 1404 cm −1 are attributed to the presence of Ti 3 C 2 T x , and the peak at 1548-1398 cm −1 is attributed to the aromatic ring belonging to MOF. In addition, compared with NH 2 -Ni/ Co-BDC MOF, the peak of the N-H group is missing in the spectrum of MOF@Ti 3 C 2 T x MXene composite, which indicates that the composite is not just simple combination of two materials, some kind of chemical reaction exists in the composite electrodes. The XPS test (Figure 2d-f and Figure S6, Supporting Information) was carried out for the composition and valence state analysis. MOF@Ti 3 C 2 T x MXene have all elements of NH 2 -Ni/Co-BDC MOF (C, O, N, Co, Ni) and Ti 3 C 2 T x (C, O, F, Ti). However, due to the small proportion of MOF in the composite material and the strong shielding effect of Ti 3 C 2 T x on the metal elements in MOF, Co and Ni are not displayed in the energy spectrum of all elements. Therefore, we carried out separate energy spectrum detection of Ni and Co in the composite material, and it can be found that, in the Ni 2p spectra of the composite, Ni 2+ and Ni 3+ are located at 855.2 and 873.8 eV, www.advmatinterfaces.de respectively. Correspondingly, in the Co 2p spectrum, Co 2+ and Co 3+ are situated at 782.2 and 797.0 eV, respectively. [30] In order to verify that the composite has an covalent heterojunction structure formed by Schiff reaction between amino groups on MOF and carbonyl groups on the surface of Ti 3 C 2 T x , we compared XPS spectra of NH 2 -Ni/Co-BDC MOF and MOF@Ti 3 C 2 T x MXene, as shown in Figure 2g,h. Compared with the binding energy at 399.1 eV of the N 1s spectrum of NH 2 -Ni/Co-BDC MOF, the N 1s spectrum of MOF@Ti 3 C 2 T x MXene composite shows three different binding energies, which are 399.8 eV belongs to aromatic ring CN, 401.1 eV belongs to CN, and 402.1 eV belongs to graphitic N, respectively. In addition to the two new peaks, the CN bond energy of the aromatic ring in the composite also has a redshift compared with the CN bond energy of the NH 2 -Ni/Co-BDC MOF, which can be attributed to the strong electron donor effect on the surface of Ti 3 C 2 T x . Figure 3a displays the SEM image of NH 2 -Ni/Co-BDC MOF with a 2D rhomboid sheet structure. Figure 3b shows the  Figure S7 (Supporting Information) provide the elemental mapping images of MOF@Ti 3 C 2 T x MXene composite, demonstrating the presence of Ti, N, Ni, and Co, which further confirms the successful hybridization of NH 2 -Ni/Co-BDC MOF with Ti 3 C 2 T x . The electrochemical properties of MOF@Ti 3 C 2 T x MXene electrodes with different MOF contents were examined by CV and EIS measurements in a three-electrode system with 1 m H 2 SO 4 electrolyte. As shown in Figure S8 (Supporting Information), when the MOF amount accounts for 9.1% of the mass of Ti 3 C 2 T x , the composite electrodes exhibit the largest specific mass capacitance of 70.13 F g −1 , which is 180% higher than pure Ti 3 C 2 T x . The CV of MOF-9.1%@Ti 3 C 2 T x MXene composite at different sweep rates (5-100 mV s −1 ) was also tested, as shown in Figure S9 (Supporting Information). EIS was measured to further investigate the electron transfer and ions diffusion at the electrode/electrolyte interface, and Nyquist plots are displayed in Figure S10 and Table S4 (Supporting Information). The semicircle at the high frequency is attributed to the charge-transfer resistance (Rct), and the intercept at the real axis indicates the equivalent series resistance (Rs). [32] Obviously, the Rs of samples declines with the increase of MOF content, such as the values of Rs is 15.99, 20.00, and 20.39 Ω for MOF-9.10%@Ti 3 C 2 T x MXene, MOF-16.70%@Ti 3 C 2 T x MXene, and MOF-23.10%@Ti 3 C 2 T x MXene, respectively. The low Rct of all samples is attributed to the nanosheet-like and mesoporous structure that provides a rapid transfer pathway for ions and electron. [33] Figure 4 shows the thermal-chargeable performance of MOF-9.10%@Ti 3 C 2 T x MXene electrodes and C 3 N 4 -0.16%@ PAMPS/PAA hydrogel electrolyte-based TCSC. In order to meet the needs of human wearable devices, a home-built vertical temperature gradient supply system consisting of heating, cooling, and temperature feedback was made ( Figure S1, Supporting Information), which realized the output and storage of voltage under the condition of temperature difference applied by the upper and lower electrodes of a symmetrical "sandwich type" TCSC. Figure 4a depicts that the ions in the solid electrolyte have a thermal diffusion effect complying with the Soret mechanism, namely ions migrate from the hot end to the cold end when different temperatures are applied to the electrodes at both ends of TCSC, which leads to a large accumulation of ions on the side of the solid electrolyte affected by the cold end. As a result, ions with the opposite sign gather on the electrode, which generates a voltage between the two electrodes. [9] This is a thermionic-charging process; The two electrodes in TCSC are connected to realize the short-circuit state, to simulate the connection of external load. This process belongs to the forward electronic working process; The short-circuit state is removed, and the electrochemical workstation continues to be connected to both ends of TCSC to receive the voltage generated by thermionic movement, but the two poles of the electrochemical workstation are opposite from Stage 1, in order to test the reverse thermal-recharging performance of TCSC; The two electrodes in TCSC have connected again, and the electrons flow out from the capacitor electrode, which belongs to the reverse electronic working process. The thermal cycle consisting of these four stages can realize the conversion of heat energy to electricity in Stages 2 and 4. Figure 4b shows the interaction between components and proton migration in the solid electrolyte. On the one hand, Lewis acid-base pairs are formed between the amino and imine groups on the surface of g-C 3 N 4 and the sulfonic/carboxylic acid groups on the anionic www.advmatinterfaces.de polymers PAMPS and PAA. This interaction weakens the associations between the acid groups and the protons, thus generating more mobile protons and participating in the thermal diffusion process. [34,35] On the other hand, g-C 3 N 4 can also serve as a platform for proton migration. [27] Hydrogen bonds are formed between nitrogen atoms on the surface of g-C 3 N 4 and free protons, and proton migration channels are formed together with acid root sites, so that protons have more energy to carry out migration movement based on thermal diffusion based on the Grotthuss mechanism and accumulate at the cold end. [14] Then, the thermal-chargeable behavior of TCSC under different temperature differences (dT) was tested. As shown in Figure 4c, during the test, the charging time was defined as the 60 s, which is the shortest defined charging time of the currently developed ionic thermal-chargeable supercapacitor, reflecting that the TCSC has more outstanding thermal-response performance. With the gradual increase of dT, the TCSC shows the change of output voltage, which indicates that the thermalrecharging behavior is strongly dependent on the dT gradient. dT = 3 K is the minimum temperature difference applied to both ends of TCSC in this test, and 55.68 mV thermal-voltage is generated after the 60 s, which is the maximum thermal-voltage generated by proton-type thermal-chargeable supercapacitor so far under similar dT. In addition, it is worth mentioning that the relative humidity (RH = 11-17%) of the test environment and the cold/heat source temperatures applied to the ends of the supercapacitor in this study were much lower than those in other studies, which represent practicality to some extent. When dT = 23 K, the thermal-voltage generated reaches 131.16 mV. Figure 4d summarized the thermal-voltage generated by TCSC at different dTs during the test and calculated the Seebeck coefficient under temperature gradient. With the increase of dT applied to both ends of TCSC, the Seebeck coefficient showed a decreasing trend and reached 18.56 mV K −1 when the lowest temperature difference dT = 3 K. The minimum Seebeck coefficient in the whole process even reached 5.38 mV K −1 at dT = 21 K, which reflects the outstanding ionic Seebeck effect of the TCSC. In addition, as shown in Figure S11 (Supporting Information), MOF-9.10%@Ti 3 C 2 T x MXene electrode-based TCSC shows more prominent thermoelectric property under various temperature differences, and the device's average Seebeck coefficient reaches 11.6, which is 3.2 times that of graphite electrode-based and 4.2 times that of conductive stainless steel electrode-base thermoelectric output devices. This breakthrough thermoelectric property is due to the abundant hydroxyl groups on the surface of Ti 3 C 2 T x MXene flakes, which can also be used as a proton source to participate in the ionic thermal diffusion process based on the Soret effect. Furthermore, the thickness of the hydrogel electrolyte has a significant influence on the thermoelectric performance of the TCSC device. By controlling the thickness of the hydrogel growth, we get a range of TCSC devices of different thicknesses and tested their thermoelectric behavior in turn. As shown in Figures S13-S23 (Supporting Information), as the thickness of the hydrogels increases gradually, a larger Seebeck coefficient is displayed, which is due to the accumulation of more mobile protons in the system on the surface of the cold electrode, thus achieving a considerable potential output. However, the further increase of the thickness of the hydrogels increases the diffusion distance of protons from the hot end to the cold end, and the low thermal conductivity of the hydrogel (0.074 W m −1 K −1 ) also leads to the larger temperature difference between the two ends of the device, which both results in the reduction of the Seebeck coefficient of the TCSC devices. Notably, the TCSC device achieves maximum Seebeck coefficient output when the hydrogel thickness is up to 2 mm. In order to further reflect the positive promotion of MOF-9.10%@Ti 3 C 2 T x MXene electrode on the thermoelectric behavior of the TCSC device, we selected graphite and conductive stainless steel plates as electrodes to form thermoelectric output devices with 2-mm thick g-C 3 N 4 -0.16%@PAA/PAMPS hydrogel electrolytes, respectively. According to the reported literature, the difference between human skin temperature and external temperature is about 5-3 K, which belongs to an extremely low-grade temperature difference. [11] Therefore, we respectively tested the multicycle repetitive thermal-rechargeable/discharge process of the TCSC in dT = 5 and 3 K to emphasize its satisfactory stability and reproducibility. As can be seen from Figure 4e, when dT = 5 K, the thermal-voltage of TCSC after five cycles, still maintained the original 92.2%; When dT = 3 K, the thermal-voltage of the TCSC also maintains 85.9%, which not only proves the stability of the device but also demonstrates its real potential in thermoelectric conversion from intermittent heat sources.
The electric energy conversion efficiency (η), Figure-of-merit (ZT), and power density (P) can be used to evaluate the performance of thermal-rechargeable supercapacitors, which are jointly determined by the ion conductivity, thermal conductivity, average Seebeck coefficient, cross-sectional area of the material, and the capacitance of the device. [36] Through testing and calculation, the ion conductivity of TCSC is 12.5 mS cm −1 , the capacitance is 44.32 mF, the average Seebeck coefficient is 11.6 mV K −1 , the cross-sectional area of electrolyte is 1.54 cm 2 , and the thermal conductivity coefficient is 0.074 W m −1 K −1 , relatively. The electric energy conversion efficiency (η) can be obtained from E/Q, where E and Q are the electric energy stored in the device and the quantity of heat provided to the device, and the result of calculation were shown in the Figure 5a. Here, E = CdV 2 /2, where C is the capacitance and V is the output thermal voltage of the TCSC. Q was obtained from Q = k A dt dT, where k, A, dt and dT are the effective thermal conductivity, cross-sectional area of electrolyte, thermal-charging time, temperature difference between the hot and cold side of the electrolyte, respectively. Therefore, as shown in Figure 5b, with the gradual increase of dT, η decreases and then increases, where η 23K = 2.5%, η 5K = 2.5%, and η 3K = 3.4%. Besides, in Figure 5c, the power density P of TCSC can be obtained according to E and P = E/Adt, where P 23K = 0.0413 W m −2 , P 5K = 0.0091 W m −2 and P 3K = 0.0075W m −2 . We further calculated a Figure-of-merit (ZT) of the TCSC in the same way that conventional thermoelectric materials are adopted on the basis of ZT = S ′2 σT/k, where S′, σ, T, and k are the average Seebeck coefficient, electrical conductivity, absolute temperature, and effective thermal conductivity, respectively. The ZT value of TCSC at room temperature and relative humidity (T = 296 K, RH = 13%) is 0.68. To demonstrate practicability, as shown in Figure 5d, we showed that the TCSC was thermal-charged for 5 min under different dTs and the was quickly connected to the positive and negative poles of the voltmeter. The voltage value displayed on the voltmeter indicates that the TCSC has thermalrechargeable performance and energy-storage performance. We soaked the sponge sheet in monolayer-Ti 3 C 2 T x MXene aqueous solution for 20 min, and then dried it in vacuum at 60 °C to obtain the material with impedance varying with pressure, and assembled it with gold electrode into a simple pressure sensor. One side of TCSC was tightly fixed with human skin, and after balancing for 10 min, the TCSC was connected in series with the pressure sensor and electrochemical workstation, and the TCSC was testified capable of driving the pressure sensor to work under the temperature difference between human skin and external environment. The i-t curve in Figure 5e can prove that, driven by the temperature difference between the body surface temperature and the external environment, the TCSC conducts thermal-rechargeable behavior and stores the electric energy as power supply equipment to provide power support for various human health-testing equipment.

Conclusion
In summary, g-C 3 N 4 -doped PAMPS/PAA hydrogel electrolyte and 2D NH 2 -Ni/Co-BDC MOF@Ti 3 C 2 T x MXene heterojunction electrodes-based TCSC was fabricated to collect thermal energy and store electrical energy, which then served as a power source for a flexible wearable pressure sensor. When the minimum temperature difference dT = 3 K is applied in the two electrodes of TCSC, the thermal-voltage generated by the device is 55.68 mV, the Seebeck coefficient reaches 18.56 mV K −1 , and the energy conversion is 3.4%. Upon dT = 23 K, the devices exhibit a thermal-voltage of 131.16 mV, Seebeck coefficient of 5.38 mV K −1 , and energy conversion of 2.5%. In addition, the TCSC has considerable cyclic thermal-recharging performance, and the ability to store electric energy and drive wearable sensors, which represents the practicality of the TCSC in thermalchargeable-integrated wearable electronics.

Experimental Section
Synthesis of Mono and Fewer Layers Ti 3 C 2 T x MXene: The selective etching method was carried out according to our previous work. [18] Synthesis of NH 2 -Ni/Co-BDC MOF: 0.287 mmol Ni(NO 3 ) 2 ·6H 2 O and 0.143 mmol Co(NO 3 ) 2 ·6H 2 O were dissolved in 30 mL DI water. Meantime, 0.207 mmol APTA was dissolved in 20 mL mixed ethanol/ DMF solution (v:v = 1:1). After the two solutions were mixed and stirred for 30 min to form transparent and homogeneous solution, which was then poured into a Teflon-lined stainless-steel autoclave and heated at 150 °C for 20 h. After the reaction, the product was centrifuged at 8000 rpm for 5 min and the reactants were cleaned several times with DMF and ethanol until the supernatant was clear. The final samples were obtained with purple after vacuum-drying overnight at 60 °C. [30] Synthesis of bimetallic MOF@Ti 3 C 2 T x MXene electrode: 5 mL of Ti 3 C 2 T x MXene suspensions (2 mg mL −1 ) was added in the MOF contained aqueous solution (1 mg NH 2 -Ni/Co-BDC MOF, 1 mL water) and stirred vigorously in 0 °C under dark situation for 4 h to prepare www.advmatinterfaces.de MOF-9.1%@Ti 3 C 2 T x MXene. The MOF-16.7%@Ti 3 C 2 T x MXene (2 mg MOF) and MOF-23.1%@Ti 3 C 2 T x MXene (3 mg MOF) were prepared at the same conditions. The MOF@ Ti 3 C 2 T x MXene electrodes were prepared via the spraying coating method on the flexible PET film.
Synthesis of g-C 3 N 4 -0.16%@PAMPS/PAA hydrogel: 5 g of AMPS and 0.05 g of NH 4 S 2 O 8 were dissolved in 20 mL DI water and heated at 70 °C for 5 h to obtain 20 wt% PAMPS aqueous solution. 5 g of dicyandiamide was heated at 550 °C for 5 h to prepare g-C 3 N 4 . [26] 0.01 g NH 4 S 2 O 8 and 0.01 g MBAA were then added under ultrasound for 20 min. 0.3 g of AA was added to 2.5 g 20 wt% PAMPS solution and stirred for 15 min, the prepared g-C 3 N 4 dispersion (1.3 mg g-C 3 N 4 , 1 mL DI water) was then added drop by drop and stirred for 12 h to obtain the precursor, which was placed on the oven at 65 °C for 4 h to obtain g-C 3 N 4 -0.16%@ PAMPS/PAA hydrogel. Pure PAMPS/PAA hydrogel, g-C 3 N 4 -0.33%@ PAMPS/PAA, and g-C 3 N 4 -0.65%@PAMPS/PAA composite hydrogel were prepared by the same method.
Fabrication of the TCSC: Two MOF@ Ti 3 C 2 T x MXene film electrodes were coated on both sides of the prepared hydrogel with a small amount of precursor, which was then heated at 65 °C for 15 min; after cooling down, the TCSC was fabricated.
Calculation: The specific capacitance (C) of the TCSC: where dt is discharging time in GCD testing (s), dV is voltage window (V), and I is discharging current (A). The Seebeck coefficient (S) under different temperature differences: The energy density (E) and power density (P): The heat energy (Q): d d Q kA t T = (4) Figure 5. a) The electric energy stored (E) in the device, quantity of heat provided to the device (Q), b) energy conversion efficiency (η), and c) power density (P) of the TCSC under different dTs in 60 s. d) The voltage value of the TCSC after thermal-charging as measured by voltmeter under dT = 0/3/5/23 K. e) Display drawing of the TCSC driving the pressure sensor under the temperature difference between human skin and environment with the addition of i-t curve generated by pressure sensor under 20 g weight. f) Schematic of the TCSC as a wearable thermally chargeable power supply.

www.advmatinterfaces.de
The energy conversion efficiency (η dT ): The ionic conductivity (σ, S cm −1 ) for a solid electrolyte: The Figure-of-merit (ZT, 1) for a traditional thermoelectric device: where dV is the thermal-voltage value (mV), dT represents the temperature difference (K), C is capacitance in the presence of a temperature difference (F), dt is thermal charging time in thermalchargeable test (s), A is the surface area of the supercapacitor (cm 2 ), k is the thermal conductivity coefficient of solid electrolyte (W m −1 K −1 ), L is the thickness of the solid electrolyte (cm), R is impedance of the solid electrolyte according to EIS test (Ω), A′ is the cross-sectional area of the solid electrolyte (cm 2 ), and S′ is average Seebeck coefficient of TCSC.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.