“One stone two birds” or “you can't have your cake and eat it too”? Effects of device dimensions and position of the thermoelectric module on simultaneous solar-driven water evaporation and thermoelectric generation

Abstract

Widespread interest has emerged in addressing the energy crisis that makes use of abundant solar energy to simultaneously evaporate water and produce electricity. Photothermal conversion materials (PCMs) and device structures are the two key platforms of photothermal conversion in the process of water evaporation (WE) and thermoelectric (TE) generation, and thus it is very important to reasonably design the structure of PCMs and the corresponding devices. Herein, three-dimensional (3D) and two-dimensional (2D) flexible carbon melamine foam (CMF) combined with Cu particles with excellent plasma effect and CuS with excellent UV-vis light absorption performance were assembled as a PCM (CMF/Cu@CuS). This functionalized porous CMF/Cu@CuS foam can not only realize WE but also simultaneously generate TE via the Seebeck effect. We found that the 3D structure favors WE and the 2D structure favors TE generation. Additionally, the TE generation is significantly impacted by the position of the TE module. Specifically, when the TE module is placed at the bottom of the PCM, the photothermal conversion efficiency (PCE) of the 3D structure reaches 86.8%, while that of the 2D structure is only 68.3% under 1.0 kW m⁻² light irradiation. However, the maximum output TE power of the 2D structure reaches 69.3 μW cm⁻², which is significantly higher than that by the 3D structure (2.4 μW cm⁻²). This difference in performance is due to the different thermal management abilities of the devices. The 3D structure can absorb extra heat from the surrounding environment, which is conducive to WE. However, the 3D structure leads to a small temperature difference between the hot and cold sides of the TE module, which is not conducive to TE generation. The 2D structure cannot absorb extra heat from the environment, which is not conducive to WE, but it causes a higher temperature difference between the hot and cold sides of the TE module, and is thus conducive to TE generation. A high WE rate can be attained as well as high TE power produced when the TE module is positioned close to the top surface of the PCM. According to this study, a properly adjusted TE module position is necessary to simultaneously achieve a high WE rate and output TE power.

---

1. Introduction

Clean water and green electricity are basic elements related to life, economic development, and social progress and have become significant global challenges.^1,2 To solve the shortage of water and electrical energy, several technologies have been developed and applied, including reverse osmosis,³ triboelectric nanogenerators,¹ flat-panel solar TE generation⁴ and WE-induced electricity generation.^2,5,6 Solar energy has become a promising energy source to replace traditional energy, attracting significant attention. Among the various solar energy utilization technologies, solar-driven photothermal conversion is used to meet the demands of various evolving applications, such as solar-driven WE and desalination^7–9 and electricity generation.^2,5,6 For photothermal conversion WE, early approaches to improve PCE focused on optimizing PCMs with excellent solar absorption performances.^10–13 However, a bottleneck in efficiency has been reached using material development alone owing to the heat losses caused by thermal convection, conduction and radiation. Therefore, recent research has focused on designing devices with different structures.^9,13–15 For example, the WE area has been increased by introducing a well-designed three-dimensional (3D) evaporator,^16–18 allowing additional energy to be absorbed from the surroundings, thereby effectively increasing the PCE beyond the theoretical limit.^19–23

An increasing number of PCMs and evaporation devices has been developed,²⁴ where carbon-based materials are the most important type of PCMs. Among the carbon-based PCMs tested to date, carbonized melamine foam (CMF) stands out, not only because of its excellent hydrophilicity and low thermal conductivity, but also due to its highly porous structure given that it can readily transport water to every active WE site, maintaining a continuous water supply and enhancing the vapor escape rate.^25–27 In addition, the modification of MF with strong light-absorbing materials is also a promising approach for the fabrication of PCMs.^28–30 However, almost all these technologies are focused on a single application although they can actually achieve the combined use of solar energy to meet different energy needs.

Recently, various strategies for solar-driven electric energy generation and fresh water production have been proposed, such as WE-induced salinity gradient,³¹ storing and recycling the steam enthalpy from interfacial solar WE^32,33 and water-flow-induced electric energy generation.^2,34 In addition, solar energy also can be converted into electric energy via the Seebeck effect.^35–40 Niu and Gong report a facile and cost-effective strategy to prepare MnO₂-decorated cotton cloth as an advanced PCM for simultaneous steam and electricity generation.⁴¹ The hybrid device exhibited an extremely high WE rate of 2.24 kg m⁻² h⁻¹ under 1 kW m⁻² irradiation. More importantly, the hybrid device produced an open-circuit voltage of 0.3 V and power output of 1.6 W m⁻² under 3 sun irradiation, outperforming most of the previously reported solar-driven electricity generation devices. Subsequently, they obtained N/Co-doped carbon and MnO/C hybrid nanoparticles through the simple pyrolysis of metal–organic frameworks.^42,43 The flexible MnO/C nanoparticle-based bifunctional solar evaporator presented a high evaporation rate of 2.38 kg m⁻² h⁻¹ and solar-to-vapor conversion efficiency of 98.4% under 1 sun irradiation. More importantly, the evaporator recovered low-grade heat from water evaporation to perform thermoelectric conversion, achieving the maximum output voltage of 330 mV under 3 sun irradiation, together with the output power of 4.65 mW and power density of 2.9 W m⁻².⁴³ These works realized the integration of WE and TE generation, but the effect of device structure and the position of the TE module on the WE rate and TE power has not been systematically studied.

On account of the local surface plasmon resonance (LSPR) effect, metal nanoparticles can effectively absorb different wavelengths of solar light, which are ideal candidates for solar energy collection materials. Many metals nanoparticles, such as Au,⁴⁴ Ag,⁴⁵ Al⁴⁶ and Cu,^47,48 have been used to achieve efficient photothermal conversion WE. Among them, nano-Cu has attracted significant attention due to its low price, easy preparation and excellent WE performance.^47–50 If porous carbon materials and nano-metal particles with the LSPR effect are combined, their advantages can be fully utilized to improve the PCE. In addition, some semiconductors, especially copper chalcogenide nanoparticles, are also used as PCMs based on their merits of simple preparation, low cost and adjustable absorption capacity.^51–53 Tao et al. assembled a 3D CuS/poly(vinylidene fluoride) membrane as a PCM and achieved a PCE of 90.4% ± 1.1% under 1 sun irradiation.^51,54 Shang et al. constructed a porous CuS/polyethylene hybrid membrane, which exhibited a PCE of 63.9%.⁵⁵ Wu et al. prepared a cotton–CuS nanocage–agarose aerogel, which presents strong absorption over the entire UV-vis region and the WE rate reached 1.63 kg m⁻² h⁻¹.⁵⁶ These results indicate that CuS is an effective PCM that can be used for solar-driven WE.

There have been many reports on solar-driven photothermal WE and simultaneous TE generation by the Seebeck effect, but little attention has been paid to the effect of the device dimensions and the position of the TE module on the performance of both.^35–43 Thus, to study the influence of the device dimensions and the position of the TE module on simultaneous WE and TE generation, in this work, we prepared a 3D and 2D flexible CMF/Cu@CuS device for synchronous solar-driven WE and TE generation. We chose CMF/Cu@CuS as the PCM, exhibiting a favorable porous structure, thermal management capability, broadband light absorption and strong LSPR effect. Moreover, due to the low thermal conductivity of bulk water, its temperature is low. Therefore, a large temperature difference will occur between CMF/Cu@CuS and the bulk water, resulting in simultaneous WE and TE generation due to the Seebeck effect. Furthermore, when the TE module was placed at the bottom of the PCM, the WE rate of the 3D device was 1.85 kg m⁻² h⁻¹ and the PCE was as high as 86.8% under 1 kW m⁻² light illumination. The maximum output TE power was 2.4 μW cm⁻². Surprisingly, compared with the 3D devices, the WE rate and PCE of the 2D devices were lower, which were only 1.32 kg m⁻² h⁻¹ and 68.3%, but their output TE power could reach 69.3 μW cm⁻². Subsequently, the performance difference between the 2D and 3D structures and the position of the TE module in photothermal WE and TE generation were analyzed. This study provides a useful reference for designing devices that can simultaneously perform photothermal WE and TE generation.

2. Experimental

2.1. Preparation of CMF and CMF/Cu foam

CMF was quickly prepared from commercial MF by annealing at 400 °C in air for 10–15 min. The white MF turned black after carbonization, but still maintained its good mechanical properties. To obtain CMF/Cu, 0.6 g copper acetate, 12 mL ethylene glycol amine, 20 mL ethylene glycol and 50 mL deionized water were mixed and stirred at 5 °C for 10 min. The obtained CMF was added to the above solution for 30 min, and then sodium borohydride solution added dropwise (0.1 g sodium borohydride in 10 mL of ethylene glycol and 40 mL deionized water). Finally, the product was removed and washed with ethanol and deionized water several times. The acquired product was denoted as CMF/Cu.

2.2. Preparation of CMF/Cu@CuS and CMF/CuS foam

Briefly, 0.08 g copper sulfate was dissolved in 100 mL water and the prepared CMF/Cu was placed in the copper sulfate solution. After 10 min, CMF/Cu was removed, and then immersed in sodium sulfide solution for 10 min (1.45 g sodium sulfide nonahydrate dissolved in 100 mL water). This was repeated several times and the obtained product was marked as CMF/Cu@CuS. For comparison, CMF/CuS was prepared by replacing CMF@Cu with CMF.

The detailed materials and physical characterization are included (ESI (1) Experimental section).†

2.3. Solar-driven vaporization experiments

The 3D or 2D CMF/Cu@CuS foam was used as a PCM. Polystyrene (PS) foam with lower thermal conductivity served as an insulation layer and supported under the CMF/Cu@CuS foam. The PS foam was wrapped with gauze to enable the water to be transported to the PCM layer. The mass changes in water were measured using an electronic scale. The temperatures were measured using a Data Logger Thermometer (NTC, 10K/3435). The surface temperature and infrared photographs were captured and recorded using an IR camera (TESTO 869, Germany). The simulative solar light was supplied by a solar simulator (PLS-SXE300, Beijing Trusttech Co. Ltd, China). Different solar light intensities were obtained by adjusting the output current.

2.4. Fabrication of the device for the generation of TE power

The commercial TE-module (TEP1-126T200, 4 cm × 4 cm × 0.4 cm) was selected as the TE generation device (Fig. S1†). 3D and 2D CMF/Cu@CuS foams with a size of ∼5.0 cm × 5.0 cm × 3 cm and ∼5.0 cm × 5.0 cm × 0.3 cm, respectively, were assembled on the top the TE module with a wall connected in bulk water. A PS foam frame with gauze was used to control the floating of the TE module to make its lower part submerged in the water and reduce the heat loss. An ammeter and multimeter were applied to measure the output current and voltage, respectively. All experiments were typically conducted at the ambient temperature of 25–28 °C and humidity of ∼70%.

3. Results and discussion

3.1. Properties of the prepared samples

The process for the preparation of the modified flexible CMF with Cu@CuS is shown in Fig. 1. Here, because of its low cost, high porosity, and excellent hydrophilicity, CMF prepared from pure MF in a muffle furnace at 400 °C for 10–15 min was used as the substrate (Fig. 1a). The CMF foam was immersed in copper acetate solution (containing the organic ingredients of ethylene glycol amine stabilizer and ethylene glycol). The reducibility of the glycol amine stabilizer is beneficial to eliminate surface oxides. Then, NaBH₄ as a reducing agent was dissolved in ethylene glycol and added to the above-mentioned solution. To control the reactive rate, the temperature of the mixing solution was maintained at 5 °C. After washing the above-mentioned product several times with ethanol and deionized water, the CMF/Cu composite was obtained (Fig. 1b). The XRD results show that CMF/Cu composite was obtained after NaBH₄ reduction (Fig. 1d). The XRD pattern displayed the (111) and (002) diffraction peaks at 2θ = 43.1° and 50.1°, respectively, which can be perfectly indexed as face-centered cubic copper (JCPDS no. 04-0836).⁴⁹ Subsequently, CuS was prepared by successive ionic layer absorption and reaction (SILAR) method. Initially, CMF/Cu was impregnated in CuSO₄ solution, removed 10 min later, impregnated in Na₂S solution for 2–5 times, and then rinsed with distilled water to obtain the CMF/Cu@CuS composite (Fig. 1c). The XRD results show that CuS was grown on CMF/Cu after SILAR (Fig. 1d). The diffraction peaks at 2θ values of 29.3°, 31.8°, 32.9°, and 47.9° can be assigned to the (102), (103), (006) and (110) planes of the hexagonal covellite CuS crystal phase, respectively, matching well with the standard card of JCPDS no. 65-3588.⁵¹ CMF/Cu@CuS with strong LSPR effect and excellent UV-vis light absorption performance can absorb sunlight and convert light energy into heat energy for photothermal WE and TE generation.

Fig. 1 Schematic illustration for the preparation of (a) CMF, (b) CMF/Cu and (c) CMF/Cu@CuS composite. (d) XRD patterns of CMF, CMF/Cu and CMF/Cu@CuS.

For morphologies and composition analyses, the different composites, including pure MF, CMF, CMF/Cu, CMF/CuS and CMF/Cu@CuS, were studied via FESEM, TEM and EDX. Fig. 2a shows the FESEM image of pure MF. It can be seen that MF has a porous network structure with a smooth surface. After rapid heat treatment, CMF still maintained the porous structure (Fig. 2b), as shown in the enlarged FESEM image (Fig. S2a†). The EDX results show that CMF contained abundant C, N and O elements (Fig. S2b†). When the Cu particles were loaded on CMF, its surface became rough and a large number of small rod-like particles was loaded on its skeleton (Fig. 2c and d). The EDX results showed that the sample contained Cu elements (Fig. S2c†). The inset in Fig. 2d shows the TEM image of the suspension obtained by ultrasonic treatment of CMF/Cu in anhydrous ethanol. It can be seen that the product consisted of irregular particles and nanorods. The nanorods had a diameter of around 14 nm and length of about 100 nm, with the average diameter of the nanoparticles being around 21 nm. Combined with the XRD results of the samples (Fig. 1d), these small rod-like particles should be Cu particles, that is, the CMF/Cu composite was successfully prepared. After SILAR, the surface of the porous structure became rougher and the number of particles loaded on the skeleton increased, and the S element appeared in the EDX results (Fig. 2e, f and S2d,† respectively). The TEM result in the inset in Fig. 2f shows that the product consisted of irregular nanoparticles, nanorods and nanosheets. Compared to the TEM result of CFM/Cu, the emerging nanosheets with a size of about 40 nm should be CuS. Combined with the XRD results of the samples (Fig. 1d), it is shown that the CuS nanosheets were successfully loaded on CMF/Cu, that is, the CMF/Cu@CuS composite was obtained. The mass of CMF (10 × 6 × 3 cm) was 1.1732 g, which increased to 1.2388 g after Cu loading, and then 1.2511 g after five cycles of CuS loading. These results indicate that the contents of Cu and CuS were 0.0656 g and 0.0123 g, respectively. When pure CMF was used as the substrate for SILAR, CMF/CuS composites with a rough surface could be prepared, but the CuS particles were large and their loading on the CMF skeleton was uneven (Fig. 2g and h). It can be seen from the inset TEM image in Fig. 2h that the size of the nanosheet was very uneven. This indicates that the Cu particles on the CMF skeleton contributed to the loading and uniform growth of CuS, and thus Cu@CuS nanomaterials embedded in CMF with the LSPR effect and excellent UV-vis light absorption properties were obtained. In addition, the XPS survey spectrum of CMF/Cu@CuS in Fig. S3† shows peaks corresponding to Cu 3p, S 2p, C 1s, N 1s and O 1s, which agrees with the EDX result. The CMF/Cu@CuS foam had a good channel structure, which can increase the light scattering, improve the utilization rate of incident light,^25–27 and contribute to the transmission of water to the upper layer during solar-driven WE.

Fig. 2 FESEM images of (a) pure MF, (b) CMF, (c and d) CMF/Cu, (e and f) CMF/Cu@CuS, and (g and h) CMF/CuS. Inset images in (d), (f) and (h) are the TEM images of Cu, Cu@CuS and CuS, respectively.

Strong and extensive light absorption is a prerequisite for efficient photothermal WE and TE generation. The optical properties of CMF/Cu@CuS were measured using a UV-vis-NIR spectrophotometer. For comparison, pure MF, CMF, CMF/Cu and CMF/CuS were also analyzed. As shown in Fig. 3a, MF foam has very low absorbance in the entire UV-vis-NIR region. After carbonization, the CMF foam enhanced the light absorption in a wide range. The CMF/CuS compound had an optical absorption in the visible and near-IR region owing to the bandgap (∼2.0 eV) absorption and LSPR of free holes in CuS.⁵¹ CMF/Cu@CuS exhibited the strongest absorption in the near-IR region due to the interaction of the LSPR of Cu and CuS.^48,57 The particles of Cu@CuS had a large number of different shapes and diameters, and thus their plasmon oscillations cover a wide frequency range.⁵⁸ CMF/Cu@CuS exhibited the strongest light absorption performance in both the visible and near-infrared regions, which could enhance the PCE and facilitate a faster WE rate and higher TE power output. In addition, the water-absorbing ability results indicated that the prepared CMF/Cu@CuS has excellent water adsorption capability (Fig. 3b and S4†). The mass of the dry MF was 0.6086 g. The dry MF was completely soaked in deionized water, and then taken out. The mass of MF after soaking was 66.8053 g. Thus, the water-absorbing ability of MF is 66.1967 g. The same steps were performed to test the water-adsorption capability of CMF/Cu@CuS, which was determined to be 67.2682 g.

Fig. 3 (a) UV-vis-NIR absorption spectra of MF, CMF, CMF/Cu, CMF/CuS and CMF/Cu@CuS. (b) Water-absorbing ability of pure MF and CMF/Cu@CuS composite.

During the use of a PCM, various operations such as assembly and washing are performed frequently on it. Therefore, the flexibility, stretchability and strength of PCMs are very important. Both the pure CMF and CMF/Cu@CuS composite could be folded and unfolded several times without breaking or being damaged, indicating that the foam is free-standing, flexible, and elastic (Fig. S5†). Consequently, the solid impurities accumulated on the foam could be easily washed away by ordinary hand washing processes after WE. The compression and tensile testing results of the prepared CMF/Cu@CuS are shown in Fig. S6.† The CMF/Cu@CuS foam had good compression and tensile properties, which could bear a weight of 200 g without prominent deformation.

Thus, the above-mentioned results indicated that CMF with a rich pore structure can be obtained through the rapid heat treatment of MF. On the one hand, after rapid heat treatment, CMF still had loose porous structures, which are conducive to water transfer and light scattering.^25–27 On the other hand, the mechanical properties of CMF, such as flexibility, compressibility and tensile property, were guaranteed, which are conducive to the subsequent operation. Subsequently, Cu@CuS with the LSPR effect and excellent UV-vis light absorption performance as a sensitizer was used to modify the flexible CMF as a PCM, thus fully improving the utilization rate of visible light and near-infrared light. It is expected that all the above-mentioned merits of the CMF/Cu@CuS composite will make it promising and suitable for solar-driven WE and TE power generation applications.

3.2. Single WE performance of 2D and 3D devices

Wide and strong light absorption usually results in an excellent photothermal conversion performance. Thus, to evaluate the photothermal conversion performance, single WE devices was designed. The top view of the device and the length, width and height of the CMF/Cu@CuS foam as the PCM are presented in Fig. S7.† Both the length and width of the WE device were 5 cm, and the thickness of the 3D and 2D CMF/Cu@CuS foam was 3 cm and 0.3 cm, respectively. In addition, it is well known that the thermal conductivity of materials can seriously affect heat transfer.⁵⁹ Large thermal conductivity will cause rapid heat dissipation from the PCM to bulk water, which will reduce the WE rate. It was found that the thermal conductivity of the PS foam is very low (0.08 W m⁻¹ K⁻¹), and thus it can be used as a thermal insulation layer for photothermal WE. We wrapped PS with gauze as the hydrophobic layer, and then put the CMF/Cu@CuS foam on the PS foam as the light-absorbing layer. The 3D or 2D CMF/Cu@CuS floated on the water and its top surface kept moist by spontaneously pumping water from the bottom.

The solar-to-vapor generation ability of the setup for tap water was studied. The surface temperatures of 3D CMF, CMF/Cu, CMF/CuS and CMF/Cu@CuS foam were recorded using an IR camera under 1.0 kW m⁻² light illumination and the results are shown in Fig. 4a. Obviously, the CMF/Cu@CuS foams showed the highest surface temperature, which increased faster than that of the other foams. After 60 min, the average surface temperature of CMF/Cu@CuS reached 42.2 °C (the average surface temperature of CMF, CMF/Cu and CMF/CuS was 37.8 °C, 39.2 °C and 41.1 °C, respectively). As the intensity of the light increased, the surface temperature became higher (Fig. 4b). The average top-surface temperature of the CMF/Cu@CuS foams increased to 46.1 °C at 1.5 kW m⁻² and 48.5 °C at 2 kW m⁻² light illumination (Fig. S8†). This high temperature increase can be attributed to the enhanced light absorption of Cu@CuS, confirming that Cu@CuS in CMF can quickly and efficiently convert light to heat.

Fig. 4 (a) Surface temperature changes of 3D CMF, CMF/Cu, CMF/CuS, and CMF/Cu@CuS under 1.0 kW m⁻² light illumination. (b) Surface temperature changes of 3D CMF/Cu@CuS under different light irradiation. (c) Mass changes of pure water, 3D MF, CMF, CMF/Cu, CMF/CuS and CMF/Cu@CuS under 1.0 kW m⁻² light illumination. (d) WE rate in the dark and 1.0 kW m⁻² light illumination using 2D and 3D devices with CMF/Cu@CuS as the PCM. (e) Comparison of WE performance with previous reports using CMF, Cu or CuS as the PCMs. (f) Efficiency stability of the 3D and 2D CMF/Cu@CuS devices.

The WE rate and PCE could be accurately checked by recording the change in the mass of the water in a container under light illumination. The typical curves of the water mass change over time for the various PCMs are shown in Fig. 4c. The WE of the blank water was similar to that of the pure MF. In contrast, CMF, CMF/Cu, CMF/CuS and CMF/Cu@CuS exhibited satisfactory WE performances. Among them, the 3D CMF/Cu@CuS showed the best WE performance. To systematically evaluate the influence of the device dimensions on the WE performance, we measured and calculated the WE rate of the 2D and 3D CMF/Cu@CuS. The WE rate and PCE (η) calculations are described in detail in the ESI (2).† PCE and WE rate calculation, and the results are shown in Table 1 and Fig. 4d. For the 2D CMF/Cu@CuS device, the single WE rate was calculated to be 1.48 kg m⁻² h⁻¹ and the PCE was 77.2%. When the 3D CMF/Cu@CuS was used as the PCM, a steady-state WE rate up to 1.86 kg m⁻² h⁻¹ was achieved, corresponding to a PCE of 89.6% (Fig. S9†). The recently reported of WE rate of 3D evaporators using CMF,^25–27 modified-MF,^28–30 Cu^47,48,50 and CuS^51,54–56 as PCMs is summarized and compared in Fig. 4e. It can be seen that although our results are not optimal, they exceed most of the reported devices in terms of WE rate and PCE. Even the 2D devices outperformed some of the reported 3D devices in terms of WE and PCE. This indicates that the use of Cu particles with LSPR effect and CuS with excellent UV-vis light absorption ability to sensitize CMF can greatly improve the absorption of light of materials and improve the PCE.

Table 1 Zero-order kinetic equations, single WE rate and PCE of CMF/Cu@CuS as the PCM under 1.0 kW m⁻² light illumination

Zero-order kinetic equation	WE rate (kg m^−2 h^−1)	PCE (η, %)	Different devices
y = 1.86249x − 0.00956	1.86	89.6	3D-WE
y = 1.48307x − 0.00565	1.48	77.2	2D-WE

---

Usually, service life is an important indicator of absorber performance. Therefore, the entire system was placed under 1 kW m⁻² light illumination for 8 cycles (one cycle for an hour). The solar-to-vapor WE rate of the 3D CMF/Cu@CuS foam remained above 1.84 kg m⁻² h⁻¹ during 8 cycles (Fig. 4f). In addition, after 8 cycles, the CMF/Cu@CuS foam maintained its original elasticity, suggesting its long service life. The impurities on the surface of the foam could be washed with distilled water and it could be reused.

The above-mentioned results show that when the same PCM is used as the light absorption layer for photothermal WE, the dimensions of the PCM has a great influence on the WE rate and PCE. It has been reported that 3D devices can obtain energy from the warm surrounding environment during the photothermal WE process.^{19,22,60–63} Thus, to determine whether the 3D CMF/Cu@CuS device can absorb additional heat from the surrounding environment, we tested the temperature of different parts of the device and a schematic diagram of the 2D and 3D devices is presented in Fig. 5a. Fig. 5b shows the change in the lateral, ambient and surface temperature of the 3D device with illumination time under 1.0 kW m⁻² light. The inserted pictures show the test points of the lateral and ambient temperature. The temperature of the different parts of the device and the ambient temperature increased with the extension of the illumination time. However, the ambient temperature of the 3D CMF/Cu@CuS device was higher than the lateral temperature, which indicates that when the 3D CMF/Cu@CuS foam is used as the PCM, its lateral area can obtain extra energy from the ambient environment during the WE process. Fig. 5c shows the surface infrared images of the 3D and 2D devices during the WE process. It is obvious that the surface temperature of the 3D structure was lower than that of the 2D structure. This is because the 3D devices have a larger lateral surface area than the 2D devices, providing a greater cold evaporation area. The average temperature of the top surface can be reduced by increasing the lateral surface area, resulting in a higher surface temperature in the 2D structure than 3D structure.²³ In the process of photothermal WE, heat exchange occurs between the PCMs in bulk water and the environment through heat convection, radiation and conduction, and the dimensions of the device have an important influence on the heat exchange.^62,63 A lower surface temperature of the 3D structure can reduce the energy loss of the top evaporation surface under light irradiation. In addition, the lateral infrared images results of the 3D and 2D devices in the WE process further evidence that the lateral temperature of the 3D structure was significantly lower than the ambient temperature (Fig. 5d). However, the lateral area of the 2D device is very small, and thus it can be assumed that it does not absorb heat from the environment. Heat exchange exists only on the surface of the 2D device, and the higher the surface temperature, the more heat will be dissipated into the environment and the bottom bulk water, resulting in a lower WE rate and PCE than the 3D device.

Fig. 5 (a) Schematic diagram of 2D and 3D devices and heat exchange during the WE process. (b) Temperature change using 3D CMF/Cu@CuS as the PCM under 1.0 kW m⁻² light illumination (insets are the test point for lateral and ambient temperature). (c) Surface infrared images of the 2D and 3D CMF/Cu@CuS under 1.0 kW m⁻² light illumination for 5, 10, 30 and 60 min. (d) Lateral infrared images of the 3D and 2D CMF/Cu@CuS under 1.0 kW m⁻² light illumination for 3 and 60 min.

The photothermal conversion WE process includes light absorption, water transfer, heat convection and radiation, etc.^62,63 The porous PCMs also have an effect on the enthalpy of WE. Thus, to fully consider the heat transfer of the porous CMF/Cu@CuS structure in the WE process, a detailed calculation of the heat loss was carried out, which is presented in the ESI (3) Heat loss analysis.† We performed differential scanning calorimetry (DSC) experiments and demonstrated the enthalpy of vaporization on the relative temperature compared with that of the bulk water (Fig. S10†).²⁶ For 3D devices, the lateral temperature is lower than the ambient temperature and the lateral part of the device will draw heat from the ambient environment (Fig. 5b and d). Therefore, when calculating the efficiency of 3D and 2D devices, not only the heat loss to the ambient environment, but also the heat gained from the ambient environment should be considered. The net energy (E_ambient) gained from the ambient environment can be estimated using the following equation:²⁹

E_ambient = −A₁εσ(T₁⁴ − T_E⁴) − A₂εσ(T₂⁴ − T_E⁴) − A₁h(T₁ − T_E) − A₂h(T₂ − T_E) − CmΔT

where A₁ (25 cm²) is the surface area of the device and A₂ (60 cm²) is the lateral area of the 3D device. The 2D device did not exchange heat with the environment due to its small lateral area and higher temperature than the surrounding environment. ε is the emissivity of the material, which is assumed to be the maximum emissivity of 1, σ is the Stefan–Boltzmann constant (5.67 × 10⁻⁸ W m⁻² K⁻⁴), T₁ is the average surface temperature, and T₂ is the lateral temperature of the materials at a solar power density of 1.0 kW m⁻². T_E is the ambient temperature at a solar power density of 1.0 kW m⁻². h is the convection heat transfer coefficient (10 W m⁻² K⁻¹). C is the specific heat capacity of water (4.2 J °C⁻¹ g⁻¹), m (175 g) is the weight of the bulk water used in the experiment, and ΔT represents the temperature change before and after the WE of the bulk water. The heat losses of the 3D and 2D devices at 1.0 kW m⁻² are presented in Table 2. The results indicate that the 3D CMF/Cu@CuS device had a higher net energy than the 2D device, which is because the lateral side of the 3D structure could absorb extra heat from the surrounding environment. These results further demonstrate that it is possible to obtain additional energy from the surrounding environment to improve the PCE and WE rate when a 3D-structured evaporator is employed during photothermal WE.^62,63

Table 2 Heat loss of the 3D and 2D devices at a solar power density of 1.0 kW m⁻²

	Top-surface radiation loss	Lateral heat extraction	Top-surface convection	Lateral convection	Heat conduction	Net energy
3D CMF/Cu@CuS	−0.0998 W	0.0755 W	−0.0148 W	0.0282 W	−0.3675 W	−0.3784 W
2D CMF/Cu@CuS	−0.1743 W	—	−0.252 W	—	−0.245 W	−0.6713 W

---

3.3. Simultaneous WE and TE energy generation performance of 2D and 3D devices

In addition to using solar energy to obtain clean water, TE energy generation through the Seebeck effect is also another important application.^35,39,41,43 It is worth noting that a temperature difference is formed between the surface of the PCM and the bulk water during water evaporation. This temperature difference enables the system to generate TE power based on the Seebeck effect.^35,41,43 Therefore, we integrated TE generation ability with WE. A photograph of the TE model is shown in Fig. S1.† The top of the TE module is covered by 3D or 2D CMF/Cu@CuS foam, thereby generating a hot end by localized heat (Fig. S11†). The TE module was placed on the supporting PS foam and wrapped in gauze to making the device stably float on the water surface and draw water into the cold side of the TE module. This facilitated a wider temperature difference and generated a higher TE power.

The TE power generation performance was tested during the WE process. As shown in Fig. 6a and b, once the light was turned on, both the output voltage and current rapidly increased until they reached a steady state. Under 1.0 kW m⁻² light illumination, the output voltage was about 6.5 mV and 66.8 mV and the current was 5.9 mA and 16.6 mA for the 3D and 2D CMF/Cu@CuS, respectively. Once the light was turned off, both the output current and voltage decreased, but they did not disappear immediately. In addition, we further studied the influence of light intensity on the output current and voltage. The results indicate that as the light intensity increased, the output current and voltage also increased. The output voltage was about 7.4 mV and 84.2 mV and the current was 6.7 mA and 18.2 mA at 1.5 kW m⁻², and 9.4 mV and 99.5 mV and 7.5 mA and 19.8 mA at 2 kW m⁻² for the 3D and 2D CMF/Cu@CuS foam, respectively. The maximum output power of the 2D and 3D CMF/Cu@CuS under different light intensities was calculated. As shown in Fig. 6c, the maximum output power of the 2D CMF/Cu@CuS device reached 69.3 μW cm⁻², 95.78 μW cm⁻², and 123.13 μW cm⁻² under 1.0 kW m⁻², 1.5 kW m⁻² and 2 kW m⁻² light illumination, which is higher than that of the 3D CMF/Cu@CuS device of 2.4 μW cm⁻², 3.1 μW cm⁻², and 4.41 μW cm⁻² under 1.0 kW m⁻², 1.5 kW m⁻² and 2 kW m⁻² light illumination, respectively. Thus, these results fully demonstrate that the 2D-structured device is beneficial for TE generation when the same PCM is used in the photothermal WE process.

Fig. 6 (a and b) Output voltages and current when the light was turned on and off under different light irradiation intensities. (c) Maximum output power under different light illumination intensities. (d) WE rate of the single WE device and integrated WE/TE device under 1.0 kW m⁻² light illumination. (e) WE rate under different light illumination using the 2D and 3D device. (f) Temperature difference between the hot and cold sides of the TE module in the 2D and 3D devices.

Obviously, while generating TE power, WE also occurred. Thus, we compared the WE rate of a single WE device and integrated a WE/TE device under different light illumination (Fig. 6d and e). Interestingly, for both the 2D and 3D devices, the WE rate of the integrated WE/TE devices was lower than that of the single WE devices. In the case of the 3D devices, the single WE rate was 1.86 kg m⁻² h⁻¹, corresponding to a PCE of 89.6%, while the WE rate of the integrated 3D WE/TE device was 1.85 kg m⁻² h⁻¹, corresponding to a PCE of 86.8%. The PCE of the single WE and integrated WE/TE devices decreased by 2.8%. In the case of the single 2D WE devices, the WE rate was 1.48 kg m⁻² h⁻¹, corresponding to a PCE of 77.2%, while that of the integrated WE/TE device was 1.32 kg m⁻² h⁻¹, corresponding to a PCE of 68.3%. The PCE was reduced by 8.9%, but the TE generation greatly improved. These results indicated that the dimensions of the PCMs had an effect on WE and TE generation when the TE module was located at the bottom of the PCM. In addition, the WE rate increased with an increase in the light intensity. Under 1.5 sun light irradiation, the WE rate of the 2D-structured device reached 1.63 kg m⁻² h⁻¹ and 2.15 kg m⁻² h⁻¹ for the 3D-structured device. Under 2 sun light intensity, the WE rate of the 2D-structured device reached 1.98 kg m⁻² h⁻¹, while that of the 3D structure was as high as 2.51 kg m⁻² h⁻¹ (Fig. 6e).

According to the TE energy generation mechanism through the Seebeck effect, it should be caused by the temperature difference between the hot side and the cold side of the TE module. Thus, we measured the temperature difference between the hot side and cold side of the TE module, and the results are shown in Fig. 6f. The bottom end of the CMF/Cu@CuS was connected to the hot side of the TE module. In the case of the 3D CMF/Cu@CuS foam, the temperature difference between the top end and the bottom end was large, and thus the temperature difference between the hot side and the cold side of the TE module was small. Conversely, the temperature difference between the two ends of TE module was large in the 2D structure. When the 2D CMF/Cu@CuS was used as the PCM, the temperature difference between the hot side and cold side of the TE module was very large, which is conducive to generating TE power. However, the lateral area of the 2D device could hardly absorb heat from the environment, and thus it is not conducive to WE. In the 3D CMF/Cu@CuS structure, the temperature difference between the hot side and cold side of the TE module was very small, and thus it was unfavorable to generate TE power. However, because the lateral temperature is lower than the ambient temperature, it can gain extra energy from the ambient environment, and thus the change in the WE rate between the single 3D WE and integrated WE/TE devices was very small.³⁵

In addition, the WE rate of both the 2D and 3D devices was lower than that of the single WE when WE and TE power generation occurred simultaneously. Especially for the 2D device, the WE rate greatly decreased. We believe that this is attributed to two reasons. Firstly, after the TE module was added, the bottom of the PCM could not fully contact with the water, only depending on the gauze around the TE module for water transmission, which reduced the water transmission area and water supply capacity. Secondly, the PCM layer of the 2D structure device was thin and its upper surface temperature is higher than that of the 3D structure device, and thus the heat dissipation is quick, resulting in large heat loss in the 2D structure devices.²³ Therefore, the WE rate of the 2D structure devices decreased greatly.

The above-mentioned results indicate that the flexible CMF/Cu@CuS PCM and well-designed devices could generate TE power during the WE process under solar irradiation. The output TE power of the 2D device was significantly higher than that of the 3D device, but its WE rate was lower than that of the 3D device when the TE module was placed at the bottom of the PCM. Fig. 7a presents a schematic of the simultaneous WE and TE power generation. PS foam with good thermal insulation properties was used as the insulation layer, and flexible CMF/Cu@CuS as the PCM. To transport water to the CMF/Cu@CuS layer continuously, the PS foam was wrapped with gauze, which not only could transport the water from the bottom to the upper but also effectively reduce the heat loss. In the semi-closed system, the temperature of the CMF/Cu@CuS layer increased as the duration of illumination increased, while the temperature at the bottom of the bulk water changed slightly due to the low heat conductivity of the PS foam.

Fig. 7 (a) Schematic of simultaneous WE and TE power generation. (b) Different temperatures of the 3D and 2D devices.

The Seebeck effect is a TE phenomenon of the voltage difference between two different electrical conductors or semiconductors caused by temperature differences. The larger the temperature difference on the hot side and cold side of the TE module, the higher the voltage and current output. As shown in Fig. 6f and 7b, for the 3D and 2D structures, the surface temperature (T₁) increased due to the absorption of solar energy, respectively. Due to the low thermal conductivity of CMF, the temperature of the lower surface (T₂) was significantly lower than T₁ when the 3D structure was used. In the case of the 2D structure, due to its thin thickness, the temperature of the lower surface (T₂) did not differ much from T₁, and thus ΔT_2D was explicitly larger than ΔT_3D, resulting in higher voltage and higher current output. In the case of the 3D-structured device, given that the temperature of its lateral side may be lower than the ambient temperature, it could absorb additional heat from the ambient environment during the WE process, thus accelerating the WE. Consequently, its WE rate was higher than that of the 2D-structured device, but it could not generate a high voltage and large current. Therefore, in the current device structure, it is difficult to simultaneously obtain a high WE rate and high TE power. Specifically, you cannot have “one stone two birds”, only “you cannot have your cake and eat it too” when the TE module is placed at the bottom of the PCM.

To verify this mechanism, a PCM with a thickness of 1.8 cm was selected for simultaneous WE and TE power generation. The experimental setup is shown in the inset picture of Fig. 8a, where TE module was placed at the bottom of the PCM. Under 1.0 kW m⁻² light irradiation, the WE rate reached 1.78 kg m⁻² h⁻¹, which was larger than that of the device with a thickness of 0.3 cm. After about 4 min of illumination, the device maintained a stable voltage and current output with an output voltage of 58.3 mV (ESI: Movie 1†), and a current of 10.56 mA (ESI: Movie 2†), and the output power reached 38.48 μW cm⁻². Although the output power was lower than that of the 0.3 cm device (69.3 μW cm⁻²), it was much higher than the 3 cm device (2.4 μW cm⁻²). By measuring the temperature of the hot side and the cold side of the TM module, the results show that with the prolonging of the illumination time, the temperature difference between the hot side and the cold side of the TE module gradually increased, and finally stabilized at about 7 °C (Fig. 8b), which is higher than that of the 3 cm device, but lower than that of the 0.3 cm device (Fig. 6f). This result fully demonstrates the rationality of the mechanism, as shown in Fig. 7. The infrared image shows that compared with the 3D structure device with a thickness of 3 cm, the lateral temperature of the PCM with a 1.8 cm thickness had little difference from the ambient temperature (Fig. S12†), and thus less heat was obtained from the environment and the WE rate was lower than that of the 3 cm device.

Fig. 8 (a) WE rate under 1.0 kW m⁻² light irradiation by the device with a thickness of 1.8 cm (inset: photograph of set-up). (b) Temperature difference between the hot and cold sides of the TE module for the 1.8 cm PCM. (c) Photographs of the experimental setup with the TE module near the upper surface of the PCM. (d) Temperature difference between the hot and cold sides of the TE module (inset: temperature measurement set-up).

The above-mentioned results indicate that the thicker the PCM, the better of WE rate because additional heat can be obtained from the environment. To obtain a large output power, the temperature difference between the hot side and the cold side of the TE module should be as large as possible. If the TE module is placed at the bottom of the PCM, the thicker PCM is not conducive to generating a large temperature difference between the hot side and cold side of the TE module, and thus a large output power could not be obtained. When the TE module was placed near the top surface of the PCM, not only the temperature difference between the hot side and the cold side of the TE module increased but also the PCM could still obtain additional energy from the surrounding environment, thereby increasing the output TE power, and simultaneously having a fast WE rate. Thus, to verify this idea, we assembled a device using a 3 cm thick PCM, but placed the TE module near the top surface of the PCM for simultaneous photothermal WE and TE power generation. The experimental setup is shown in Fig. 8c. The thickness of the PCM on the upper part of the TE module was 0.5 cm. The results show that under 1.0 kW m⁻² light irradiation, not only a high WE rate was achieved but also a large output power was obtained. The WE rate reached 1.84 kg m⁻² h⁻¹ (ESI: Movie 3†). The WE rate decreased slightly when the TE module was placed near the upper surface of the PCM compared to placing the TE module at the bottom of the PCM. We believe that this is mainly due to the fault zone in the PCM, which leads to a reduction in the water transmission area. In addition, the TE module absorbs a large amount of heat, which is one of the other reasons why the WE rate decreased.

When the TE module was placed near the upper surface of the PCM, the device could sustain a stable voltage and current output after about 5 min of illumination. The output voltage reached 286.6 mV (ESI: Movie 4†), the current reached 50 mA (ESI: Movie 5†), and the output power reached 895.62 μW cm⁻². This is much higher than the result when the TE module was placed at the bottom of the PCM. The temperature of the upper layer and bottom layer of the TE module were measured, and the findings revealed that as the illumination period was extended, the temperature difference between the two layers quickly increased and stabilized at about 17.3 °C (Fig. 8d; the inset is the temperature measurement photograph). Due to the large temperature difference between the upper layer and the bottom layer of the TE module, a higher TE output power could be achieved. This result indicates that with the proper assembly of the device, the function of “one stone two birds” can be realized.

A summary of the performance of other PCMs for simultaneous WE and TE power generation is presented in Table S1.† Although our results are not optimal, they have unique advantages. In comparison with the reported results, the merits of our work are as follows: (1) the CMF obtained by rapid heat treatment of MF not only can maintain a loose porous structure, but also facilitate water transfer. In addition, the mechanical properties of CMF, such as elasticity, compressibility and tensile properties, are ensured, which are conducive to the subsequent operation. (2) The Cu particles on the CMF skeleton contribute to the loading and uniform growth of CuS, and thus the uniform Cu@CuS particles were embedded in CMF. The 3D porous structure of the flexible CMF/Cu@CuS can be used as a PCM to make full use of incident light and improve the utilization rate of visible light and near-infrared light. (3) Most importantly, the lateral temperature of this 3D structure is lower than the ambient temperature, and thus additional energy can be obtained from the surrounding environment, improving the WE rate and PCE.

4. Conclusion

In this contribution, we developed a super-hydrophilic and light absorption property of flexible CMF/Cu@CuS as a PCM for simultaneous WE and TE power generation. The CMF/Cu@CuS had effective full-spectrum absorption, excellent photothermal conversion performance, and good hydrophilicity. We found that for the same PCM, the WE rate of the integrated WE/TE devices as less than that of the single WE devices. When the device realized simultaneous WE and TE power generation, the surface temperature of the device was slightly lower than that of the single WE system. It is advantageous to get more heat from the environment and limit heat loss when the surface temperature is lower. Furthermore, “one stone two birds” could not be realized when the TE module was placed at the bottom of the PCM; instead, it was only “you can't have your cake and eat it too”. When the TE module was placed close to the upper surface of the 3D device, it could simultaneously achieve a fast WE rate and high output TE power, fulfilling the “one stone, two birds” function. This research can aid in the development of devices that generate both WE and TE power by photothermal conversion.

Author contributions

Haiwen Wang and Yanying Shi: writing – original draft, investigation, data curation, validation. Tiefeng Liu and Xiuwen Zheng: investigation, and validation. Shanmin Gao and Jun Lu: conceptualization, supervision, project administration, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors appreciate the support from the National Natural Science Foundation of China (22075122), Natural Science Foundation of Shandong Province (ZR2019MB019) and the Research Foundation for Talented Scholars of Linyi University (Z6122010).

Notes and references

1. W. L. Liu, Z. Wang, G. Wang, G. L. Liu, J. Chen, X. J. Pu, Y. Xi, X. Wang, H. Y. Guo, C. G. Hu and Z. L. Wang, Nat. Commun., 2019, 10, 1426.
2. G. B. Xue, Y. Xu, T. P. Ding, J. Li, J. Yin, W. W. Fei, Y. Z. Cao, J. Yu, L. Y. Yuan, L. Gong, J. Chen, S. Z. Deng, J. Zhou and W. L. Guo, Nat. Nanotechnol., 2017, 12, 317–321.
3. L. F. Greenlee, D. F. Lawler, B. D. Freeman, B. Marrrot and P. Moulin, Water Res., 2009, 43, 2317–2348.
4. D. Kraemer, B. Poudel, H. P. Feng, J. C. Caylor, B. Yu, X. Yan, Y. Ma, X. W. Wang, D. Z. Wang, A. Muto, K. McEnaney, M. Chiesa, Z. F. Ren and G. Chen, Nat. Mater., 2011, 10, 532–538.
5. Y. T. Wu, H. Q. Huang, W. M. Zhou, C. T. You, H. L. Ye, J. Chen, S. Zang, J. H. Yun, X. Q. Chen, L. W. Wang and Z. H. Yuan, ACS Appl. Mater. Interfaces, 2022, 14, 29099–29110.
6. C. H. Huang, J. X. Huang, Y. H. Chiao, C. M. Chang, W. S. Hung, S. J. Lue, C. F. Wang, C. C. Hu, K. P. Lee, H. H. Pan and H. Y. Lai, Adv. Funct. Mater., 2021, 31, 2010422.
7. X. H. Han, S. Q. Ding, H. Y. Hu and S. R. Wang, J. Mater. Chem. A, 2022, 10, 18509–18541.
8. L. L. Zhu, M. M. Gao, C. K. N. Peh and G. W. Ho, Nano Energy, 2019, 57, 507–518.
9. X. Z. Min, B. Zhu, B. Li, J. L. Li and J. Zhu, Acc. Mater. Res., 2021, 2, 198–209.
10. L. B. Zhang, B. Tang, J. B. Wu, R. Y. Li and P. Wang, Adv. Mater., 2015, 27, 4889–4894.
11. Y. C. Wang, C. Z. Wang, X. J. Song, S. K. Megarajan and H. Q. Jiang, J. Mater. Chem. A, 2018, 6, 963–971.
12. W. Fang, L. Zhao, X. He, H. Chen, W. X. Li, X. H. Zeng, X. D. Chen, Y. Shen and W. H. Zhang, Renewable Energy, 2020, 151, 1067–1075.
13. C. J. Chen, Y. D. Kuang and L. B. Hu, Joule, 2019, 3, 1–36.
14. D. N. Thoai, Q. T. H. Ta, T. T. Truong, H. V. Nam and G. V. Vo, J. Clean. Prod., 2021, 293, 126122.
15. C. X. Li, S. J. Cao, J. Lutzki, J. Yang, T. Konegger, F. Kleitz and A. Thomas, J. Am. Chem. Soc., 2022, 144, 3083–3090.
16. X. Su, D. Z. Hao, M. Y. Sun, T. S. Wei, D. K. Xu, X. C. Ai, X. L. Guo, T. Zhao and L. Jiang, Adv. Funct. Mater., 2022, 32, 2108135.
17. X. P. Li, X. F. Li, H. G. Li, Y. Zhao, J. Wu, S. K. Yan and Z. Z. Yu, Adv. Funct. Mater., 2022, 32, 2110636.
18. Y. J. Hu, H. Y. Ma, M. M. Wu, T. Y. Lin, H. Z. Yao, F. Liu, H. H. Cheng and L. T. Qu, Nat. Commun., 2022, 13, 4335.
19. P. Sun, W. Zhang, I. Zada, Y. X. Zhang, J. J. Gu, Q. L. Liu, H. L. Su, D. Pantelić, B. Jelenković and D. Zhang, ACS Appl. Mater. Interfaces, 2019, 12, 2171–2179.
20. Y. Bian, Q. Q. Du, K. Tang, Y. Shen, L. C. Hao, D. Zhou, X. K. Wang, Z. H. Xu, H. L. Zhang, L. J. Zhao, S. M. Zhu, J. D. Ye, H. Lu, Y. Yang, R. Zhang, Y. D. Zheng and S. L. Gu, Adv. Mater. Technol., 2019, 4, 1800593.
21. Y. Y. Shi, C. F. Zhang, Y. H. Wang, Y. M. Cui, Q. Y. Wang, G. J. Liu, S. M. Gao and Y. F. Yuan, Desalination, 2021, 507, 115038.
22. Y. Shi, R. Y. Li, Y. Jin, S. F. Zhuo, L. Shi, J. Chang, S. Hong, K. C. Ng and P. Wang, Joule, 2018, 2, 1171–1186.
23. B. Shao, Y. Wang, X. Wu, Y. Lu, X. F. Yang, G. Y. Chen, G. Owens and H. L. Xu, J. Mater. Chem. A, 2020, 8, 11665–11673.
24. M. M. Gao, L. L. Zhu, C. K. Peh and G. W. Ho, Energy Environ. Sci., 2019, 12, 841–864.
25. P. F. Liu, L. Miao, Z. Y. Deng, J. H. Zhou, Y. F. Gu, S. Y. Chen, H. F. Cai, L. X. Sun and S. Tanemura, Appl. Energy, 2019, 241, 652–659.
26. F. Gong, H. Li, W. B. Wang, J. G. Huang, D. X. Xia, J. X. Liao, M. Q. Wu and D. V. Papavassiliou, Nano Energy, 2019, 58, 322–330.
27. X. F. Lin, J. Y. Chen, Z. K. Yuan, M. J. Yang, G. J. Chen, D. S. Yu, M. Q. Zhang, W. Hong and X. D. Chen, J. Mater. Chem. A, 2018, 6, 4642–4648.
28. X. Zhao, X. J. Zha, J. H. Pu, L. Bai, R. Y. Bao, Z. Y. Liu, M. B. Yang and W. Yang, J. Mater. Chem. A, 2019, 7, 10446–10455.
29. S. Meng, X. Zhao, C. Y. Tang, P. Yu, R. Y. Bao, Z. Y. Liu, M. B. Yang and W. Yang, J. Mater. Chem. A, 2020, 8, 2701–2711.
30. C. W. Li, D. G. Jiang, B. B. Huo, M. C. Ding, C. C. Huang, D. D. Jia, H. X. Li, C. Y. Liu and J. Q. Liu, Nano Energy, 2019, 60, 841–849.
31. P. H. Yang, K. Liu, Q. Chen, J. Li, J. J. Duan, G. B. Xue, Z. S. Xu, W. K. Xie and J. Zhou, Energy Environ. Sci., 2017, 10, 1923–1927.
32. X. Q. Li, X. Z. Min, J. L. Li, N. Xu, P. C. Zhu, B. Zhu, S. N. Zhu and J. Zhu, Joule, 2018, 2, 2477–2484.
33. L. L. Zhu, M. M. Gao, C. K. N. Peh, X. Q. Wang and G. W. Ho, Adv. Energy Mater., 2018, 8, 1702149.
34. P. Xiao, J. He, F. Ni, C. Zhang, Y. Liang, W. Zhou, J. C. Gu, J. Y. Xia, S. W. Kuo and T. Chen, Nano Energy, 2020, 68, 104385.
35. L. L. Zhu, T. P. Ding, M. M. Gao, C. K. N. Peh and G. W. Ho, Adv. Energy Mater., 2019, 9, 1900250.
36. Y. X. Zhang, S. K. Ravi and S. C. Tan, Nano Energy, 2019, 65, 104006.
37. Z. M. Zheng, H. Y. Li, X. D. Zhang, H. Jiang, X. M. Geng, S. M. Li, H. Y. Tu, X. R. Cheng, P. Yang and Y. F. Wan, Nano Energy, 2020, 68, 104298.
38. Y. Xu, Z. Z. Guo, J. Wang, Z. H. Chen, J. C. Yin, Z. X. Zhang, J. M. Huang, J. W. Qian and X. B. Wang, ACS Appl. Mater. Interfaces, 2021, 13, 27129–27139.
39. Y. Y. Cui, J. Liu, Z. Q. Li, M. Y. Ji, M. Zhao, M. H. Shen, X. Han, T. Jia, C. L. Li and Y. Wang, Adv. Funct. Mater., 2021, 31, 2106247.
40. X. J. Mu, J. H. Zhou, P. F. Wang, H. Chen, T. T. Yang, S. Y. Chen, L. Miao and T. Mori, Energy Environ. Sci., 2022, 15, 3388–3399.
41. J. X. Ren, Y. Ding, J. Gong, J. P. Qu and R. Niu, Energy Environ. Mater., 2022 DOI:10.1002/eem2.12376.
42. P. P. He, H. Y. Bai, Z. F. Fan, L. Hao, N. Liu, B. Y. Chen, R. Niu and J. Gong, J. Mater. Chem. A, 2022, 10, 13378–13392.
43. Z. F. Fan, J. X. Ren, H. Y. Bai, P. P. He, L. Hao, N. Liu, B. Y. Chen, R. Niu and J. Gong, Chem. Eng. J., 2023, 451, 138534.
44. K. Bae, G. Kang, S. Y. Cho, W. Park, K. Kim and W. J. Padilla, Nat. Commun., 2015, 6, 10103.
45. Z. Y. Sun, J. J. Wang, Q. L. Wu, Z. Y. Wang, Z. Wang, J. Sun and C. J. Liu, Adv. Funct. Mater., 2019, 29, 1901312.
46. L. Zhou, Y. L. Tan, J. Y. Wang, W. C. Xu, Y. Yuan, W. S. Cai, S. N. Zhu and J. Zhu, Nat. Photonics, 2016, 10, 393–398.
47. J. J. Xu, F. Xu, M. Qian, Z. Li, P. Sun, Z. L. Hong and F. Q. Huang, Nano Energy, 2018, 53, 425–431.
48. L. T. Ren, X. L. Yi, Z. S. Yang, D. F. Wang, L. Q. Liu and J. H. Ye, Nano Lett., 2021, 21, 1709–1715.
49. Y. W. Lin, Z. H. Chen, L. Fang, M. Meng, Z. F. Liu, Y. S. Di, W. D. Cai, S. S. Huang and Z. X. Gan, Nanotechnology, 2019, 30, 015402.
50. C. Tian, C. C. Li, D. L. Chen, Y. F. Li, L. Xing, X. L. Tian, Y. Cao, W. Huang, Z. X. Liu and Y. J. Shen, J. Mater. Chem. A, 2021, 9, 15462.
51. F. J. Tao, Y. L. Zhang, S. J. Cao, K. Yin, X. T. Chang, Y. H. Lei, R. H. Fan, L. H. Dong, Y. S. Yin and X. B. Chen, Mater. Today Energy, 2018, 9, 285–294.
52. F. Yu, Z. Z. Guo, Y. Xu, Z. H. Chen, M. S. Irshad, J. W. Qian, T. Mei and X. B. Wang, ACS Appl. Mater. Interfaces, 2020, 12, 57155–57164.
53. Q. H. Zhang, X. M. Yin, C. B. Zhang, Y. M. Li, K. J. Xiang, W. L. Luo and X. Z. Qiao, Small, 2022, 18, 2202867.
54. F. J. Tao, Y. L. Zhang, K. Yin, S. J. Cao, X. T. Chang, Y. H. Lei, D. S. Wang, R. H. Fan, L. H. Dong, Y. S. Yin and X. B. Chen, ACS Appl. Mater. Interfaces, 2018, 10, 35154–35163.
55. M. Y. Shang, N. Li, S. D. Zhang, T. T. Zhao, C. Zhang, C. Liu, H. F. Li and Z. Y. Wang, ACS Appl. Energy Mater., 2018, 1, 56–61.
56. X. Wu, M. E. Robson, J. L. Phelps, J. S. Tan, B. Shao, G. Owens and H. L. Xu, Nano Energy, 2019, 56, 708–715.
57. M. W. Zhu, Y. J. Li, F. J. Chen, X. Y. Zhu, J. Q. Dai, Y. F. Li, Z. Yang, X. J. Yan, J. W. Song, Y. B. Wang, E. Hitz, W. Luo, M. H. Lu, B. Yang and L. B. Hu, Adv. Energy Mater., 2018, 8, 1701028.
58. P. Wang, B. B. Huang, X. Y. Qin, X. Y. Zhang, Y. Dai, J. Y. Wei and M.-H. Whangbo, Angew. Chem., Int. Ed., 2008, 47, 7931–7933.
59. H. Ghasemi, G. Ni, A. M. Marconnet, J. Loomis, S. Yerci, N. Miljkovic and G. Chen, Nat. Commun., 2014, 5, 4449.
60. X. Z. Hu and J. Zhu, Adv. Funct. Mater., 2019, 29, 1907234.
61. L. Wu, Z. C. Dong, Z. R. Cai, T. Ganapathy, N. X. Fang, C. X. Li, C. L. Yu, Y. Zhang and Y. L. Song, Nat. Commun., 2020, 11, 521.
62. Y. Wang, X. Wu, P. Wu, J. Y. Zhao, X. F. Yang, G. Owens and H. L. Xu, Sci. Bull., 2021, 66, 2479–2488.
63. Y. Wang, X. Wu, X. F. Yang, G. Owens and H. L. Xu, Nano Energy, 2020, 78, 105269.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2ta07946a

---