Skip to main content
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Feb 11;21(11):2411798. doi: 10.1002/smll.202411798

Transient Liquid Induced Hierarchical Structure Contributes to High Thermoelectric Performance in Ag2Se

Peeranut Kitthonbancha 1,2, Wei‐Di Liu 2, Meng Li 2, Supree Pinitsoontorn 1,3,, Zhi‐Gang Chen 2,
PMCID: PMC11922017  PMID: 39935139

Abstract

Owing to the intrinsic high thermoelectric performance, Ag2Se is a promising alternative for traditional near‐room temperature Bi2Te3‐based materials. In this study, a Na2SO3 solution has been used as the transient liquid during the modified cold sintering process to induce a hierarchical structure, including micropores, nanopores, sub‐nanopores, and additional nanoscale Na2SO3 residuals. Such a hierarchical structure contributes to an ultralow lattice thermal conductivity of 0.18 W m−1 K−1 at ≈300 K in the Ag2Se‐30%Na2SO3. Additionally, extra Se vacancies further optimize the carrier concentration to ≈5.6 × 1018 cm−3, leading to a high power factor of ≈25 µW cm−1 K−2 at ≈300 K in the Ag2Se‐30%NS. Consequently, due to the synergistic effects of high power factor and low lattice thermal conductivity, an ultrahigh room‐temperature figure of merit of 1.04 in the Ag2Se‐30%Na2SO3. The study demonstrates that introducing transient liquid solutions in the modified cold sintering process can effectively achieve specific structural engineering and high thermoelectric performance.

Keywords: cold sintering process, hierarchical structure, silver selenide, thermoelectric, transient liquid


Na2SO3 solution has been used as the transient liquid during the modified cold sintering process to induce a hierarchical structure, including micropores, nanopores, sub‐nanopores, and additional nanoscale Na2SO3 residuals. Such a hierarchical structure contributes to an ultralow lattice thermal conductivity of 0.18 W m−1 K−1 and an ultrahigh figure of merit of 1.04 at ≈300 K in the Ag2Se‐30%Na2SO3.

graphic file with name SMLL-21-2411798-g006.jpg

1. Introduction

Owing to the ability to directly convert heat into electricity, thermoelectric technology emerges as a promising sustainable energy technology.[ 1 ] Such a thermoelectric technology has unique characteristics, including being noise‐free, emission‐free, and maintenance‐free, which make it suitable for a wide range of applications, such as microelectronics,[ 2 , 3 ] medical sensors,[ 4 , 5 , 6 ] and wearable devices.[ 7 , 8 , 9 ] Thermoelectric performance is commonly assessed by the dimensionless figure of merit, ZT = (S2σT)/κ, where S is the Seebeck coefficient, σ is the electrical conductivity, κ is total thermal conductivity, and T is the absolute temperature. The κ is comprised of the electronic thermal conductivity (κ e) and the lattice thermal conductivity (κ l). As can be seen, achieving a high ZT requires both maximizing the power factor (S2σ) and minimizing κ.

Generally, thermoelectric materials can be categorized into three types based on their operating temperature range: high‐temperature (>873 K), mid‐temperature (473–873 K), and near‐room‐temperature (<473 K) ones.[ 10 ] For near‐room‐temperature applications, Bi2Te3‐based materials are highly attractive due to their outstanding thermoelectric performance at this temperature range.[ 11 , 12 ] However, the scarcity of tellurium[ 13 ] and the unsatisfactory mechanical properties of Bi2Te3 have limited their applications.[ 14 ] As an alternative, silver selenide (Ag2Se) is a promising high‐performance near‐room‐temperature thermoelectric material. Ag2Se is an n‐type thermoelectric material with a narrow band gap of 0.04–0.2 eV.[ 15 ] Ag2Se is orthorhombic structured with a P212121 space group at room temperature,[ 16 ] which undergoes a phase transition into a superconducting cubic phase at ≈407 K.[ 17 , 18 ] Pristine Ag2Se exhibits high σ [ 19 ] and low κ,[ 20 , 21 ] having an impressively high ZT of ≈0.7 at room temperature.[ 22 ] In addition, due to its tunable ductility, Ag2Se can be exploited as efficient flexible thermoelectric applications.[ 23 , 24 ]

Currently, various strategies have been investigated to enhance the thermoelectric performance of Ag2Se‐based materials, including compositing,[ 25 , 26 ] introducing nano‐inclusions,[ 19 , 27 ] creating point defects,[ 28 ] densifying grain boundaries,[ 29 , 30 , 31 , 32 , 33 ] and incorporating nanopores.[ 34 , 35 ] For instance, Nan et al.[ 36 ] composited Ag2Se with 1.0 wt.% Bi2S3, which decreased the Hall carrier concentration (n H) and enhanced the S to −178.5 µV K−1 at 300 K, leading to a high ZT of 0.96 at 370 K. Tee et al.[ 37 ] developed ternary Cu0.021Ag1.979Se alloys, where Cu doping increased the n H and induced a high σ of 1216 S cm−1 at room temperature, leading to an enhanced ZT from 0.9 to 1.2. Wang et al.[ 16 ] optimized the n H and the Hall carrier mobility (µ H) in Ag2Se by introducing excess Se, which increased S to −136 µV K−1 at 375 K. The reduced n H also decreased κ e, resulting in a high ZT value of 1.02 at 375 K of Ag2Se1.015. Chen et al.[ 38 ] synthesized Ag2Se with a hierarchical porous structure, achieving a ZT of ≈0.7 at 300 K due to its effective multi‐wavelength phonon scattering and ultralow κ l.

Currently, high‐temperature sintering techniques easily induce grain growth and diminish various phonon scattering centers. To address this challenge, low‐temperature sintering methods that preserve nanostructures are essential. The cold sintering process (CSP), featured with the assistance of liquid phase, external pressure, and relatively low sintering temperature, has been applied for the fabrication of thermoelectric materials.[ 39 , 40 ] Lu et al.[ 41 ] compared PbTe0.94Se0.06 materials prepared by CSP and spark plasma sintering (SPS), and found that the CSP prepared PbTe0.94Se0.06 exhibits significantly less grain growth compared with that prepared by SPS. Zhu et al.[ 42 ] developed Bi2Te3‐based bulk materials using CSP at a temperature of <423 K and observed refrained grain growth and small grain sizes.

Herein, we introduce a sodium sulfite (Na2SO3) solution (up to 40 wt.%) as a transient liquid during the modified cold sintering process (mCSP)[ 43 ] of Ag2Se, as illustrated in Figure 1a. Schematical mechanisms for Ag2Se sintering during mCSPs with and without Na2SO3 solution are shown in Figure 1b. For mCSP‐ed Ag2Se without Na2SO3 solution, deionized water was used as a reference. The water evaporation during mCSP can facilitate particle diffusion, resulting in a high density of 98.5%. In contrast, using Na2SO3 solution as the transient liquid can create a hierarchical structure, including micropores, nanopores, sub‐nanopores, and additional nanoscale Na2SO3 residuals, that can effectively scatter multi‐wavelength phonons and suppress κ l. Simultaneously, additional Se vacancies induce optimized n H and enhanced S 2 σ. As a result, the thermoelectric performance of bulk Ag2Se with 30 wt.% Na2SO3 (Ag2Se‐30%NS) achieves a low κ l of ≈0.18 W m−1 K−1 at room temperature with an outstanding ZT value of 1.04. Figure 1c compares the room‐temperature ZT of our study with other state‐of‐the‐art Ag2Se‐based materials,[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ] which demonstrates the feasibility of the transient liquid‐assisted mCSP in enhancing thermoelectric properties of Ag2Se‐based materials. And a more detailed comparison is shown in Table S1 (Supporting Information).

Figure 1.

Figure 1

a) A schematic diagram illustrating the modified cold sintering process (mCSP) using Na2SO3 solution to prepare bulk Ag2Se. b) Schematic diagrams explaining the mechanism for the densification of bulk Ag2Se during mCSP without and with Na2SO3 solution. c) Comparison of the room‐temperature (T) dimensionless figure of merit (ZT) in this study and other state‐of‐the‐art values.[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ]

2. Results and Discussion

Figure 2a shows the X‐ray diffraction (XRD) patterns of the mCSP‐ed Ag2Se using different amounts of Na2SO3 solution. All observed peaks are well aligned with the standard pattern for orthorhombic Ag2Se (PDF 00‐024‐1041), without obvious secondary phases and peak shifts. This indicates that our mCSP method with Na2SO3 solvent as the transient liquid has no significant influence on the phase. Scanning electron microscope (SEM) images of as mCSP‐ed Ag2Se samples with increasing Na2SO3 content are shown in Figure S1 (Supporting Information). As can be seen, when only water is employed as the transit liquid, the Ag2Se sample is densely sintered with high relative density (Figure 2d). Increasing Na2SO3 content leads to the formation of a hierarchical porous structure (Figure S1, Supporting Information), and dramatically reduced relative density (Figure 2d), with enriched Na, S, and O at the pore surface (Figure 2b). Therefore, the decomposition of Na2SO3 should be the main reason for the formation of the hierarchical porous structure. Simultaneously, some Na2SO3 residuals can be observed in the pores (Figure 2c1; Figure S2, Supporting Information) further suggesting partial decomposition of Na2SO3 during the mCSP.

Figure 2.

Figure 2

a) X‐ray diffraction (XRD) patterns of modified cold sintering processed (mCSP) Ag2Se samples with varying quantities of Na2SO3 solution (0 – 40 wt.%). b) Scanning electron microscope (SEM) images with corresponding energy‐dispersive X‐ray spectroscopy (EDS) maps of the Ag2Se‐30%NS sample containing a porous region. c) Transmission electron microscope (TEM) images of the Ag2Se‐30%NS sample. d) The relative density of bulk Ag2Se samples with different Na2SO3 quantities.

The porous structure is hierarchical, comprised of micropore (sub‐micron size), nanopore (≈100 nm), and sub‐nanopore (<10 nm) with some nanoscale Na2SO3 residuals as shown in the transmission electron microscope (TEM) image of Figure 2c. With increasing Na2SO3 solution content, such a hierarchical porous structure induces reduced densities for the mCSP‐ed Ag2Se (Figure 2d), indicating that the formation of the hierarchical porous structure is dominated by Na2SO3 solution, which is possibly due to partial decomposition of Na2SO3 solution and the release of SO2 gas during the mCSP process as shown below:

Na2SO3+2H2OH2SO3+2NaOH (1)
H2SO3H2O+SO2g (2)

NaOH is fully dissoluble in water as depicted by Equations (1) and (2) forming a basic solution, which can facilitate the merging of particles during the mCSP process. It is then squeezed out without further affecting the thermoelectric performance. Therefore, the mCSP mainly influence the sintering process by facilitating the merging of particles, which does not influence the phase transition processes.

Figure 3 presents the thermoelectric properties of the mCSP‐ed Ag2Se. Figure 3a illustrates the temperature‐dependent σ of the mCSP‐ed Ag2Se. All samples display typical semiconducting behavior, as demonstrated by increasing σ with increasing temperature. The sample without Na2SO3 (Ag2Se‐0%NS) exhibits the highest σ value of 1.95 × 103 S cm−1 at ≈300 K. Introducing Na2SO3 during mCSP results in reduced σ, reaching the lowest σ of 1.10 × 103 S cm−1 at ≈300 K for Ag2Se‐40%NS. To understand the influence of Na2SO3 addition on σ, Hall measurement was conducted to measure the n H and the µ H, as shown in Figure 3b. The n H decreases with increasing the amount of Na2SO3, from 13.5 × 1018 cm−3 for Ag2Se‐0%NS to ≈5.6 × 1018 cm−3 for Ag2Se‐30%NS. Therefore, the reduced n H can be attributed to additional Se vacancies as indicated by the reduced Ag:Se ratios (Table S2, Supporting Information).[ 38 , 49 ] To better understand the influence of Se vacancies on n H, we perform first‐principal density functional theory (DFT) calculations of orthorhombic Ag2Se. Figure 3c,d compares the calculated electronic band structures and density of states (DOS) of Ag32Se16 (pristine Ag2Se) and Ag32Se15 (with additional Se vacancies), with corresponding crystal structures shown as insets. As can be seen, introducing Se vacancy expands the band gap of Ag2Se‐based materials, which should be the main reason for reduced n H. In contrast, the influence of nanoscale Na2SO3 residuals on electrical properties is minor. On the other hand, the µ H tends to increase with increasing the amount of Na2SO3 addition, reaching a maximum of 1324 cm2 V−1 s−1 for Ag2Se‐30%NS. The increase in µ H can be mainly attributed to significantly reduced n H. Further increasing the Na2SO3 content from 30% to 40% leads to reduced µ H (from 1324 to 1129 cm2 V−1 s−1, respectively), which should be attributed to increased n H (from 5.63 × 1018 to 6.08 × 1018 cm−3, respectively).

Figure 3.

Figure 3

Temperature (T)‐dependent (a) electrical conductivity (σ) and (b) Hall carrier concentration (n H) and Hall carrier mobility (µ H) at 300 K of Ag2Se samples prepared by modified cold sintering process (mCSP) with varying quantities of Na2SO3 solution (0 – 40 wt.%). Density functional theory (DFT) calculated electronic band structures and density of states (DOS) with schematic crystal structures of (c) Ag32Se16 and (d) Ag32Se15. e) Temperature‐Seebeck coefficient (S), and (f) power factor (S2σ) as‐prepared Ag2Se samples.

Figure 3e plots the temperature‐dependent S, all of which exhibit negative values, indicating n‐type semiconducting behavior. When Ag2Se powders are mCSP‐ed without Na2SO3 solution, the absolute Seebeck coefficient (|S|) value is as small as ≈110 µV K−1 at ≈300 K. However, with the addition of Na2SO3 solution during mCSP, the |S| values significantly increase to ≈136 – 144 µV K−1 at ≈300 K. The increase in |S| with the additional Na2SO3 solution can be mainly attributed to the reduced n H, according to Equation (3) as shown below:

S=8π2kB2mT3eh2π3nH2/3 (3)

where k B is the Boltzmann constant, h is the Planck constant, m* is the effective mass, and e is the electron charge. Additionally, the temperature‐dependent n H is shown in Figure S3 (Supporting Information). Decrease in |S| values with increasing temperature of as mCSP‐ed samples, should be attributed to the increased n H with increasing temperature. Moreover, introducing an additional hierarchical porous structure with Na2SO3 residuals successfully changed the carrier scattering mechanism to acoustic phonon scattering (µ≈T −1.5) as shown in Figure S4 (Supporting Information). Using the measured values of |S| and n H, we estimated the DOS effective mass (m*) of Ag2Se‐30%NS at room temperature and compared with other studies[ 15 , 24 , 38 , 43 , 44 , 45 , 46 ] as shown in Figure S5 (Supporting Information). The m* of our study is 0.23m e, which is closely comparable to other state‐of‐the‐art Ag2Se‐based materials.[ 15 , 24 , 38 , 43 , 44 , 45 , 46 ]

Figure 3f plots the calculated S2σ. Without Na2SO3, the Ag2Se‐0%NS sample exhibits a S2σ of 23.8 µW cm−1 K−2 at ≈300 K, which increases to 27.0 µW cm−1 K−2 at 356 K. Utilizing Na2SO3 solution as the transient liquid in the mCSP enhances the overall S2σ, which approaches the maximum value of 26.5 and 33.8 µW cm−1 K−2 at ≈300 K and 356 K, respectively, for the Ag2Se‐10%NS sample. This can be attributed to better‐optimized n H closer to the optimal level of ≈1.0 × 1018 cm−3 of Ag2Se‐based materials.[ 38 ] The cycling test of thermoelectric performance and carrier transport of the reproduced Ag2Se‐30%NS sample is shown in Figures S6 and S7 (Supporting Information), respectively, which shows the performance is reliable and reproducible. Figure S8 (Supporting Information) shows the n H‐dependent room‐temperature S 2 σ of Ag2Se in this study compared with other state‐of‐the‐art studies.[ 15 , 24 , 38 , 43 , 44 , 45 , 46 ] As can be seen, the optimized n H for high S 2 σ is from ≈3 × 1018 cm−3 to ≈1 × 1019 cm−3. And the n H of our Ag2Se‐30%NS sample is ≈5.6 × 1018 cm−3, which is closely within the optimal level, contributing to high S 2 σ.

Figure 4 displays the thermal properties of the mCSP‐ed samples. Figure 4a plots the temperature‐dependent κ, which increases linearly with increasing temperature for all mCSP‐ed samples. Ag2Se‐0%NS exhibits the highest κ from 1.77 to 2.45 W m−1 K−1 in the temperature range from ≈300 to 356 K. Employing Na2SO3 solution as the transient liquid during mCSP results in decreased κ. With increasing the amount of Na2SO3 solution from 10 to 40 wt.%, the κ substantially reduces from 1.32 to 0.73 m−1 K−1 at ≈300 K. In addition, the differential scanning calorimetry (DSC) curves, as shown in Figure S9 (Supporting Information), indicate that the Na2SO3 residual does not affect the thermal stability of as mCSP‐ed Ag2Se samples.

Figure 4.

Figure 4

Temperature (T)‐dependent (a) total thermal conductivity (κ), b) electronic thermal conductivity (κ e), and (c) lattice thermal conductivity (κ l) of Ag2Se samples prepared via modified cold sintering process (mCSP) with varying quantities of Na2SO3 solution (0 – 40 wt.%). d) Room‐temperature κ l versus relative density of as mCSP‐ed Ag2Se samples at ≈300 K.

To understand the underlying mechanism for such low κ values, the κ e was calculated by using the Wiedemann–Franz law, κ e = LσT, where L is the Lorenz number.[ 34 ] Temperature‐dependent κ e is plotted in Figure 4b. As can be seen, increasing the quantity of Na2SO3 results in significantly reduced κ e, which has approached as low as 0.51 W m−1 K−1 at ≈300 K for the Ag2Se‐40%NS sample. The κ l values were calculated by subtracting of κ e from κ, and plotted against temperature as shown in Figure 4c. Generally, with increasing Na2SO3 from 0 to 40 wt.%, the κ l reduces. An ultralow κ l is observed in the Ag2Se‐30%NS sample, which approaches as low as 0.18 W m−1 K−1 at ≈300 K.

To understand the low κ l, the influence of porosity was evaluated on room‐temperature κ l based on the gray medium model[ 50 ] (more details of the gray medium model are provided in the supporting information), as shown in Figure 4d. As can be seen, the κ l values are significantly reduced with increasing porosity, induced by employing Na2SO3 as the transient liquid for mCSP. However, the increased porosity can only explain around half of the reduced κ l, whereas the additional reduction of κ l should be attributed to the presence of nanoscale Na2SO3 residuals. With the hierarchical porous structure and nanoscale Na2SO3 residuals functioning together as effective scattering centers for multi‐wavelength phonons,[ 51 , 52 , 53 ] exceptionally low κ l values can be realized in the mCSP‐ed Ag2Se samples.

The temperature‐dependent ZT values of the mCSP‐ed Ag2Se were calculated and presented in Figure 5a. The Ag2Se‐0%NS sample exhibits the lowest ZT of ≈0.4 across the entire temperature range. Introducing Na2SO3 solution during mCSP significantly improves the ZT values, leading to the peak ZT of 1.04 at ≈300 K in the Ag2Se‐30%NS sample. Additionally, the average ZT (ZT ave) was calculated following the equation ZTave=1ΔTTcThZTdT [ 16 ] where T h and T c are the highest and lowest temperature values in the range, and compared with other literature, as shown in Figure 5b. A high ZT ave of 0.97 can be observed across the temperature range from ≈300 to 356 K, which is higher than most literature and comparable with the state‐of‐the‐art values.[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ] Figure 5c illustrates the temperature‐dependent ZT values of this study compared to other literature.[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ] Different from other studies where the peak ZT values are generally observed at ≈330 – 380 K, our study presents a peak ZT of 1.04 at near room temperature, which is comparable with other state‐of‐the‐art Ag2Se‐based materials, and the key reason for a high ZT ave in our study. The decrease in ZT of the Ag2Se‐30%NS sample with increasing temperature can be attributed to high n H (as shown in Figure S3, Supporting Information), corresponding σ and κ e. Correspondingly, at a higher temperature range, the κ e plays a more significant role compared to κ l, resulting in an overall increase in κ. Figure S10 (Supporting Information) compares the n H‐dependent ZT of Ag2Se at room temperature in this study with other state‐of‐the‐art studies.[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ] The predicted n H‐dependent ZT curves, with differenced κ l values, were reproduced based on previous literature.[ 43 ] The optimized n H for high ZT is from ≈2 × 1018 cm−3. Through Se vacancy engineering, the reduced n H is getting closer to the optimal level and comparable with other state‐of‐the‐art studies,[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ] which contributes to a high ZT together with the reduced κ l.

Figure 5.

Figure 5

a) Temperature (T)‐dependent dimensionless figure of merit (ZT) of Ag2Se samples prepared via modified cols sintering process (mCSP) with varying quantities of Na2SO3 solution (0 – 40 wt.%). Comparison of (b) average ZT (ZT ave), and c) temperature‐dependent ZT between this study and other state‐of‐the‐art values.[ 15 , 24 , 37 , 38 , 43 , 44 , 45 , 46 , 47 , 48 ]

3. Conclusion 

In this study, by introducing Na2SO3 solution as the transient liquid during mCSP, a hierarchical structure, including micropores, nanopores, sub‐nanopores, and additional nanoscale Na2SO3 residuals, can be achieved in Ag2Se, which is attributed to the partial decomposition of Na2SO3 solution during the mCSP. Such a hierarchical structure contributes to an ultralow κ l of 0.18 W m−1 K−1 at ≈300 K in the Ag2Se‐30%NS. Simultaneously, additional Se vacancies have been observed, contributing to reduced n H from 13.5 × 1018 cm−3 of Ag2Se‐0%NS to ≈5.6 × 1018 cm−3 of Ag2Se‐30%NS at 300 K, which is closer to the optimal level of ≈1.0 × 1018 cm−3.[ 38 ] The optimized n H results in a high S 2 σ of ≈25 µW cm−1 K−2 at ≈300 K in Ag2Se‐30%NS. Consequently, an untrahigh room‐temperature ZT of 1.04 at ≈300 K and a high ZT ave of 0.97 from ≈300 to 356 K can be achieved in Ag2Se‐30%NS. This study demonstrates that introducing additional transient liquid solutions in mCSP can effectively realize hierarchical structure features, contributing to ultralow κ l and high thermoelectric performance.

4. Experimental Section

Synthesis of Ag2Se Powder

Silver nitrate (AgNO3, AR, RCI‐Labscan), selenium (Se) powder (100 mesh, 99.5%, Aldrich), and sodium sulfite (Na2SO3, AR, QRec) were purchased and used as raw materials without additional purification. Ag2Se powders were synthesized via a chemical reaction route.[ 54 ] Initially, 1.5125 g of Na2SO3 was dissolved in 60 mL of deionized (DI) water, and then 0.4739 g of Se powder was added to the solution. The mixture was continuously stirred using a magnetic stirrer at 100 °C for 30 min, followed by natural cooling to room temperature. The resulting clear yellow solution indicated the formation of the Na2SeSO3 compound. Subsequently, in another beaker, 2.0386 g of AgNO3 was dissolved in 120 mL of DI water. The Na2SeSO3 solution was gradually added into the AgNO3 solution and stirred steadily at room temperature for 2 h. Ag2Se precipitate was formed following the chemical reaction described by Equation (4).[ 55 ] The resulting dark gray Ag2Se precipitate underwent several washes with DI water and ethanol before being dried in an oven at 80 °C overnight.

2AgNO3+Na2SeSO3+H2ONg2Se+2NaNO3+H2SO4 (4)

Modified Cold Sintering Process of Bulk Ag2Se

Ag2Se bulk samples were prepared using the mCSP method using a 1 M Na2SO3 solution as the transient liquid. In a typical procedure, 1.2 g of as‐synthesized Ag2Se powder was manually blended with the Na2SO3 solution in a mortar. The mixture was then placed into a 12.7 mm inner diameter die, which was subsequently heated to 170 °C at a rate of 5 °C min−1 without applying external pressure. Once the desired temperature was reached, a uniaxial pressure of 620 MPa was applied and maintained for 1 h to consolidate the Ag2Se powders. Afterward, the pressure was released, and the sample was allowed to cool naturally to room temperature. To investigate the effect of the amount of Na2SO3 solution on the microstructure and properties of the fabricated Ag2Se pellets, Na2SO3 solutions were comprised of 10, 20, 30, and 40 wt.% of Na2SO3, relative to the total mass of Na2SO3 and Ag2Se. These samples were designated as Ag2Se‐10%NS, Ag2Se‐20%NS, Ag2Se‐30%NS, and Ag2Se‐40%NS, respectively. For reference, another Ag2Se pellet was prepared via mCSP without the Na2SO3 solution, namely Ag2Se‐0%NS, using DI water equivalent to 30 wt.% of the total mass of DI water and Ag2Se.

Material Characterizations

The phase and crystalline structure of the as‐synthesized Ag2Se powder and the mCSP bulk Ag2Se samples were determined using X‐ray diffraction (XRD, Panalytical, Empyrean). Field emission scanning electron microscopy (FE‐SEM) and energy‐dispersive X‐ray spectroscopy (EDS, FEI, Helios NanoLab G3 CX) were employed to observe the morphologies and elemental composition of the Ag2Se powder and mCSP bulk pellets. Furthermore, a focused ion beam (FIB) equipped in the SEM instrument, was utilized to prepare a thin lamella for detailed microstructural investigation using a transmission electron microscope (TEM, Thermo Scientific, Talos F200X).

Measurements of Thermoelectric Properties

Electrical conductivity (σ) and Seebeck coefficient (S) were measured simultaneously using the LSR‐3 instrument (Linseis) under a He atmosphere from ≈300 to 383 K. The Hall coefficient (R H) was measured using the van der Pauw method under a magnetic field of 0 ‐1 T, employing the electron transport option (ETO) in the VersaLab instrument (Quantum Design). The Hall carrier concentration (n H) and Hall carrier mobility (µ H) were calculated using the relationships: n H = 1/(eR H) and µ H = σ/(en H), respectively. Thermal conductivity (κ) was calculated using the equation κ = αρc p, where α is thermal diffusivity measured using the laser flash method in Ar atmosphere (LFA‐500, Linseis). ρ is density determined through an Archimedes’ method in water following the relationship: ρ = W a ρ w /(W a W w), where W a, W w, and ρ w represent the weight in air, the weight in water, and the density of water, respectively. The error range of relative density in the study was determined by three repeated measurements under identical conditions. C p is a specific heat capacity of as mCSP‐ed samples (Figure S11, Supporting Information) measured by a differential scanning calorimeter (DSC, Rigaku, Thermo plus evo2).

5. Computational Details

Density‐functional theory (DFT) calculations were performed using the all electron projected augmented wave (PAW) method, as implemented in the Vienna Ab initio Simulation Package (VASP).[ 56 , 57 , 58 , 59 , 60 , 61 ] The generalized gradient approximation (GGA) with the fully relativistic Perdew‐Burke‐Ernzerhof (PBE) functional was employed to treat the exchange correlations.[ 62 ] The valence wave functions were expanded in a plan‐wave basis with a cut‐off energy of 400 eV. All atoms were allowed to relax in their geometric optimizations until the Hellmann–Feynman force is less than 1 × 10−2 eV·Å−1. The convergence criterion and the Monkhorst‐Pack k‐mesh adopted for ionic relaxation are 1 × 10−7 eV per electron and 0.03π per Å, respectively. A denser 0.02π Å−1 Monkhorst‐Pack k‐mesh was adopted for calculating density‐of‐state (DOS), and a line‐mode k‐path based on Brillouin path features indicated by the AFLOW framework was adopted for calculating band structures.[ 63 , 64 ] To precisely predict bandgap, the Hubbard U model was considered, with the on‐site coulombic (U) and the exchange (J) terms combined in a single effective U parameter of 5.8 eV for Ag_4d orbitals, determined using the linear corresponding method. To eliminate the influence of the superpositions of wavefunctions, the band unfolding method was also used to reflect the band structure in the first Brillouin zone.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

SMLL-21-2411798-s001.docx (22.9MB, docx)

Acknowledgements

P.K. and W.‐D.L. contributed equally to this work. This research project was supported by the Fundamental Fund of Khon Kaen University from the National Science, Research and Innovation Fund (NSRF), and the National Research Council of Thailand (NRCT) (N41A650078, N42A650237). ZGC thanks the financial support from the Australian Research Council and the QUT Capacity Building Professor Program. WDL acknowledges the Queensland Cyber Infrastructure Foundation Ltd (Project pe62) for providing computation allocation. This work is supported by the use of the Central Analytical Research Facility hosted by the Institute for Future Environments at QUT. The authors acknowledge the National Computational Infrastructure, supported by the Australian government, for providing computational resources and service.

Open access publishing facilitated by Queensland University of Technology, as part of the Wiley ‐ Queensland University of Technology agreement via the Council of Australian University Librarians.

Kitthonbancha P., Liu W.‐D., Li M., Pinitsoontorn S., Chen Z.‐G., Transient Liquid Induced Hierarchical Structure Contributes to High Thermoelectric Performance in Ag2Se. Small 2025, 21, 2411798. 10.1002/smll.202411798

Contributor Information

Supree Pinitsoontorn, Email: psupree@kku.ac.th.

Zhi‐Gang Chen, Email: zhigang.chen@qut.edu.au.

Data Availability Statement

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

References

  • 1. Liu Y., Wang Z., Li J., Zhang M., Liu Y., Han X., Zuo X., Cai K., Chen L., ACS Appl. Mater. Interfaces 2024, 16, 26417. [DOI] [PubMed] [Google Scholar]
  • 2. d'Angelo M., Galassi C., Lecis N., Energies 2023, 16, 6409. [Google Scholar]
  • 3. Zhang Q., Deng K., Wilkens L., Reith H., Nielsch K., Nat. Electron. 2022, 5, 333. [Google Scholar]
  • 4. Hu B., Shi X.‐L., Zou J., Chen Z.‐G., Chem. Eng. J. 2022, 437, 135268. [Google Scholar]
  • 5. Zaferani S. H., Sams M. W., Ghomashchi R., Chen Z.‐G., Nano Energy 2021, 90, 106572. [Google Scholar]
  • 6. Bahk J.‐H., Fang H., Yazawa K., Shakouri A., J. Mater. Chem. C 2015, 3, 10362. [Google Scholar]
  • 7. Liu E., Negm A., Howlader M. M. R., Mater. Today Energy 2021, 20, 100625. [Google Scholar]
  • 8. Yang D., Zhang D., Ao D., Nisar M., Mansoor A., Chen Y., Li F., Ma H., Liang G., Zhang X., Fan P., Zheng Z., Nano Energy 2023, 117, 108930. [Google Scholar]
  • 9. Wu B., Wei W., Guo Y., Hou Yip W., Tay B. K., Hou C., Zhang Q., Li Y., Wang H., Chem. Eng. J. 2023, 453, 139749. [Google Scholar]
  • 10. Guan Q.‐L., Dong L.‐Q., Hao Q., Materials 2023, 16, 1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Pei J., Cai B., Zhuang H.‐L., Li J.‐F., Natl. Sci. Rev. 2020, 7, 1856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Saberi Y., Sajjadi S. A., J. Alloys Compd. 2022, 904, 163918. [Google Scholar]
  • 13. Li Y., Lou Q., Yang J., Cai K., Liu Y., Lu Y., Qiu Y., Lu Y., Wang Z., Wu M., He J., Shen S., Adv. Funct. Mater. 2021, 32, 2106902. [Google Scholar]
  • 14. Jin M., Liang J., Qiu P., Huang H., Yue Z., Zhou L., Li R., Chen L., Shi X., J. Phys. Chem. Lett. 2021, 12, 8246. [DOI] [PubMed] [Google Scholar]
  • 15. Jood P., Chetty R., Ohta M., J. Mater. Chem. A 2020, 8, 13024. [Google Scholar]
  • 16. Wang P., Chen J.‐L., Zhou Q., Liao Y. T., Peng Y., Liang J. S., Miao L., Appl. Phys. Lett. 2022, 120, 193902. [Google Scholar]
  • 17. Mi W., Qiu P., Zhang T., Lv Y., Shi X., Chen L., Appl. Phys. Lett. 2014, 104, 133903. [Google Scholar]
  • 18. Rajkumar V. B., Chen S.‐W., Calphad 2018, 63, 51. [Google Scholar]
  • 19. Lim K. H., Wong K. W., Liu Y., Zhang Y., Cadavid D., Cabot A., Ng K. M., J. Mater. Chem. C 2019, 7, 2646. [Google Scholar]
  • 20. Conn J. B., Taylor R. C., J. Electrochem. Soc. 1960, 107, 977. [Google Scholar]
  • 21. Hou S., Liu Y., Yin L., Chen C., Wu Z., Wang J., Luo Y., Xue W., Liu X., Zhang Q., Cao F., Nano Energy 2021, 87, 106223. [Google Scholar]
  • 22. Wu H., Shi X.‐l., Duan J., Liu Q., Chen Z.‐G., Energy Environ. Sci. 2023, 16, 1870. [Google Scholar]
  • 23. Yang D., Shi X.‐L., Li M., Nisar M., Mansoor A., Chen S., Chen Y., Li F., Ma H., Liang G. X., Zhang X., Liu W., Fan P., Zheng Z., Chen Z.‐G., Nat. Commun. 2024, 15, 923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Lin S., Guo L., Wang X., Liu Y., Wu Y., Li R., Shao H., Jin M., J. Materiomics 2023, 9, 754. [Google Scholar]
  • 25. Wang H., Liu X., Zhou Z., Wu H., Chen Y., Zhang B., Wang G., Zhou X., Han G., Acta Mater. 2022, 223, 117502. [Google Scholar]
  • 26. Wu H., Shi X. L., Mao Y., Li M., Liu W. D., Wang D. Z., Yin L. C., Zhu M., Wang Y., Duan J., Liu Q., Chen Z. G., Adv. Energy Mater. 2023, 13, 2302551. [Google Scholar]
  • 27. Ahmad S., Singh A., Bhattacharya S., Navaneethan M., Basu R., Bhatt R., Sarkar P., Meshram K. N., Muthe K. P., Vitta S., Aswal D. K., ACS Appl. Energy Mater. 2020, 3, 7113. [Google Scholar]
  • 28. Liang X., Wang C., Jin D., Appl. Phys. Lett. 2020, 117, 213901. [Google Scholar]
  • 29. Ivanov O., Maradudina O., Lyubushkin R., Mater. Charact. 2015, 99, 175. [Google Scholar]
  • 30. Dylla M. T., Kuo J. J., Witting I., Snyder G. J., Adv. Mater. Interfaces 2019, 6, 1900222. [Google Scholar]
  • 31. Rahman J. U., Van Du N., Nam W. H., Shin W., Lee K., Seo W.‐S., Kim M.‐H., Lee S., Sci. Rep. 2019, 9, 8624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Lee S., Jung S.‐J., Park G. M., Hong J., Lee A. S., Baek S.‐H., Kim H., Park T. J., Kim J.‐S., Kim S. K., J. Eur. Ceram. Soc. 2023, 43, 3376. [Google Scholar]
  • 33. Takashiri M., Miyazaki K., Tanaka S., Kurosaki J., Nagai D., Tsukamoto H., J. Appl. Phys. 2008, 104, 084302. [Google Scholar]
  • 34. Huang S., Wei T. R., Chen H., Xiao J., Zhu M., Zhao K., Shi X., ACS Appl. Mater. Interfaces 2021, 13, 60192. [DOI] [PubMed] [Google Scholar]
  • 35. Xu B., Feng T., Agne M. T., Zhou L., Ruan X., Snyder G. J., Wu Y., Angew. Chem., Int. Ed. 2017, 56, 3546. [DOI] [PubMed] [Google Scholar]
  • 36. Nan B., Li M., Zhang Y., Xiao K., Lim K. H., Chang C., Han X., Zuo Y., Li J., Arbiol J., Llorca J., Ibáñez M., Cabot A., ACS Appl. Electron. Mater. 2023, 6, 2807. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Tee S. Y., Tan X. Y., Wang X., Lee C. J. J., Win K. Y., Ni X. P., Teo S. L., Seng D. H. L., Tanaka Y., Han M. Y., Inorg. Chem. 2022, 61, 6451. [DOI] [PubMed] [Google Scholar]
  • 38. Chen J., Sun Q., Bao D., Liu T., Liu W. D., Liu C., Tang J., Zhou D., Yang L., Chen Z. G., ACS Appl. Mater. Interfaces 2020, 12, 51523. [DOI] [PubMed] [Google Scholar]
  • 39. Wang D., Li L., Jiang J., Lu Z., Wang G., Song K., Zhou D., Reaney I. M., J. Mater. Res. 2021, 36, 333. [Google Scholar]
  • 40. Grasso S., Biesuz M., Zoli L., Taveri G., Duff A. I., Ke D., Jiang A., Reece M. J., Adv. Appl. Ceram. 2020, 119, 115. [Google Scholar]
  • 41. Lu X., Lu W., Gao J., Liu Y., Huang J., Yan P., Fan Y., Jiang W., ACS Appl. Mater. Interfaces 2022, 14, 37937. [DOI] [PubMed] [Google Scholar]
  • 42. Zhu B., Su X., Shu S., Luo Y., Tan X. Y., Sun J., Sun D., Zhang H., Zhang Q., Suwardi A., Zheng Y., ACS Appl. Energy Mater. 2002, 5, 2022. [Google Scholar]
  • 43. Palaporn D., Pinitsoontorn S., Kurosaki K., Snyder G. J., Adv. Mater. Technol. 2024, 9, 2301242. [Google Scholar]
  • 44. Jakhar N., Kedia D. K., Kumar A., Saurabh K., Singh S., Appl. Phys. Lett. 2023, 122, 163901. [Google Scholar]
  • 45. Palaporn D., Kurosaki K., Pinitsoontorn S., Adv. Energy Sustainability Res. 2023, 4, 2300082. [Google Scholar]
  • 46. Li D., Zhang B. L., Ming H. W., Wang L., Zu Y., Qin X. Y., ACS Appl. Mater. Interfaces 2021, 13, 34543. [DOI] [PubMed] [Google Scholar]
  • 47. Abusa Y., Yox P., Viswanathan G., Opare‐Addo J., Sarkar A., Kyveryga V., Smith E., Lebedev O. I., Kovnir K., J. Am. Chem. Soc. 2024, 146, 11382. [DOI] [PubMed] [Google Scholar]
  • 48. Kleinhanns T., Milillo F., Calcabrini M., Fiedler C., Horta S., Balazs D., Strumolo M. J., Hasler R., Llorca J., Tkadletz M., Brutchey R. L., Ibáñez M., Adv. Energy Mater. 2024, 14, 2400408. [Google Scholar]
  • 49. Feng B., Cheng Y., Liu C., Gao J., Wu G., Bai X., Si R., Li W., Guo Y., Miao L., ACS Appl. Energy Mater. 2023, 6, 2804. [Google Scholar]
  • 50. Tarkhanyan R. H., Niarchos D. G., APL Mater. 2014, 2, 076107. [Google Scholar]
  • 51. Hao Q., Zhao H., Xiao Y., Xu D., J. Appl. Phys. 2018, 123, 014303. [Google Scholar]
  • 52. Jin Q., Shi W., Zhao Y., Qiao J., Qiu J., Sun C., Lei H., Tai K., Jiang X., ACS Appl. Mater. Interfaces 2018, 10, 1743. [DOI] [PubMed] [Google Scholar]
  • 53. Liu W.‐D., Yin L.‐C., Li L., Yang Q., Wang D.‐Z., Li M., Shi X.‐L., Liu Q., Bai Y., Gentle I., Wang L., Chen Z.‐G., Energy Environ. Sci. 2023, 16, 5123. [Google Scholar]
  • 54. Palaporn D., Mongkolthanaruk W., Tanusilp S., Kurosaki K., Pinitsoontorn S., Appl. Phys. Lett. 2022, 120, 073901. [Google Scholar]
  • 55. Ge J. P., Xu S., Liu L. P., Li Y. D., Chem.–Eur. J. 2006, 12, 3672. [DOI] [PubMed] [Google Scholar]
  • 56. Kresse G., Hafner J., Phys. Rev. B 1994, 49, 14251. [DOI] [PubMed] [Google Scholar]
  • 57. Kresse G., Hafner J., Phys. Rev. B 1993, 47, 558. [DOI] [PubMed] [Google Scholar]
  • 58. Kresse G., Furthmüller J., Comput. Mater. Sci. 1996, 6, 15. [Google Scholar]
  • 59. Kresse G., Hafner J., J. Phys. Condens. Matter. 1994, 6, 8245. [DOI] [PubMed] [Google Scholar]
  • 60. Kresse G., Furthmüller J., Phys. Rev. B 1996, 54, 11169. [DOI] [PubMed] [Google Scholar]
  • 61. Kresse G., Joubert D., Phys. Rev. B 1999, 59, 1758. [Google Scholar]
  • 62. Perdew J. P., Burke K., Ernzerhof M., Phys. Rev. Lett. 1996, 77, 3865. [DOI] [PubMed] [Google Scholar]
  • 63. Setyawan W., Curtarolo S., Comput. Mater. Sci. 2010, 49, 299. [Google Scholar]
  • 64. Tran F., Blaha P., Phys. Rev. Lett. 2009, 102, 226401. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

SMLL-21-2411798-s001.docx (22.9MB, docx)

Data Availability Statement

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


Articles from Small (Weinheim an Der Bergstrasse, Germany) are provided here courtesy of Wiley

RESOURCES