1. INTRODUCTION

As energy depletion and environmental pollution have grown as critical global challenges, demand for sustainable innovations across industries has intensified. Addressing these issues requires technologies that can efficiently generate, harvest, and store energy^[1–2]. Among various candidates, thermoelectric (TE) conversion, a solid-state platform for both energy harvesting and cooling, has attracted growing interest^[3]. TE materials can directly convert thermal energy (e.g., waste heat and solar heat) into electrical energy, and their reversible operation enables both power generation and cooling. This dual functionality makes TEs a compelling route toward sustainable and energy-efficient technologies^[4–6]. The performance of TE materials at a given temperature is commonly quantified by the dimensionless figure of merit, $ZT$, defined as $ZT = \sigma S^2 T / \kappa$, where $\sigma$, $S$, $T$, and $\kappa$ denote the electrical conductivity, Seebeck coefficient, absolute temperature, and total thermal conductivity, respectively. The $\kappa$ can be further expressed as the sum of lattice ($\kappa_L$) and electronic ($\kappa_e$) contributions^{[7–9, 23]}. Because higher $ZT$ generally translates into improved performance for both TE generators (TEGs) and TE coolers (TECs), developing high-$ZT$ TE materials remains essential for enhancing the overall efficiency of TE devices (TEDs)^[10–11].

TEDs are typically realized in three main configurations: three-dimensional (3D) bulk-based, two-dimensional (2D) film-based, and one-dimensional (1D) fiber-based architectures^[12]. While 3D bulk-based devices enabled early commercialization owing to their relatively high performance^[13], their intrinsic rigidity limits implementation in flexible and wearable applications^[14]. In contrast, 2D film-based devices offer superior mechanical flexibility^[15]. However, the practical deployment of flexible TEs remains challenging due to their limited stability and insufficient output performance, which are closely tied to fabrication constraints and an incomplete understanding of material and device behavior. Despite these challenges, film-based TEDs remain promising for flexible and portable power generation and cooling in the low-to-mid temperature range, where achieving a balance between performance and flexibility is critical^[16]. Accordingly, extensive efforts have focused on developing film-based TE materials such as Sb₂Te₃ ^[17–18], Bi₂Te₃ ^{[19, 24]}, and SnSe^[20].

2D film-based TE materials can be fabricated by various approaches, including physical vapor deposition (PVD) (e.g., sputtering and thermal evaporation, and pulsed laser deposition), chemical vapor deposition (CVD), solution-based coating methods (e.g., spin coating and doctor blading), and printing techniques (e.g., screen printing and inkjet printing). Among these, electrodeposition is particularly attractive for commercial-scale manufacturing because of its simple process flow, low processing cost, and straightforward scalability to large area, high throughput film fabrication^[21]. Nevertheless, electrodeposited Sb₂Te₃ films often contain a significant fraction of amorphous phases, which severely degrades their electrical conductivity^{[22, 23]}. Although mild annealing near 373 K can reduce amorphization, the room temperature power factor ($PF = S^2 \sigma$) often remains limited, primarily due to insufficient $\sigma$. To improve charge transport, incorporating Ag during electrodeposition has been reported to enhance $\sigma$ and $PF$. After annealing, Ag can be incorporated into the Sb₂Te₃ lattice as a hole dopant, while Ag₂Te nanoprecipitates formed during annealing can induce an energy filtering effect that preferentially suppresses low-energy carriers. This mechanism can help preserve $S$ while improving overall performance^[22]. Despite these advances, further simplification of the processing route and improved transport optimization remain important for electrodeposited Sb₂Te₃ films, particularly to overcome the $\sigma$–$S$ trade off.

In this study, we fabricated a polycrystalline Sb₂Te₃-based TE film by tuning the tartaric acid content in a two-step electrodeposition process combined with low-temperature annealing. Upon annealing at 393 K, a portion of the introduced Ag reacted with the Sb₂Te₃ matrix to form Ag and Ag₂Te precipitates, yielding a stable composite microstructure. As a result, the Ag-plated film annealed at 393 K exhibited markedly improved reliability while maintaining a high $PF$.

2. EXPERIMENTAL

2.1 Fabrication of polycrystalline Sb₂Te₃ film

As shown in Fig. 1a, the electrolyte for the electrodeposition of the Sb₂Te₃ films was prepared in two parts. First, 0.8 mM Sb₂O₃ was dissolved in 33 mM tartaric acid under magnetic stirring at 300 rpm and 60 ℃. Separately, 2.4 mM TeO₂ was dissolved in 1 M HNO₃. A Cr/Au layer was deposited on a SiO₂ substrate by thermal evaporation. A Cu tape was then attached to one side of the substrate to provide an electrical contact for electrodeposition. A Pt electrode and an Ag/AgCl electrode were used as the counter and reference electrodes, respectively. Electrodeposition was performed at room temperature at a fixed potential of -0.1 V (vs Ag/AgCl) for 3 h.

Fig. 1. Process schematic for fabricating Ag/Ag₂Te-integrated Sb₂Te₃ films via two-step plating: (a) Sb₂Te₃ electrodeposition and (b) electroless Ag plating.

3. RESULTS AND DISCUSSION

To confirm phase formation in the electrodeposited polycrystalline Sb₂Te₃ film and to track the evolution of Ag- and Ag₂Te-related phases introduced by electroless plating as a function of annealing temperature, XRD measurements were performed (Fig. 2). The electrodeposited Sb₂Te₃ film (ST) matched well with the standard diffraction pattern of hexagonal Sb₂Te₃ (space group R-3m, PDF#15-0874). No obvious impurity peaks or noticeable peak broadening were observed, indicating good crystallinity. In contrast, the Ag-plated film (AST) exhibited reflections from metallic Ag (PDF#41-1402) along with additional phases, including Ag₂Te (PDF#34-1042), Sb (PDF#35-0732), Te (PDF#18-1324), and Ag₃Sb (PDF#10-0452). These phases likely arise from concurrent reactions during Ag incorporation and Ag₂Te formation during the electroless plating process, and/or surface reactions promoted by the alkaline plating solutions. After annealing at 373 K (AST373), the phase composition remained largely similar to that of AST. By contrast, after annealing at 393 K (AST393), the diffraction peaks associated with Sb and Te disappeared, while the Ag and Ag₂Te peaks were clearly retained, suggesting additional reactions and phase redistribution at 393 K. Upon annealing at 423 K (AST423), the Sb₂Te₃ reflections markedly weakened, while Ag₂Te became the dominant phase. This behavior may be associated with the α-Ag₂Te to β-Ag₂Te transition above 418 K^[24] and/or Sb volatilization (or loss) at elevated temperatures. Overall, AST393 yields a rational Ag/Ag₂Te/Sb₂Te₃ multiphase architecture, which may help alleviate the $\sigma$–$S$ trade-off by providing high mobility conductive pathways (Ag and Ag₂Te) and interfacial energy barriers that can promote energy filtering.

Fig. 2. XRD patterns of ST, AST, AST373, AST393, and AST423 samples. Reference stick patterns for Sb₂Te₃ and Ag₂Te are shown at the bottom.

The surface morphologies of all samples were further examined by FE-SEM. As shown in Fig. 3a-e, all of the films, including ST, exhibit a plate-like microstructure, which is consistent with other reports on electrodeposited polycrystalline Bi₂Te₃ films^[25]. After electroless Ag plating (AST), Ag-related particles/precipitates appear on the surface, and the AST sample (Fig. 3b) shows several micron-scale coarse features. These coarse particles became less prominent with increasing annealing temperature (Fig. 3c–e), suggesting that additional reactions and microstructural rearrangement occur during annealing. Cross-sectional FE-SEM images (Fig. 3i-l) indicate that the film thickness remains nearly unchanged up to 393 K (approximately 4.1 – 4.5 μm). In contrast, annealing at 423 K led to a sharp increase in thickness to ~16.06 μm, accompanied by pronounced structural deformation. This trend is consistent with the XRD results in Fig. 2, where Ag₂Te becomes the dominant phase at elevated temperatures. The thickness increase and deformation are likely associated with excessive Ag₂Te precipitation and the development of a porous structure, resulting in film swelling.

Fig. 3. FE-SEM images and EDS elemental maps of Sb₂Te₃-based films. Top view FE SEM images of (a) ST, (b) AST, (c) AST373, (d) AST423, and (e) AST393. EDS elemental maps of the Ag-plated film showing the spatial distributions of (f) Ag, (g) Sb, and (h) Te in the AST393 sample. Cross sectional FE-SEM images of (i) ST, (j) AST, (k) AST393, and (l) AST423, with the corresponding film thickness indicated in the images.

In addition, EDS elemental maps (Fig. 3f-h) show relatively uniform distributions of Ag, Sb, and Te, implying that the Ag₂Te formed at 393 K is more likely present as finely dispersed precipitates within the matrix rather than as micron-scale surface particles^[22].

The room-temperature electrical transport properties of the Sb₂Te₃-based films, including $\sigma$, $S$, and $PF$ are summarized in Fig. 4. The ST film exhibited relatively high $\sigma$ compared with previously reported electrodeposited amorphous Sb₂Te₃ films, however, the batch-to-batch error was substantial; therefore, further optimization for stabilization is required, even though the resulting $PF$ was relatively high. In addition, because high $\sigma$ can increase $\kappa_e$ through the Wiedemann–Franz relation, changes in $\kappa$ should be considered when assessing the overall TE performance ($ZT$).

Fig. 4. Room-temperature electrical transport properties of the Sb₂Te₃-based films: (a) σ (left side) and S (right side) and (b) PF. Error bars indicate the reliability of samples and measurement uncertainty.

For the Ag-plated films, $n$ values (1.01 – 1.04 × 10¹⁹ cm^-3) slightly increased relative to ST (~5.1 × 10¹⁸ cm^-3). This increase can be attributed to hole doping by Ag⁺ substitution at Sb³⁺ sites and a reduction in antisite defects during annealing. Nevertheless, when multiple secondary phases (e.g., Te, Sb, and Ag₃Sb) coexist, charge transport can be impeded, leading to reduced $m$. Because these secondary phases remain in AST373, as shown in Fig. 2, the $PF$ is rather decreased. When the annealing temperature was increased to 393 K (AST393), $\sigma$ decreases and $S$ increases compared with ST, due to the trade-off between $\sigma$ and $S$. Moreover, the relatively higher $S$ despite the increase in $n$ might be related to an energy filtering effect, in which Ag and Ag₂Te selectively scatter low-energy holes, thereby increasing density-of-states effective mass ($m^*$) and $S$^[26]. The $m^*$ can be estimated based on Boltzmann transport theory as follows:

The calculated $m^*$ value for AST393 (0.35 $m_0$) is higher than that of ST (0.26 $m_0$), suggesting a carrier filtering effect induced by the presence of Ag and Ag₂Te precipitates. Consequently, owing to the rational Ag/Ag₂Te/Sb₂Te₃ multiphase architecture, a competitive and stable $PF$ value of ~648 μW m^-1 K^-2 was obtained for AST393.

For AST423, a similar multiphase structure was also observed, however, as confirmed by XRD, the Sb₂Te₃ phase was substantially weakened and Ag₂Te became dominant at this higher annealing temperature. Accordingly, $\sigma$ decreased while $S$ increased significantly, resulting in the highest $PF$ of ~1671 μW m^-1 K^-2. However, the porous morphology and severe structural deformation indicate degraded film quality, and the formation of excessive secondary phases may raise concerns regarding process reproducibility.

In this work, we introduce a processing strategy that combines electroless plating with subsequent annealing, and systematically examined how Ag incorporation and phase evolution influence electrical transport. This approach provides practical guidelines for process design and optimization toward high-performance TE thin films. Moreover, the electroless plating process may be further optimized to suppress film degradation and excessive secondary phase formation, and it could also be extended to other metals (e.g., Cu and Ni) for TE film applications.

4. CONCLUSIONS

In this study, polycrystalline Sb₂Te₃-based films were fabricated by tuning the tartaric acid concentration during electrodeposition, combined with low-temperature annealing. After electroless Ag plating, partial decomposition of Sb₂Te₃ in the alkaline solution led to the formation of Sb- and Te-based secondary phases, which adversely affected electrical transport. Subsequent annealing at 393 K suppressed most of these impurity phases and produced an Ag/Ag₂Te/Sb₂Te₃ multiphase architecture. This multiphase configuration, together with the high mobility of Ag and Ag₂Te and the associated energy filtering effect, can partially alleviate the trade-off between electrical conductivity and the Seebeck coefficient. As a result, the AST393 sample achieved an electrical conductivity of 3502 S cm^-1 while maintaining a Seebeck coefficient of 41.5 μV K^-1, with improved reproducibility, yielding a reliable room temperature power factor of 648 μW m^-1 K^-2. Overall, combining the electrodeposition of polycrystalline Sb₂Te₃ with electroless metal plating and low-temperature annealing offers a practical and scalable route for enhancing thermoelectric thin films for large-area processing.