Thermoelectric conversion technology, which could convert heat into electricity for power generation or vice versa convert electrical energy into heat for refrigeration, provides a possible solution for the energy shortage and environmental concerns.^[1–3] However, the low conversion efficiency limits the large-scale application of thermoelectric technology, which is determined by the dimensionless thermoelectric figure-of-merit (ZT). ZT is defined as ZT = S²σT/κ, where S is the Seebeck coefficient, σ the electric conductivity, T the absolute temperature, and κ the thermal conductivity, respectively.

Recently, much effort has been focused on ternary I–III–VI2-type (I = Cu, Ag; III = Al, Ga, In; and VI = S, Se, Te) chalcopyrite compounds owing to their nontoxicity, low lattice thermal conductivity, and good electrical transport properties.^[4–13] So far, many compounds with high ZT values have been reported, for example, CuGaTe₂ with a ZT value of 1.4,^[14] Ag_0.95GaTe₂ with a ZT value of 0.77,^[15] Cu_{1 − x}Mn_xInTe₂ with a ZT value of 0.84^[16] and CuInTe_1.99Sb_0.01 + 1 wt% ZnO system with a ZT value of 1.61.^[4] Among these ternary chalcopyrite compounds, p-type CuInTe₂ are characterized by relatively high Seebeck coefficient and low high-temperature thermal conductivity, owing to its degenerate valence bands^[4] and highly distorted crystal structure,^[7] respectively, which makes it a promising high-temperature thermoelectric material. Nevertheless, the further application of un-doped CuInTe₂ is still retarded by its poor electrical conductivity and high room-temperature thermal conductivity. Therefore, it is desirable to develop a strategy to improve the electrical conductivity of CuInTe₂ and simultaneously suppress its lattice thermal conductivity.^[4,12] In this work, we find that Mn is a desired doping element to improve both the electrical and thermal transport properties of CuInTe₂. Through replacing part of In with Mn, the electrical conductivity of the CuIn_{1 − x}Mn_xTe₂ sample increases dramatically since Mn element acts as the acceptor to provide additional charge carriers (hole). Meanwhile, the elemental substitution leads to obviously reduced lattice thermal conductivity due to the enhanced phonon scattering. Therefore, enhanced thermoelectric performance is obtained for CuIn_{1 − x}Mn_xTe₂, resulting from synergistic effect of optimized electrical conductivity and reduced lattice thermal conductivity.

Polycrystalline samples of CuIn_{1 − x}Mn_xTe₂ (x = 0, 0.025, 0.05, 0.075, 0.1) were synthesized by a melt-annealing method. The starting materials Cu (99.99%, shot), In (99.999%, shot), Mn (99.999%, piece), Te (99.999%, shot) were weighed according to the stoichiometric ratio and sealed in evacuated quartz tubes, which were heated to 1173 K at a rate of 1 K/min and held at this temperature for 24 h. Then, the samples were slowly cooled to 923 K, annealed at 923 K for 72 h, and cooled to room temperature in the furnace. Finally, to obtain densified bulk samples, the as-prepared ingots were crushed into fine powders and then vacuum sintered in a hot-pressing furnace at 823 K for 15 min under an axial pressure of 65 MPa.

Phase identification and structure analysis were carried out by x-ray powder diffraction (XRD) using a Rigaku D/max 2200 diffractometer equipped with Cu-Kα radiation (45 KV, 250 mA). The microstructures of the polished hot-pressed samples were investigated by scanning electron microscope (Zeiss Supra 55, Germany). An electron energy x-ray dispersive spectroscopy (EDXS, Oxford) was equipped for elemental mapping and point analysis. The density D of hot-pressed samples was determined by the Archimedes’ method and found to be over 97% of the theoretical density. The heat capacity C_P was obtained using Dulong–Petit model, C_P = 3nR/M, where n is the number of atoms per formula unit, R is the gas constant, and M is the molar mass. Thermal diffusivity (λ) was measured by the laser flash method using a Netzsch LFA457 instrument. The thermal conductivity was calculated from the equation κ = C_PDλ. The resistivity ρ and Seebeck coefficient S were measured via four probe method utilizing ZEM-3 instrument (ULVAC, Japan) under low-pressure helium atmosphere. The room-temperature carrier concentrations of the samples were obtained from Hall coefficient measurements, which were conducted on a mini cryogen free measurement system (Cryogenic Limited, UK).

Figure 1(a) shows the XRD patterns of CuIn_{1 − x}Mn_xTe₂ samples (x = 0, 0.025, 0.05, 0.075, 0.1). For all these samples, their diffraction peaks can be solely indexed to the chalcopyrite phase (PDF# 81-1937) and no impurities are observed. All the XRD patterns were analyzed by Rietveld refinement. The obtained lattice parameters of the a-axis are shown in Fig. 1(b). It can be seen that the lattice parameter, a, linearly decreases with the Mn contents when x < 0.075, indicating the replacement of In³⁺ (0.80 Å) by smaller Mn²⁺ (0.67 Å). Higher Mn content did not lead to further decrease in a, ascribed to the solid solubility limit of Mn.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) (a) The XRD patterns and (b) lattice parameters of the a-axis for CuIn1 − xMnxTe2 (x = 0, 0.025, 0.05, 0.075, 0.1) samples. The gray line in panel (b) is guide for eyes.

Figure 2 presents the backscattered electron (BSE) microscopy of the polished surface for typical samples with x = 0.05 and 0.075. As shown in Fig. 2(a), the BSE image of the sample with 5% Mn reveals a typical homogeneous solid solution microstructure and no obvious second phase precipitates can be found. However, for the sample with 7.5% Mn (see Figs. 2(b) and 2(c)), particle-like precipitates with sizes from several to 200 nm can be clearly observed, which are randomly distributed in the CuInTe₂ matrices. Thus, we can conclude that the solubility of Mn is between 5% and 7.5%. The precipitates cannot be identified in the XRD patterns, possibly due to their low content and tiny size. The elemental mapping (Fig. 2(d)) reveals that the precipitates have Mn-rich compositions.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) BSE images of the samples with (a) 5% Mn and (b) 7.5% Mn. The magnified image of 7.5% Mn sample and Mn elemental mapping of boxed area are shown in panels (c) and (d), respectively.

The electrical transport properties for CuIn_{1 − x}Mn_xTe₂ samples (x = 0, 0.025, 0.05, 0.075, 0.1) are compared in Fig. 3. As shown in Fig. 3(a), the electrical conductivity of the pristine CuInTe₂ sample increases with the rising temperature, exhibiting the typical behavior of intrinsic semiconductors. However, when Mn is doped, the room-temperature electrical conductivities of CuIn_{1 − x}Mn_xTe₂ are dramatically increased, and the temperature dependence of conductivities becomes the behavior of degenerate semiconductors, indicating that the carrier concentrations of the samples were substantially enlarged by Mn doping. The measured room-temperature hole concentrations (n_h) and Hall mobilities of carriers (μ_H) for the CuIn_{1 − x}Mn_xTe₂ samples are listed in Table 1. It can be seen that the hole concentrations of the samples increase with the increase of Mn contents. It has been reported that the hole concentration of the Mn-free CuInTe₂ is about 5.75 × 10¹⁸ cm⁻³.^[12] Consequently, the room-temperature electrical conductivity of the sample with x = 0.1 is ~ 46600 S/m, which is more than 40 times larger than that of the sample free of Mn (about 1100 S/m). Furthermore, the increasing of the hole concentrations suppresses the intrinsic excitation of charge carriers. Thus, the occurrence of bipolar conduction is pushed to higher temperatures with increasing Mn contents.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) The temperature dependence of (a) electrical conductivity, (b) Seebeck coefficient, (c) power factor for CuIn1 − xMnxTe2 (x = 0, 0.025, 0.05, 0.075, 0.1) samples.

Table 1.

Table 1.

Table 1. Room temperature hole concentrations (nh) and Hall mobilities of carriers (μH) for the CuIn1 − xMnxTe2 (x = 0.025, 0.05, 0.075, 0.1) samples. . x 0.025 0.05 0.075 0.1 nh/cm−3 7.4 × 1018 6.0 × 1019 8.3 × 1019 1.8 × 1020 μH/cm2⋅V−1⋅s−1 75 38 30 33

Table 1. Room temperature hole concentrations (nh) and Hall mobilities of carriers (μH) for the CuIn1 − xMnxTe2 (x = 0.025, 0.05, 0.075, 0.1) samples. .

Figure 3(b) shows the temperature-dependent Seebeck coefficient of CuIn_{1 − x}Mn_xTe₂ samples. Positive Seebeck coefficients reveal the p-type hole dominant conduction in all the samples. The variation trend of the Seebeck coefficient with temperatures is opposite to that of the electrical conductivity. The Seebeck coefficient of the pristine CuInTe₂ sample decreases with the temperature increasing, further proving that the undoped sample is a non-degenerate semiconductor, whereas the Seebeck coefficient for the samples doped with Mn increases with the temperature increasing, exhibiting a behavior of degenerate semiconductors. As the Mn content increases, the Seebeck coefficients at 340 K reduce from 424 μV/K for the pristine sample to 130 μV/K for the sample with x = 7.5%. The Seebeck coefficients for the samples with x = 0.075 and 0.1 were found to be almost identical at lower temperature range. One possible reason is that the solid solubility of Mn in CuInTe₂ is below 0.075. The increased electrical conductivity and moderately decreased Seebeck coefficients lead to enhanced power factors (PF) in the temperature rangingfrom room temperature to 720 K (Fig. 3(c)). A maximum PF of 11.2 μV⋅cm⁻¹⋅K⁻² at 625 K is achieved for the sample with x = 0.05, which is much larger than that for CuInTe₂ at 625 K.

Figure 4 depicts the temperature-dependent thermal transport properties of the samples. As shown in Fig. 4(a), the introduction of Mn can substantially reduce the total thermal conductivities (κ_tot). Lattice thermal conductivity (κ_L) can be obtained from κ_tot by subtracting the electronic part (κ_e) according to the relationship of κ_tot = κ_L + κ_e. Electronic thermal conductivity κ_e can be calculated from the electrical conductivity by the celebrated Wiedemann–Franz law κ_e = LσT, where L is the Lorenz number.^[17] The Lorenz number in degenerate semiconductors can be calculated by Seebeck coefficient within a single parabolic model, using the following equations:^[18]

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) The temperature dependence of (a) total thermal conductivities, (b) Lorentz numbers, (c) electronic thermal conductivities, and (d) lattice thermal conductivities for CuIn1 − xMnxTe2 (x = 0, 0.025, 0.05, 0.075, 0.1) samples. The dashed line in panel (d) is the curve for T−1 relationship.

The temperature-dependent dimensionless thermoelectric figure-of-merit ZT of CuIn_{1 − x}Mn_xTe₂ compounds is calculated based on the above transport data and is plotted in Fig. 5. The ZT values increase when approaching high temperature, and reach a maximum value of ~ 0.8 at 813 K for the sample with x = 0.075, which is slightly higher than that of the pristine CuInTe₂ sample. Nevertheless, the average ZT is also an important factor for actual thermoelectric applications.^[20] The calculated average ZT for all the samples is shown in the inset of Fig. 5. It can be seen that the average ZT of all the samples with Mn incorporation is much higher than that of the pristine one and is up to 0.34 for the samples with x = 0.075 and 0.1, owing to their moderate power factors and low lattice thermal conductivities in lower temperature range.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) Influence of Mn contents on figure of merit ZT for CuIn1 − xMnxTe2 samples. The inset is the average ZT calculated from the integration using the formula .

In summary, polycrystalline chalcopyrite CuIn_{1 − x}Mn_xTe₂ compounds (x = 0, 0.025, 0.5, 0.075, 0.1) has been successfully synthesized. The solubility of Mn in CuInTe₂ is determined to be below 7.5%. By substituting In, Mn acts as an effective acceptor dopant and can dramatically improve the electrical transport properties of the pristine CuInTe₂ system. In addition, κ_L of the CuInTe₂ system can be substantially decreased by the point defects and nanosized precipitates introduced by the incorporation of Mn, and thus resulting in a considerably reduction in κ_tot. As a result, an enhanced average ZT value up to 0.34 is achieved for sample CuIn_0.925Mn_0.075Te₂, which is 41% higher than that of the undoped CuInTe₂, making it a promising candidate for thermoelectric applications. Furthermore, it is reported that anions (P³⁻, Sb³⁻) doping on the Te sites can effectively increase the Seebeck coefficient and consequently enhance the PF of the system.^[4] Thus, introducing anion dopants plus some additional microstructure engineering in our Mn-doped CuInTe₂ system may bring its ZT value to that of the state-of-the-art thermoelectric materials.