Excellent Thermoelectric Performance Realized in Copper Sulfide Magnetic Nanocomposites Via Modified Solid States Reaction

Abstract

Solid‐state reactions, which are the basic reactions involved in the preparation, processing, and application of materials, are ubiquitous in material science and chemistry research. In the fields of metallurgy and geology, a significant number of complex chemical reactions occur during the smelting of ores. Inspired by the smelting of copper concentrate, this work applies modified metallurgical chemical reactions to the field of (thermoeletric) TE materials. By controlling the reaction temperature and composition, porous copper sulfide magnetic nanocomposites can be formed. Regulating the composition generates numerous precipitates and Cu‐rich phases, and controlling the microstructure facilitates the formation of porous structures. The second phase and porous structure effectively decreased thermal conductivity. Furthermore, the introduction of ferromagnetic Fe₃O₄ particles plays a role in reducing carrier concentration and forming potential barrier scattering low energy carriers, which improves the Seebeck coefficient of the samples. Ultimately, the optimum figure of merit (ZT) of ≈1.3 at 773 K for the Cu_1.8S + 2 wt.% Fe₃O₄ bulk sample and an average ZT of 0.57 over the entire operating temperature range. The modified solid states reaction between oxides and sulfides could be employed to optimize electrical and thermal transport properties for sulfide TE material, as well as sulfide batteries, sulfide photoelectric materials, and sulfide catalytic materials.

Introduction

The new epoch has been developed since the late 20th century because of technological progress in semiconductors, such as transistors, diodes, solar cells, piezoelectric transducers, and TE materials. However, advances in technology are inseparable from the basic chemical reactions. It is precisely because of the study of the ubiquitous chemical reaction that scientific research can be further developed by revealing the mechanism behind the reaction. TE materials enable the conversion of heat to electricity and vice versa, thus facilitating new energy generation and solid-state cooling.^[1] The theoretical sources of TE materials are the Seebeck and Peltier effects,^[2] the optimization process of TE performance is also accompanied by many chemical reactions. Based on this theory, the dimensionless figure of merit (ZT) is used to evaluate the TE performance of materials, namely, ZT = S²σT/κ, where S is the Seebeck coefficient, σ is the electrical conductivity, T is the absolute temperature, and κ is the thermal conductivity. The thermal conductivity (κ) is composed of the electronic thermal conductivity (κ_e) and lattice thermal conductivity (κ_l). The most straightforward strategy for maximizing the TE performance is either enhancing the power factor (PF = S²σ) or reducing the thermal conductivity (κ = κ_e + κ_l), or both at the same time.

Copper sulfide (or selenide) is a new liquid-like TE material that possesses properties consistent with the phonon-liquid electron-crystal (PLEC) concept.^[3] The mobile superionic Cu ions are kinetically disordered in the rigid sublattice formed by sulfur atoms. This trait provides a new strategy for the design of low-thermal conductivity materials. Numerous studies have been conducted on optimizing TE performance. Common strategies include stoichiometric ratio regulation,^[4] element doping,^[5] entropy engineering,^[6] compositing,^[7] ion blocking,^[8] band engineering,^[9] nanoprecipitation,^[10] lattice engineering,^[11] and boundary engineering.^[12] The main response of these methods is to optimize the electrical conductivity or to decrease the thermal conductivity. The limitations of the TE performance of Cu_1.8S are mainly attributed to its low Seebeck coefficient and high thermal conductivity. The primary research objective is to effectively suppress thermal conductivity and simultaneously enhance the Seebeck coefficient by regulating the carrier concentration.

In the field of metallurgy, 82% of copper is smelted via pyrometallurgy, which is accompanied by numerous chemical reactions.^[13] The chalcocite (Cu₂S) is the main mineralogical species in the concentrate, and the regular chemical reaction during pyrometallurgy is as follows:^[14] 1 $$$\begin{equation}C{{u}_2}S\left( s \right) + 2CuO\left( s \right)\mathop \to \limits^\Delta 4Cu\left( s \right) + S{{O}_2}\left( g \right)\end{equation}$$$

In this reaction, the element S was oxidized, the valance of S changed from −2 to +4, and the element Cu was reduced from +1 and +2 to 0. The Cu₂S is the oxidizing agent and also the reducing agent. The smelting temperature range of the above reaction is above 1100–1350 °C. A similar reaction occurs between Bi₂S₃ (bismuthinite) and Bi₂O₃, as shown below:^[15]2 $$$\begin{equation}2B{{i}_2}{{S}_3}\left( s \right) + B{{i}_2}{{O}_3}\left( s \right)\mathop \to \limits^\Delta Bi\left( l \right) + S{{O}_2}\left( g \right)\end{equation}$$$

The two regular solid states reactions inspired the new strategy to enhance the thermoelectric properties of metal sulfides: 1) the addition of the oxides could modify the concentration of metal ions in the matrix for optimizing the carrier concentration; 2) the SO₂ gas could introducing the pores for reducing the lattice thermal conductivity.

Here, we designed a modified solid states reaction for improving the thermoelectric and mechanical properties of Cu_1.8S-based materials with not only the two advantages mentioned above but also element doping and second phases.

The magnetic Fe₃O₄ particles were added to the Cu_1.8S matrix according to a stoichiometric ratio of x wt.% Fe₃O₄ (x = 0, 0.5, 1, 1.5, 2, 2.5) and sintered by spark plasma sintering to form a magnetic nanocomposite, as shown in Figure  1. The addition of Fe₃O₄ causes a series of in situ solid-state reactions during the sintering process to control the composition and microstructure. The reaction parameters (such as temperature and composition) are very important for the redox reaction and directly determine the thermoelectric properties. In the field of metallurgy, the smelting temperature range of the above reaction is above 1100–1350 °C. However, in this work, the sintering temperature is fixed at 500 °C because 500 °C is an optimized temperature for sintering Cu_1.8S powder to bulk. The controlled reaction is not conducted completely due to the low reaction temperature, which is why there are still Fe₃O₄ particles in the sintered bulk. However, the incomplete reaction is good for formatting the complex microstructure of the sintered samples and benefits from the high thermoelectric properties. The additional contents of Fe₃O₄ in Cu_1.8S have been well investigated and discussed in the manuscript. With the increase of Fe₃O₄, the content of pores is increased, and the Fe doping content in the matrix is also increased. However, there is a limitation of Fe in the Cu_1.8S lattice, it is therefore the reaction is finished when the Fe₃O₄ contents are beyond 1.5 wt.% (Figure  2), and the excess Fe₃O₄ remains in the matrix as a dispersed second phase. The second phase (Cu_1.96S and Fe₃O₄) and porous structure effectively decreased the thermal properties and ferromagnetic particles Fe₃O₄ increased the Seebeck coefficient. The related reaction equation can be roughly given as follows: 3 $$$\begin{equation}C{{u}_{1.8}}S\left( s \right) + xF{{e}_3}{{O}_4}\left( s \right)\mathop \to \limits^\Delta C{{u}_{1.8}}F{{e}_x}{{S}_{1 - 2x}}\left( s \right) + 2xS{{O}_2}\left( g \right)\end{equation}$$$

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

According to the smelting reaction of ore, the addition of oxides will generate SO₂. Therefore, the addition of Fe₃O₄ to copper sulfide can not only release SO₂ to form the porous structure but also realize Fe doping. The carrier concentration was modified by Fe doping and the generating of S vacancies, resulting in the 26% enhancement of the average power faction (Figure  1b). The introduction of Fe₃O₄ particles affects the electrical transport properties of non-magnetic Cu_1.8S to a certain extent. The reason is that the magnetic particles scatter some carriers and affect the carrier mobility and electron conductivity. Due to the generating of SO₂, the pores remained in the Cu_1.8S. The kinetics of redox reactions do not change during the SPS process because the content of reactant Fe₃O₄ is very low compared to the Cu_1.8S. The remaining Fe₃O₄ in the Cu_1.8S matrix is due to the lower sintering temperature and the limitation of Fe in the matrix, which is beneficial to improve the thermoelectric properties. The second phases and pores benefit from reducing the lattice thermal conductivity (70% decrement, as shown in Figure  1c). As a result, the ZT value of the Cu_1.8S + 2 wt.% Fe₃O₄ sample gets a 180% increment compared to that of the pristine sample, reaching 1.3 at 773 K. It is worth mentioning that the average ZT from (323 to 773 K) is 0.57 of the magnetic nanocomposites, this is an ultra-high value in copper sulfides based thermoelectric materials (as shown in Figure  1e). This work demonstrated that the addition of oxide particles is beneficial to modifying the composition and introducing pores and second phases in metal sulfides. Our strategy could be also utilized to enhance TE properties for other sulfide materials, such as Bi₂S₃, Ag₂S, PbS, Cu₃SbS₄, and Cu₂ZnSnS₄.

Results and Discussion

Phase Structure

The phase structures of the Cu_1.8S + x wt.% Fe₃O₄ (x = 0, 0.5, 1, 1.5, 2, 2.5) bulk samples were characterized by X-ray diffraction (XRD) (Figure  2). The peak of the pure sample corresponded to the hexagonal Cu_1.8S phase (PDF#23-0962), i.e., the room temperature (RT) phase. With the addition of Fe₃O₄, the sample diffraction peak began to detect some Cu_1.96S (PDF#29-0578), which we attributed to the volatilization of sulfur.^[17] The previous study demonstrated that S volatilization can generate an S vacancy, and the presence of an S vacancy induces the collapse of the lattice frame.^[18] The rejection of cations in anion vacancies to the greatest extent possible is a distinctive phenomenon in sulfides; an anion vacancy induces lattice expansion.^[19] The evolution of the diffraction peak matches this phenomenon well. With the increase of Fe₃O₄, the diffraction peaks gradually shift toward a low angle. When the addition contents to 1.5 wt.%, the solid solution limit of samples is reached and the diffraction peaks no longer shift. Besides, when the additional contents are above 1.5 wt.%, there is an obvious diffraction peak of Cu_1.96S appearing in the XRD patterns. Subsequently, the 2 wt.% doped sample was analyzed using the Rietveld refinement method, and the lattice constant of all samples is shown in Figure S8 (Supporting Information).

Thermoelectric Performance

The temperature dependence in the TE performance of the Cu_1.8S + x wt.% Fe₃O₄ (x = 0, 0.5, 1, 1.5, 2, 2.5) bulk samples are shown in Figure  3. The electrical conductivity (σ) initially increased with temperature and then decreased, demonstrating that the semiconductor behavior changed to degenerate semiconductor behavior. The pure sample exhibited ultrahigh electrical conductivity, whereas the Fe₃O₄-doped sample exhibited decreased electrical conductivity at all temperatures after the addition of Fe₃O₄. In this study, Fe₃O₄ was added to the matrix, because it acts as a p-type dopant and the carrier scattering center, which reduces the carrier concentration and mobility (Figure S7, Supporting Information), so the electrical conductivity decreased slightly. The inflection point of the electrical conductivity represents the phase transition (PT) temperature point in the copper sulfides, and the PT temperature of the Cu-rich phase (namely, Cu_1.96S and Cu₂S) shifts toward higher temperatures. This phenomenon corresponds to the XRD results. With the addition of Fe₃O₄, Cu_1.96S was formed. When the Fe₃O₄ content reached a certain extent, the electrical conductivity no longer deteriorated. The Seebeck coefficient (S) is shown in Figure  3b. Positive Seebeck coefficients for all samples demonstrated p-type conductive behavior, with the main carrier being the hole. The Seebeck coefficient increased with the addition of Fe₃O₄, which can be attributed to the role of the interfaces scattering some low-energy carriers and the increased density of states. The discussion of interface barrier scattering carriers is detailed in the magnetic properties part, and the increased density of states is detailed in the Supporting Information. To further determine the effects of the nanoprecipitate phase on the carrier transport properties of copper sulfides, Hall tests were performed at 323 K. Figure S7 (Supporting Information) shows the carrier concentration (n) and mobility (μ) of the Cu_1.8S + x wt.% Fe₃O₄ (x = 0, 0.5, 1, 1.5, 2, 2.5) bulk samples at RT. The ultrahigh electrical conductivity of pure Cu_1.8S can be attributed to a higher hole concentration. The value of n decreased from 5.56 × 1020 cm⁻³ for the pure sample to 2.49 × 1020 cm⁻³ for the Cu_1.8S + 2 wt.% Fe₃O₄ sample. With the addition of Fe₃O₄, the value of μ gradually decreased from 33.4 to 18 cm² V⁻¹ s⁻¹ for the Cu_1.8S + 2 wt.% Fe₃O₄ sample. Owing to the Seebeck coefficient being optimized effectively over the entire temperature range, the average power factor was 26% higher than that of the pure Cu_1.8S sample.

[IMAGE OMITTED. SEE PDF]

The sound velocities of all samples were measured by the ultrasonic reflection method. The results are presented in Figure  4a. The longitudinal sound speed (vl) initially decreased and then increased after adding Fe₃O₄, with the 2 wt.% doped samples exhibiting the lowest vl. The phonon transmission speed indicates the introduction of interfaces and phase boundaries, which effectively reduce the sound velocity. In addition, the transverse sound speed (vt) and average sound speed (va) were maintained at comparable levels. The related elastic properties and calculation formulas of the parameters are presented in Table S1 (Supporting Information). The previous study proved that porous interfaces are more stable than nanoprecipitates at high temperatures and strengthen phonon scattering.^[11d] The presence of porous structure effectively enhanced phonon scattering to optimize thermal conductivity in Cu_1.8S + 2 wt.% Fe₃O₄. After doping 2 wt.% Fe₃O₄ in Cu_1.8S, iron oxide nanoprecipitates were observed in the matrix. Different from the smelting of copper concentrate (>1100 °C), this is because the reaction is not complete due to insufficient temperature. Therefore, the continuous addition of Fe₃O₄ will lead to more iron oxide left in the matrix. It is well known that the thermal conductivity of oxides is high.^[20] Based on the Debye–Callaway model, $$\tau _B^{ - 1} = \frac{v}{d}$$, where τB, v, and d are the phonon relaxation time, sound velocity, and average grain size, respectively. The larger grain size corresponds to a longer phonon relaxation time, which weakens phonon scattering and increases thermal conductivity.^[21] Unfortunately, there is no accurate way to detect carrier conductivity in Cu_1.8S material systems because of the existence of ion conductivity and because we failed to obtain the lattice thermal conductivity. Because of the influence of various defects on phonon transport, the total thermal conductivity of Cu_1.8S + 2 wt.% Fe₃O₄ was lower than that of all samples over the entire temperature range, and the thermal conductivity was reduced to 0.58 W m⁻¹ K⁻¹ at 773 K. Thereafter, the dimensionless figure of merit ZT was optimized over the entire temperature range after adding Fe₃O₄. Ultimately, the optimal ZT value of the Cu_1.8S + 2 wt.% Fe₃O₄ sample was 1.3, which was 180% higher than that of the pure sample at 773 K. The theoretical efficiency of the Cu_1.8S + 2 wt.% Fe₃O₄ sample was 9.6% (Figure S7, Supporting Information). More critically, the average ZT of 0.57 for the Cu_1.8S + 2 wt.% Fe₃O₄ sample over the entire temperature range was a relatively high value for Cu_1.8S material systems.

[IMAGE OMITTED. SEE PDF]

To investigate the valence states of the Cu_1.8S + 2 wt.% Fe₃O₄ sample, XPS was performed, and the results are shown in Figure  5. The full survey scanning results (Figure  5a) demonstrate the existence of Cu 2p, Fe 2p, Cu Lmm, O 1s, and S 2p. Figure  5b shows the Cu 2p spectrum containing two binding energy peaks, Cu 2p_1/2 and 2p_3/2. Because the positions of the binding energies of Cu are similar, visually distinguishing the valence of Cu is usually difficult. The characteristic of monovalent Cu is that it exhibits a small satellite peak. Metal Cu showed no satellite peak, whereas divalent Cu showed a strong satellite peak. To thoroughly determine the valence state of Cu, the Auger spectrum peak was examined, and the results are shown in Figure  5c. The results indicate that Cu was monovalent in the Cu_1.8S + 2 wt.% Fe₃O₄ sample. Figure  5d shows the binding energy peaks of S 2p_1/2 and S 2p_3/2 at 162.8 and 161.6 eV, respectively. In addition, Fe 2p_1/2, satellite, and 2p_3/2 peaks were detected, demonstrating the coexistence of divalent and trivalent iron (Figure  5e). Furthermore, an O 1s peak was detected (Figure  5f). The results show the coexistence of Fe²⁺ and Fe³⁺. These results confirmed the existence of Fe₃O₄ and Cu-rich phases (Cu_1.96S).

[IMAGE OMITTED. SEE PDF]

TEM Analysis

To verify the existing defects at the nanoscale, the Cu_1.8S + 2 wt.% Fe₃O₄ sample was analyzed by transmission electron microscopy (TEM). Enlarged high-angle annular dark-field (HAADF) images are shown in Figure  6a. Different contrast regions were observed in the images, with the inset corresponding to fast Fourier transform (FFT) images. The calibration results of the diffraction spots indicate that the second phase was Fe₃O₄, which is consistent with the EDS mapping results. Figure  6b shows the high-resolution TEM (HRTEM) images of the interface region and the corresponding FFT images. The results demonstrate that Fe₃O₄ was embedded in the matrix (the triangular precipitated phase in region 2). The inverse fast Fourier transform (IFFT) of regions 1 (Cu_1.8S) and 2 (Fe₃O₄) are shown in Figure  6c,d, respectively. As can be observed, no high-density dislocations exist. The lattice spacing was measured from IFFT images. Regions 1 and 2 had lattice spacings of 2.099 and 1.813 Å, respectively. Figure  6e shows the HAADF-STEM images of the Cu_1.8S + 2 wt.% Fe₃O₄ sample. The different contrast regions were due to composition variations. The elemental distribution of iron and oxygen was relatively obvious, and copper and sulfur were uniformly distributed in the matrix. In addition, the Cu-rich phase (Cu_1.96S) was also observed in the matrix, as shown in Figure S1 (Supporting Information).

[IMAGE OMITTED. SEE PDF]

Figure  7a shows that the Cu_1.8S + 2 wt.% Fe₃O₄ sample exhibited a twinning-like boundary in the matrix, which was confirmed by the FFT patterns presented in Figure  7d. In this study, the cause of the twinning-like boundary is mainly due to the lattice strain caused by the second phase. For example, the influence of stress or defects in the process of crystal growth leads to local dislocation of the arrangement of atoms, thus forming a twinning-like boundary. The stress analysis of the entire twinning-like boundary region (Figure  7e–g) also proves that the twinning-like boundary has two different stress distributions. The checkerboard-like diffraction spots correspond to the Na structure in the Cu─S system. The marked spots (yellow circles) are coherent with the hexagonal phase. Through IFFT, we observed that dislocations existed, as depicted in Figure  7b,c. The geometric phase analysis (GPA) results are shown in Figure  7e–g, and the stress distribution around the twin boundary can also be observed. Local concentrated stress existed around the dislocations, which effectively scattered the mid-frequency phonons and decreased the lattice thermal conductivity over the entire temperature range.^[22] The structural evolution caused lattice stress due to the formation of additional nanoprecipitates and grain refinement, which introduced numerous interfaces. This resulted in grain extrusion and dislocations distributed at the interfaces. In addition, the existence of dislocations effectively enhances the mechanical properties of reinforced composites. The results are detailed below.

[IMAGE OMITTED. SEE PDF]

Magnetic Properties

The magnetic properties and interface diagrams of the pure and Cu_1.8S + 2 wt.% Fe₃O₄ samples are shown in Figure  8. Br and Hc represent the remnant magnetization and coercive force, respectively. The M─H curves of the Cu_1.8S + 2 wt.% Fe₃O₄ sample demonstrated that the M value increased with H and reached saturation when H reached a certain level. As a typical soft magnetic material, the addition of Fe₃O₄ triggered the magnetic transformation of Cu_1.8S. Cu_1.8S exhibited weak magnetism owing to the presence of Fe₃O₄. The previous study showed that the scattering parameter of the matrix (ϒ0) is beneficial for the Seebeck coefficient (S) but does not affect the effective mass. The relationship between ϒ0 and S was calculated using the following equation: $$S = [ {\frac{{8{{\pi }^2}k_B^2}}{{3e{{h}^2}}}} ]{{m}^*}T{{[ {\frac{\pi }{{3n}}} ]}^{\frac{2}{3}}}( {\Upsilon + 1} )$$.^[23] The work function of magnetic Fe₃O₄ (φ = 5.5 eV)^[24] was larger than that of the Cu_1.8S matrix (φ = 5.27 eV),^[25] which may have created barriers for hole carriers to reduce the pure sample's high carrier concentration (≈5.56 × 1020 cm⁻³) to the ideal interval (≈2.5 × 1020 cm⁻³).^[26] The difference in work function between Cu_1.8S and Fe₃O₄, thus a depleted layer would be generated near the interface when the carrier transports. This potential barrier would scatter low-energy carriers while allowing high-energy transport, resulting in the slightly improved Seebeck coefficient of the samples. Through X-ray photoelectron spectroscopy (XPS) characterization, we observed a chemical shift from a high binding energy to a low binding energy, and the results demonstrate that charge transfer occurred.^[23a,27] As depicted in the diagram, the charge transfer forms an interface potential barrier. In this study, the Seebeck coefficient after adding Fe₃O₄ at RT was consistent with the above trend, indicating that iron oxide is beneficial to the Seebeck coefficient of copper sulfides. These results demonstrate that adding a promoted number of magnetic Fe₃O₄ particles to the matrix not only introduces extra electrons and decreases the carrier concentration but also enhances the scattering of electrons. Therefore, as the Fe₃O₄ content increased, the carrier concentration and mobility gradually decreased, as shown in Figure S7 (Supporting Information).

[IMAGE OMITTED. SEE PDF]

Mechanical Properties

The mechanical performances of the pure and Cu_1.8S + 2 wt.% Fe₃O₄ samples were characterized via nanoindentation using the Nano-Blitz 3D method (Figure  9). The average hardness of the pure sample was 0.864 GPa, which was lower than the hardness of the Cu_1.8S + 2 wt.% Fe₃O₄ sample. After adding Fe₃O₄, the modulus of the sample also improved. The overall mechanical properties of the Cu_1.8S+2 wt.% Fe₃O₄ sample were increased by 35%. The mechanical performance was optimized by second-phase strengthening. A large number of interfaces and phases were introduced via mechanical alloying, which effectively hindered dislocation movement and thus enhanced the mechanical properties. The improved mechanical performance contributes to the work stability of p-type legs in copper sulfide TE devices.

[IMAGE OMITTED. SEE PDF]

Conclusion

In this study, we propose a strategy for constructing copper sulfide magnetic nanocomposites by drawing lessons from the smelting process of Cu minerals. In situ, solid-state chemical reactions achieve the modulation of carrier concentration, followed by the composition and microstructure control. The ferromagnetic Fe₃O₄ facilitated charge transfer, and engineered the interface barrier potential, effectively enhancing the Seebeck coefficient. Benefiting from the enhanced Seebeck coefficient, the average power factor was increased by 26% over the entire operating temperature range. In addition, phase boundaries, dislocations, and precipitates effectively suppress the thermal conductivity of porous magnetic nanocomposites over the entire temperature range. The thermal conductivity of the doped sample was 70% lower than that of the pure sample. The mechanical properties were effectively enhanced by 35%, as characterized by nanoindentation. Based on the optimization of the power factor and thermal conductivity over the entire temperature range, the ZT value of the Cu_1.8S + 2 wt.% Fe₃O₄ sample was 1.3 at 773 K, which was 180% higher than that of the pure sample. The average ZT was 0.57 in the temperature range of 323–773 K, which is the highest average ZT value of the Cu─S material system. This work demonstrated that the use of oxides is a high-efficiency strategy to regulate the carrier concentration of metal sulfides, and this strategy can also be used in other fields of sulfide functional materials.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 52162029), the Outstanding Youth Fund of Yunnan Province (Grant No. 202201AV070005), National Key R&D Program of China (Grant No. 2022YFF0503804), Yunnan Science and Technology Program (202401AT070403).

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

a) S. Ayachi, X. He, H. J. Yoon, Adv. Energy Mater. 2023, 13, [eLocator: 2300937];

b) J. F. Cai, R. Y. Wang, S. Zhuang, F. Gao, M. H. Zhang, Z. W. Zhang, X. J. Tan, G. Q. Liu, J. Jiang, Adv. Funct. Mater. 2024, 34, [eLocator: 202311217];

c) B. Qin, D. Wang, T. Hong, Y. Wang, D. Liu, Z. Wang, X. Gao, Z. H. Ge, L. D. Zhao, Nat. Commun. 2023, 14, 1366;

d) K. P. Zhao, A. B. Blichfeld, H. Y. Chen, Q. F. Song, T. S. Zhang, C. X. Zhu, D. D. Ren, R. Hanus, P. F. Qiu, B. B. Iversen, F. F. Xu, G. J. Snyder, X. Shi, L. D. Chen, Chem. Mater. 2017, 29, 6367;

e) Y.‐K. Zhu, Y. Sun, X. Dong, L. Yin, M. Liu, M. Guo, H. Wu, F. Li, Z. Guo, X. Wang, K. Yu, F. Guo, Q. Zhang, Z. Liu, W. Cai, J. Sui, Joule 2024, 8, 2412;

f) Y. Liu, Q. H. Zhang, A. B. Huang, K. Y. Zhang, S. Wan, H. Y. Chen, W. S. Zuo, Y. Z. Wang, X. Cao, L. J. Wang, U. Lemmer, W. Jiang, Nat. Commun. 2024, 15, 2141.

a) D. M. Rove, in Thermoelectrics Handbook, (Ed: D. W. Rowe), Boca Raton, Florida 2006, 1, pp. 1–10;

b) M. Jiang, Y. T. Fu, Q. H. Zhang, Z. L. Hu, A. B. Huang, S. L. Wang, L. J. Wang, W. Jiang, Natl. Sci. Rev. 2023, 10, [eLocator: nwad095].

a) H. Liu, X. Shi, F. Xu, L. Zhang, W. Zhang, L. Chen, Q. Li, C. Uher, T. Day, G. J. Snyder, Nat. Mater. 2012, 11, 422;

b) M. Beekman, D. T. Morelli, G. S. Nolas, Nat. Mater. 2015, 14, 1182;

c) Y. Sun, L. Xi, J. Yang, L. Wu, X. Shi, L. Chen, J. Snyder, J. Yang, W. Zhang, J. Mater. Chem. A 2017, 5, 5098.

a) W. Zhou, H. Z. Li, Z. H. Shan, R. Zhang, S. K. Lu, J. Pei, Z. H. Ge, M. Zhou, Y. B. Wang, B. P. Zhang, Sci. China Mater. 2023, 66, 2051;

b) Y. J. Wang, Z. Long, Y. Cheng, M. Zhou, H. Y. Chen, K. P. Zhao, X. Shi, Mater. Today Phys. 2023, 32, [eLocator: 101028];

c) W. D. Liu, X. L. Shi, H. Gao, R. Moshwan, S. D. Xu, Y. Wang, L. Yang, Z. G. Chen, J. Zou, J. Mater. Chem. C 2019, 7, 5366;

d) K. P. Zhao, A. B. Blichfeld, E. Eikeland, P. F. Qiu, D. D. Ren, B. B. Iversen, X. Shi, L. D. Chen, J. Mater. Chem. A 2017, 5, [eLocator: 18148].

a) S. Zhao, H. Chen, X. Zhao, J. Luo, Z. Tang, G. Zeng, K. Yang, Z. Wei, W. Wen, X. Chen, Y. Sun, Mater. Today Phys. 2020, 15, [eLocator: 100271];

b) H. C. Tang, H. L. Zhuang, B. W. Cai, Asfandiyar, J. F. D, F. H. Sun, J. F. Li, J. Mater. Chem. C 2019, 7, 4026;

c) P. F. Qiu, T. Mao, Z. F. Huang, X. G. Xia, J. C. Liao, M. T. Agne, M. Gu, Q. H. Zhang, D. D. Ren, S. Q. Bai, X. Shi, G. J. Snyder, L. D. Chen, Joule 2019, 3, 1538;

d) Y. T. Fu, Q. H. Zhang, Z. L. Hu, M. Jiang, A. B. Huang, X. Ai, S. Wan, H. Reith, L. J. Wang, K. Nielsch, W. Jiang, Energy Environ. Sci. 2022, 15, 3265.

a) B. B. Jiang, W. Wang, S. X. Liu, Y. Wang, C. F. Wang, Y. N. Chen, L. Xie, M. Y. Huang, J. Q. He, Science 2022, 377, 208;

b) X. Liang, D. Jin, F. H. Dai, Adv. Electron. Mater. 2019, 5, [eLocator: 1900486];

c) W. Zhang, Y. Lou, H. Dong, F. Wu, J. Tiwari, Z. Shi, T. Feng, S. T. Pantelides, B. Xu, Chem. Sci. 2022, 13, [eLocator: 10461];

d) C. R. McCormick, R. E. Schaak, J. Am. Chem. Soc. 2021, 143, 1017.

a) H. C. Tang, F. H. Sun, J. F. Dong, Asfandiyar, H. L. Z, Y. Pan, J. F. Li, Nano Energy 2018, 49, 267;

b) R. S. Sundari, V. Vijay, V. Shalini, S. Harish, M. Shimomura, J. Archana, S. Ponnusamy, M. Navaneethan, J. Alloy. Compd. 2023, 960, [eLocator: 170796];

c) M. Yang, X. Liu, B. Zhang, Y. Chen, H. Wang, J. Yu, G. Wang, J. Xu, X. Zhou, G. Han, ACS Appl. Mater. Interfaces 2021, 13, [eLocator: 39541];

d) R. Zhang, J. Pei, Z.‐J. Han, Y. Wu, Z. Zhao, B.‐P. Zhang, J. Adv. Ceram. 2020, 9, 535.

a) T. Mao, P. Qiu, P. Hu, X. Du, K. Zhao, T. R. Wei, J. Xiao, X. Shi, L. Chen, Adv. Sci. (Weinh) 2020, 7, [eLocator: 1901598];

b) H. H. Hu, B. Su, X. D. Liu, H. C. Thong, Y. L. Jiang, H. Z. Li, J. W. Li, H. L. Zhuang, Z. R. Han, J. C. Yu, B. L. Mehdi, J. F. Li, Joule 2024, 8, 416;

c) P. Qiu, M. T. Agne, Y. Liu, Y. Zhu, H. Chen, T. Mao, J. Yang, W. Zhang, S. M. Haile, W. G. Zeier, J. Janek, C. Uher, X. Shi, L. Chen, G. J. Snyder, Nat. Commun. 2018, 9, 2910;

d) T. Mao, P. Qiu, J. Liu, X. Du, P. Hu, K. Zhao, D. Ren, X. Shi, L. Chen, Phys. Chem. Chem. Phys. 2020, 22, 7374;

e) D. Yang, X. Su, J. Li, H. Bai, S. Wang, Z. Li, H. Tang, K. Tang, T. Luo, Y. Yan, J. Wu, J. Yang, Q. Zhang, C. Uher, M. G. Kanatzidis, X. Tang, Adv. Mater. 2020, 32, [eLocator: 2003730].

a) Y. Pei, X. Shi, A. LaLonde, H. Wang, L. Chen, G. J. Snyder, Nature 2011, 473, 66;

b) X. Shi, S. Song, G. Gao, Z. Ren, Science 2024, 384, 757;

c) B. Qin, D. Wang, X. Liu, Y. Qin, J. F. Dong, J. Luo, J. W. Li, W. Liu, G. Tan, X. Tang, J. F. Li, J. He, L. D. Zhao, Science 2021, 373, 556;

d) R. Nunna, P. F. Qiu, M. J. Yin, H. Y. Chen, R. Hanus, Q. F. Song, T. S. Zhang, M. Y. Chou, M. T. Agne, J. Q. He, G. J. Snyder, X. Shi, L. D. Chen, Energ. Environ. Sci. 2017, 10, 1928.

a) L. Hu, Y. B. Luo, Y. W. Fang, F. Y. Qin, X. Cao, H. Y. Xie, J. W. Liu, J. F. Dong, A. Sanson, M. Giarola, X. Y. Tan, Y. Zheng, A. Suwardi, Y. Z. Huang, K. Hippalgaonkar, J. Q. He, W. Q. Zhang, J. W. Xu, Q. Y. Yan, M. G. Kanatzidis, Adv. Energy Mater. 2021, 11, [eLocator: 2100661];

b) Z. Zhou, Y. Huang, B. Wei, Y. Yang, D. Yu, Y. Zheng, D. He, W. Zhang, M. Zou, J. L. Lan, J. He, C. W. Nan, Y. H. Lin, Nat. Commun. 2023, 14, 2410;

c) J. Q. Zheng, Y. Wen, S. N. Wang, Y. C. Li, S. Q. Wang, Z. Zhao, S. Liu, S. B. Liu, X. Gao, L. D. Zhao, Adv. Funct. Mater. 2024.

a) Y. Qin, B. Qin, T. Hong, X. Zhang, D. Wang, D. Liu, Z. Y. Wang, L. Su, S. Wang, X. Gao, Z. H. Ge, L. D. Zhao, Science 2024, 383, 1204;

b) D. Liu, D. Wang, T. Hong, Z. Wang, Y. Wang, Y. Qin, L. Su, T. Yang, X. Gao, Z. Ge, B. Qin, L. D. Zhao, Science 2023, 380, 841;

c) H. Su, P. Lu, C. Zhu, W. Qiu, X. Qiu, K. Zhao, X. Lu, Y. Wei, X. Shi, L. Chen, F. Xu, Acta Mater. 2023, 248, [eLocator: 118764];

d) H. Hu, H. L. Zhuang, Y. Jiang, J. Shi, J. W. Li, B. Cai, Z. Han, J. Pei, B. Su, Z. H. Ge, B. P. Zhang, J. F. Li, Adv. Mater. 2021, 33, [eLocator: 2103633].

a) Y. He, P. Lu, X. Shi, F. Xu, T. Zhang, G. J. Snyder, C. Uher, L. Chen, Adv. Mater. 2015, 27, 3639;

b) M. Li, D. L. Cortie, J. X. Liu, D. H. Yu, S. M. K. N. Islam, L. L. Zhao, D. R. G. Mitchell, R. A. Mole, M. B. Cortie, S. X. Dou, X. L. Wang, Nano Energy 2018, 53, 993;

c) Y. Jiang, J. Dong, H. L. Zhuang, J. Yu, B. Su, H. Li, J. Pei, F. H. Sun, M. Zhou, H. Hu, J. W. Li, Z. Han, B. P. Zhang, T. Mori, J. F. Li, Nat. Commun. 2022, 13, 6087.

Z. Zulhan, I. M. Fauzian, T. Hidayat, J. Mater. Res. Technol. 2020, 9, [eLocator: 13625].

a) I. Wilkomirsky, R. Parra, F. Parada, E. Balladares, E. Seguel, J. Etcheverry, R. Díaz, Metall. Mater. Trans. B 2020, 51, 1540.

b) S. Zhou, X. Guo, B. Tian, B. Li, Y. Wei, Metals 2020, 11, 19.

W. Jin, S. Yang, C. Tang, Y. Li, C. Chang, Y. Chen, ACS Sustainable Chem. Eng. 2023, 11, 9932.

a) Y. Yao, B.‐P. Zhang, J. Pei, Q. Sun, G. Nie, W.‐Z. Zhang, Z.‐T. Zhuo, W. Zhou, ACS Appl. Mater. Interfaces 2018, 10, [eLocator: 32201];

b) Y. X. Zhang, Q. Lou, Z. H. Ge, S. W. Gu, J. X. Yang, J. Guo, Y. K. Zhu, Y. Zhou, X. H. Yu, J. Feng, J. Q. He, Acta Mater. 2022, 233, [eLocator: 117972];

c) Z. H. Ge, X. Y. Liu, D. Feng, J. Y. Lin, J. Q. He, Adv. Energy Mater. 2016, 6, [eLocator: 1600607];

d) D. D. Liang, B. P. Zhang, L. Zou, J. Alloy. Compd. 2018, 731, 577;

e) Y.‐X. Zhang, J. Feng, Z.‐H. Ge, Chem. Eng. J. 2022, 428;

f) Q. Jiang, H. Yan, J. Khaliq, Y. Shen, K. Simpson, M. J. Reece, J. Mater. Chem. A 2014, 2, 9486;

g) Y. Wang, Z. Long, Y. Cheng, M. Zhou, H. Chen, K. Zhao, X. Shi, Mater. Today Phys. 2023, 32;

h) Y. X. Zhang, Q. Y. Huang, X. Yan, C. Y. Wang, T. Y. Yang, Z. Y. Wang, Y. C. Shi, Q. Shan, J. Feng, Z. H. Ge, Nat. Commun. 2024, 15, 2736;

i) Y. Lou, X. Li, Z. Shi, H. Zhou, T. Feng, B. Xu, Nano Lett. 2023, 23, 5317;

j) X. Li, Y. Lou, K. Jin, L. Fu, P. Xu, Z. Shi, T. Feng, B. Xu, Angew. Chem. Int. Ed. Engl. 2022, 61, [eLocator: e202212885].

a) S. W. Gu, P. Qin, Y. X. Zhang, J. Guo, Q. Shan, J. Feng, Z. H. Ge, J. Alloy. Compd. 2021, 852, [eLocator: 156972];

b) Z. H. Ge, Y. X. Zhang, D. S. Song, X. Y. Chong, P. Qin, F. S. Zheng, J. Feng, L. D. Zhao, J. Mater. Chem. A 2018, 6, [eLocator: 14440].

a) Y. X. Zhang, T. Y. Yang, Z. Y. Wang, J. Feng, Z. H. Ge, Mater. Today Phys. 2022, 27, [eLocator: 100808];

b) L. Zou, B.‐P. Zhang, Z.‐H. Ge, C. Gao, D.‐B. Zhang, Y.‐C. Liu, Phys. Status Solidi A 2013, 210, 2550.

Y. X. Zhang, Z. H. Ge, J. Feng, J. Alloy. Compd. 2017, 727, 1076.

S. Sailler, R. Bueno Villoro, S. Bayesteh, H. Schlörb, M. O. Cichocka, K. Nielsch, S. Zhang, N. Pérez, Mater. Today Phys. 2024, 46, [eLocator: 101477].

Y. Gong, W. Dou, B. Lu, X. Zhang, H. Zhu, P. Ying, Q. Zhang, Y. Liu, Y. Li, X. Huang, M. F. Iqbal, S. Zhang, D. Li, Y. Zhang, H. Wu, G. Tang, Nat. Commun. 2024, 15, 4231.

a) S. I. Kim, K. H. Lee, H. A. Mun, H. S. Kim, S. W. Hwang, J. W. Roh, D. J. Yang, W. H. Shin, X. S. Li, Y. H. Lee, G. J. Snyder, S. W. Kim, Science 2015, 348, 109;

b) P. Carruthers, Phys. Rev. 1959, 114, 995.

a) W. Zhao, Z. Liu, Z. Sun, Q. Zhang, P. Wei, X. Mu, H. Zhou, C. Li, S. Ma, D. He, P. Ji, W. Zhu, X. Nie, X. Su, X. Tang, B. Shen, X. Dong, J. Yang, Y. Liu, J. Shi, Nature 2017, 549, 247;

b) F. Luo, C. Zhu, J. Wang, X. He, Z. Yang, S. Ke, Y. Zhang, H. Liu, Z. Sun, ACS Appl. Mater. Interfaces 2022, 14, [eLocator: 45503].

M. Zervos, Z. Viskadourakis, G. Athanasopoulos, R. Flores, O. Conde, J. Giapintzakis, J. Appl. Phys. 2014, 115, [eLocator: 033709].

a) G. Liu, T. Schulmeyer, J. Brötz, A. Klein, W. Jaegermann, Thin Solid Films 2003, 431, 477;

b) Y. Wang, S. Shen, M. Liu, G. He, X. Li, J. Colloid Interf. Sci. 2024, 655, 187.

M. Hong, J. Zou, Z. G. Chen, Adv. Mater. 2019, 31, [eLocator: 1807071].

a) W. Zhao, Z. Liu, P. Wei, Q. Zhang, W. Zhu, X. Su, X. Tang, J. Yang, Y. Liu, J. Shi, Y. Chao, S. Lin, Y. Pei, Nat. Nanotechnol. 2017, 12, 55;

b) C. C. Li, S. F. Ma, P. Wei, W. T. Zhu, X. L. Nie, X. H. Sang, Z. G. Sun, Q. J. Zhang, W. Y. Zhao, Energ. Environ. Sci. 2020, 13, 535.