Thermoelectric materials have attracted significant attention as eco-friendly and sustainable energy production technologies because of their ability of convert heat to electricity through the Seebeck effect [1-3]. The thermoelectric efficiency of materials is expressed as a dimensionless figure of merit, zT = S²σT/κ_tot, where σ, S, κ_tot, and T are the electrical conductivity, Seebeck coefficient, thermal conductivity, and absolute temperature, respectively. To enhance zT, σ and S must be increased while κ_tot must be decreased. However, it is not simple to enhance zT because of trade-off relationship between σ and S. Therefore, a lot of tactics such as band structure engineering, carrier scattering and nanostructing have been studied to increase zT [4-6].

The Bi-Te alloy is a representative chalcogenide with high thermoelectric performance in the room-temperature range. Because of their low thermal conductivity and relatively high electrical conductivity, Bi-Te alloys are used for cooling systems in small-scale thermoelectric generation modules in the low-temperature range [7]. The Bi₂Te₃ in Bi-Te alloys has a trigonal crystal structure with a quintuple-layer structure that is stacked repeatedly in the order of Te-Bi-Te-Bi-Te in the c-axis direction. Because it has strong covalent bonds and ionic bonds in the ab-plane direction and Van der Waals bonds in the c-axis direction, Bi₂Te₃ has outstanding electrical performance with high σ, high S, and low κ_tot.

Bismuth selenide (Bi₂Se₃), which is one of the chalcogenides and has the same rhombohedral structure as Bi₂Te₃, has been widely studied as a thermoelectric material as well [8-9]. Kang et al. reported that polycrystalline Bi₂Se₃ exhibited a zT value of 0.12 at room temperature [10]. Sun et al. reported that single-layer-based Bi₂Se₃ exhibited a zT value of 0.1 at room temperature and a maximum zT value of 0.35 at 400 K [11]. Hong et al. reported that Bi₂Se₃ nanosheets exhibited a zT value of 0.17 at room temperature and a maximum zT value of 0.48 at 427 K [12]. Because the experimentally reported thermoelectric performance of Bi₂Se₃ is low, several attempts have been made to increase that thermoelectric performance, with one of them by doping [13-14]. Li et al. doped Bi₂Se₃ with Sn using the composition Bi_1.93Sn_0.07Se₃ and obtained a zT increase of 60% compared with that of pure Bi₂Se₃ [15]. P. Janíček et al. reported that doping Bi₂Se₃ with thallium enhanced the power factor by 25% [16].

Meanwhile, adding Cu to Bi₂Te₃ has been reported to enhance the thermoelectric performance [17-18]. Kim et al. reported that Bi₂Te₃ with Cu addition enhanced thermoelectric performance and improved electronic transport properties [19]. In this study, the thermoelectric transport properties of Cu-added Bi₂Se₃ polycrystalline alloys were investigated because Bi₂Se₃ has the same structure as Bi₂Te₃ in the bismuth chalcogenide family. The thermoelectric transport properties, including σ, S, and κ_tot, were measured, and the phenomenological transport parameters, including density of state effective mass (m_d^*), weighted mobility (μ_w), and thermoelectric quality factor (B), were calculated to analyze the electronic transport properties of the Cu-added Bi₂Se₃.

Cu_xBi₂Se₃ (x = 0, 0.004, 0.008, 0.012, and 0.016) polycrystalline samples were stoichiometrically synthesized via a solid-state reaction. High-purity elemental Bi (99.999%), Se (99.999%), and Cu (99.998%) powders were mixed and loaded into vacuum quartz tubes. After melting at 800 °C for 4 h, the samples were cooled to room temperature in a furnace. The synthesized ingot samples were pulverized into powders by high-energy ball milling (SPEX 8000D, SPEX). The samples were sintered via spark plasma sintering (SPS, SPS-1030, Sumitomo Coal Mining Co., Ltd., Japan) under vacuum at 350 °C for 10 min at 70 MPa. The crystalline phase of the samples was identified by X-ray diffraction (XRD, D8 Discover, Bruker) with Cu K_α1 radiation at 40 kV and 40 mA. The values of S and σ were measured using a thermoelectric evaluation system (ZEM-3M8, Advance Riko, Japan) along the perpendicular direction of the SPS pressing direction in a He atmosphere. Hall measurement was performed at room temperature in the same direction using the Van der Pauw method to determine the carrier concentration (n_H) and mobility (μ_H). The κ_tot value of the samples was calculated using the theoretical density (ρ_s), heat capacity (C_p), and thermal diffusivity (λ) (κ_tot = ρ_sC_pλ). The λ value of each sample was measured by laser flash system analysis (LFA457, Netsch). The zT values of the samples were calculated based on measured data.

Figure 1(a) and (b) show the XRD patterns of the sintered Cu_xBi₂Se₃ samples (x = 0, 0.004, 0.008, 0.012, and 0.016). As shown in Figure 1(a), a single-phase Bi₂Se₃ with a rhombohedral structure was synthesized without impurities. Figure 1(b) shows the lattice parameters a and c calculated from the XRD data with error bars. No variation was observed in the lattice parameter a with the varying Cu content. However, the lattice parameter c varied with the Cu content, increasing from 28.64 to 28.67 Å as x varied from 0 to 0.008, and then exhibiting a decreasing tendency beyond x = 0.008. This behavior was also observed in the study of Cu doping of Bi₂Te₃ [19]; Cu was intercalated into the van der Waals gap of Bi₂Te₃ (lattice parameter c increased) with small Cu additions, while Cu also entered the substitutional site of Bi (lattice parameter c decreased). Thus, it was speculated that the increase in the lattice parameter c up to x = 0.008 can be attributed to Cu intercalation, and the decrease beyond x = 0.008 is due to Cu substitution of Bi site.

Figure 2(a) shows the temperature dependence of σ. Semiconducting conduction behavior was observed for all samples, as σ increased with temperature. Moreover, the value of σ for the Cu_xBi₂Se₃ samples (x = 0.004, 0.008, 0.012, 0.016) was higher than that of the pristine Bi₂Se₃. At 300 K, the value of σ for the pristine Bi₂Se₃ sample was 164 S/cm, and it increased to 245, 286, 239, and 214 S/cm for the samples with x = 0.004, 0.008, 0.012, and 0.016, respectively. The inset of Figure 2(a) shows the σ values as a function of x at 520 K, indicating that the σ values of all the Cu-added samples were higher than that of the pristine Bi₂Se₃ sample.

Furthermore, the results in the inset of Figure 2(a) are similar to those of a previous study that showed that the increase in the value of σ up to x = 0.008 was due to Cu occupying the interstitial sites, and the decrease beyond x = 0.008 was attributed to the substitution of Cu in the Bi site [19]. Figure 2(b) shows the temperature dependence of the S of the samples, the negative values of S indicate that all the samples are n-type semiconductors. The S values of all the Cu-added samples were smaller than that of the pristine Bi₂Se₃ sample over the entire temperature range. The inset of Figure 2(b) shows the S values as a function of x at 520 K, indicating that the S values of the Cu-added samples were lower than that of pristine Bi₂Se₃. Similar results were obtained in a previous study that showed that the decrease in S values up to x = 0.008 indicates interstitial site doping, and the increase beyond x = 0.008 indicates interstitial and substitutional doping [19].

Figure 2(c) shows the power factor values of the samples calculated from the measured σ and S values. The power factor values of the Cu-added samples were higher than that of the pristine Bi₂Se₃ sample at all measured temperatures. The maximum power factor values were observed for Cu_0.008Bi₂Se₃ sample 0.83 mW/mK² in the lower temperature range (T < 360 K). However, in the high temperature range, the maximum power factor values were observed for Cu_0.016Bi₂Se₃ sample (T > 360 K). The power factor values of the pristine Bi₂Se₃ sample improved, from 0.54 to 0.83 mW/mK² (x = 0.008), at room temperature and from 0.59 to 0.80 mW/mK² (x = 0.016) at 520 K (inset of Figure 2(c).

Figure 3(a) and (b) show the measured Hall carrier concentration (n_H) and Hall mobility (μ_H) of the samples at 300 K. As shown in Figure 3(a), n_H increased from 1.94 × 10¹⁸ to 2.81 × 10¹⁸ cm^-3 owing to a small amount of Cu doping (Cu_0.008Bi₂Te₃), then decreased to 2.17 × 10¹⁸ cm^-3 beyond x = 0.008. Based on a previous study, it can be speculated that excess Cu atoms are preferentially located at the van der Waals gap in the sample with lower Cu content (x ≤ 0.008), while incorporated Cu atoms at Bi-site occur in the samples with higher Cu content beyond x = 0.008 [19]. Figure 1(b) shows good agreement with these results because the ionic radii of Cu¹⁺ and Cu²⁺are smaller than that of Bi³⁺. Figure 3(b) shows the μ_H values measured at 300 K. The μ_H of the samples were 605, 755, 658, 741, and 685 cm²/Vs for x = 0, 0.004, 0.008, 0.012, and 0.016, respectively.

Figure 3(c) shows the weighted mobility (μ_w) as a function of temperature for the samples. μ_w is proportional to the maximum power factor that a sample can attain when n_H is optimized. The μ_w value of each sample was calculated from the measured σ and S using a simple analytical form that approximates the exact Drude-Sommerfield free-electron model given in Equation (1) for |S|>20 μV/K [20].

where m_e is the electronic mass. The calculated μ_w values at 520 K were 29.5, 37.3, 38.3, 39.5; and 42.7 cm²/Vs for x = 0, 0.004, 0.008, 0.012; and 0.016, respectively. The μ_w value increases as the Cu content increases, which is in agreement with the power factor trend over the whole temperature range.

Figure 3(d) shows the calculated m_d^* of the Cu_xBi₂Se₃ (x = 0, 0.004, 0.008, 0.012, and 0.016) samples with respect to x. The inset of Figure 3(d) shows log₁₀n_H as a function of |S|. The m_d^* values were calculated based on a single parabolic band model using an acoustic phonon-scattering mechanism [20].

The m_d^* values of the samples were 0.187, 0.177, 0.205, 0.176 and 0.202 m₀ for x = 0, 0.004, 0.008, 0.012 and 0.016, respectively. It can be speculated that the m_d^* value increased because of Cu intercalation for x = 0 to at x = 0.008, and that it decreased because of Cu substitution for x = 0.012 and 0.016, based on a previous study [19].

Figure 4(a) and (b) show the κ_tot and lattice thermal conductivity (κ_latt) as a function of temperature for the Cu_xBi₂Se₃ (x = 0, 0.004, 0.008, 0.012, 0.016) samples. The inset of Figure 4(a) shows the electronic thermal conductivity (κ_elec), calculated using the Wiedemann Franz law [21]: κ_elec = LσT, where L is the Lorenz number. L is calculated using the equation L = 1.5 + exp(-|S|/116) (where L is in 10⁻⁸ WΩK⁻² and S in μV/K) [22]. The κ_latt value was obtained by subtracting κ_elec from κ_tot. The κ_latt values for the Cu-added samples were lower than that of the pristine Bi₂Se₃ sample over the entire temperature range. The κ_latt value significantly decreased from 0.96 W/mK for the pristine sample to 0.52–0.61 W/mK for Cu-added samples at 520 K. The decrease in the κ_latt values with Cu content is due to point defect phonon scattering caused by the Cu addition.

Figure 5(a) shows the temperature dependence of zT for each sample calculated from the σ, S, and κ_tot values. As shown in Figure 5(a), the zT value of the Cu-added Bi₂Se₃ samples were higher than that of the pristine Bi₂Se₃ sample. The highest zT value (0.38) was achieved for the sample with x = 0.016 at 520 K, which was ~48% higher than that of the pristine Bi₂Se₃ sample. The increase in zT for all samples over whole temperature range was attributed to the increase in the power factor and decrease in κ_latt.

Figure 5(c) shows the dimensionless thermoelectric quality factor (B) of each sample, calculated from μ_w and κ_latt. B is related to the maximum zT that a material can achieve when n_H is optimized; B was calculated using Equation (3) [20] as follows:

The value of B at 520 K were 0.083, 0.19, 0.19, 0.17 and 0.21 for x = 0, 0.004, 0.008, 0.012 and 0.016, respectively. Therefore, a zT improvement can be expected for all Cu-added samples, and was probably maximized at 520 K for x = 0.012 when n_H is optimized.

The thermoelectric transport properties of Cu-added Bi₂Se₃ (Cu_xBi₂Se₃, x = 0, 0.004, 0.008, 0.012, and 0.016) were investigated. A rhombohedral Bi₂Se₃ phase was synthesized without secondary phases. Cu doping increases the electrical conductivity and carrier concentration. As a result, the power factor values of the samples increases from 0.54 (x = 0) to 0.83 mW/mK² (x = 0.008) at room temperature and 0.59 (x = 0) to 0.80 mW/mK² (x = 0.016) at 520 K. A decrease in thermal conductivity was observed in the Cu-added samples because of the decrease in the lattice thermal conductivity. Consequently, the zT values of all Cu-added samples (Cu_xBi₂Se₃, x = 0.004, 0.008, 0.012, and 0.016) were enhanced compared with that of the pristine Bi₂Se₃ sample over the entire temperature range. The maximum zT value of 0.38 was observed for the Cu_0.016Bi₂Se₃ (x = 0.16) sample at 520 K, which exhibited a 48% improvement in zT compared with that of pristine Bi₂Se₃.