Effect of doping indium into a Bi₂Te₃ matrix on the microstructure and thermoelectric transport properties

Abstract

Indium (In) as an unconventional doping element for Bi₂Te₃ is applied in this work to investigate the effect of In doping on the microstructure and thermoelectric transport properties of Bi₂Te₃-based alloys prepared via a unique high-pressure method (high pressure and high temperature, HPHT). The synthesis time of the sample is acutely shortened compared with conventional preparation methods. Doping In into a Bi₂Te₃ matrix can induce multiple textures and microstructures, effectively scattering different frequency phonons. Additionally, the electrical transport properties are also significantly modulated due to the role of In atoms as electron donors. As a result, the lattice thermal conductivity and electrical resistivity are dramatically reduced, in particular a decrease of 64–54% in 305–585 K of the lattice thermal conductivity of the samples with In content from 1 to 2 at%. An acceptable ZT value of 0.65 at 385 K is achieved from the as-prepared Bi_1.9In_0.1Te₃ bulk materials with In as a dopant in the Bi₂Te₃ matrix.

---

1. Introduction

With an increasingly severe energy crisis and environmental pollution, the development of human society urgently needs a green energy conversion material to prevent the situation continuing to deteriorate. The discovery of thermoelectric materials attracts considerable attention as an important energy conversion material, which can directly convert waste heat into usable electricity and vice versa without any pollution. This excellent characteristic can be extensively used in the fields of power generation and refrigeration. However, the large-scale applications are restricted for now due to the low efficiency of thermoelectric conversion, which is evaluated by the dimensionless figure of merit (ZT): ZT = S²T/κρ, where S, ρ, κ and T are the Seebeck coefficient, electrical resistivity, total thermal conductivity (including the lattice and electron thermal conductivity) and kelvin temperature, respectively.^1,2

In the family of thermoelectric materials, several classes of thermoelectric materials are being investigated for renewable power generation applications including lead/germanium tellurides,^3,4 half-Heuslers,^5,6 silicides⁷ and bismuth telluride, of which bismuth telluride (Bi₂Te₃) based alloys are regarded as the best potential thermoelectric materials since they came under observation for refrigeration and low-temperature power generation.^8,9 Bi₂Te₃-based alloys not only have essentially good electrical properties, but also exhibit lower thermal conductivity. The improvement of ZT values for Bi₂Te₃-based alloys generally requires lower lattice thermal conductivity and the ideal electrical properties by microstructure design and composition optimization. Long-term research suggests that Sb and Se elements respectively substitute for Bi and Te that can constitute the best p- and n-type Bi₂Te₃-based alloys.^10–12 However, the effect of other elements doping on the microstructures and thermoelectric properties of Bi₂Te₃ materials is less reported. In addition, with the development of preparative techniques, the hydrothermal synthesis, zone melting, spark plasma sintering, etc. exhibit respective advantages in terms of optimizing thermoelectric performances. Among the technological means, our previous work indicates that high pressure and high temperature (HPHT) method for materials synthesis has a series of unique advantages, which are rapid synthesis, low-cost, one-step, large-scale production, etc.^13,14 In addition, the HPHT method can induce reduction in resistivity and thermal conductivity, which are attributed to the effect of high pressure on crystal/band structure, carrier concentration and microstructure.^14–17 Therefore, the HPHT method is employed to conduct the preparation and investigation of doped In in a Bi₂Te₃ matrix in this work.

The In-doped Bi₂Te₃-based alloys with various atomic ratios are successfully synthesized via a HPHT method for the first time. Surprisingly, the processing time is sharply reduced from a few days to 30 min in comparison with other methods. The thermoelectric properties and microstructures of the In-doped Bi₂Te₃-based alloys are investigated in detail. The textures and microstructures are characterized, revealing an evident difference with In content. Meanwhile, the as-prepared (Bi,In)₂Te₃ alloys exhibit a substantial decreases of lattice thermal conductivity and electrical resistivity. Consequently, an acceptable ZT value of 0.65 is achieved at 385 K, which is comparable with the maximum ZT value of the state-of-the-art In-doped Bi₂Te₃-based alloys.¹⁸

2. Experimental section

2.1 Sample preparation

Precursor samples with the nominal composition of Bi_1.95In_0.05Te₃ (In-1 at%), Bi_1.93In_0.07Te₃ (In-1.4 at%), Bi_1.9In_0.1Te₃ (In-2 at%) and Bi_1.8In_0.2Te₃ (In-4 at%) were prepared to be a ϕ 10.5 × 11 mm cylinder by cold compression. All the chemical compositions employed the high-pure powders of Bi (5N), Te (5N) and In (4N). The precursor sample was assembled into a sample chamber of 23 mm with a graphite crucible as a heating unit, following by preparation at 900 K and under the pressure of 2 GPa for 30 min via a cubic anvil high-pressure apparatus (SPD 6 × 1200). The pressure was calibrated by the change in resistance of standard substances. The synthesis temperature was measured by a Pt RH/Pt Rh6 thermocouple placing near the sample.

2.2 Characterization

The phase structures of the as-prepared samples were identified by X-ray diffraction (XRD, D/MA X-RA) with Cu K_α radiation. The morphologies were observed by field-emission scanning electron microscopy (FESEM, Magellan-400 FEI microscope). An energy dispersive spectrometer (EDS) system in FESEM equipment was employed for measuring the atomic ratio of the composition. Microstructures were obtained by a high-resolution transmission electron microscope (HRTEM, JEOL JEM-2200FS). The electrical properties (S and ρ) were measured simultaneously using ZEM-3 apparatus (ULVAC-RIKO, Inc.). The thermal conductivity was calculated by the formula of κ = C_pλD, where C_p, λ and D are the specific heat, thermal diffusivity and relative density, respectively. The C_p measurement by DSC conversion was obtained using a Linseis STA PT-1750 equipment with the sapphire revision, the λ was measured based on a laser flash technique (Netzsch LFA-427), the D was obtained by the Archimedes method. To consider the anisotropy of Bi₂Te₃-based alloys, the parameters above were measured along the cross-plane. In addition, the errors in S, ρ and κ were retested and did not exceed ±3–5%.

3. Results and discussion

3.1 XRD patterns

Fig. 1 shows the XRD patterns of the as-prepared samples with In-1, 1.4, 2, 4 at% and the standard Bi₂Te₃ data card (PDF#15-863). It can be observed that the characteristic peaks of the as-prepared samples can be indexed to the standard diffraction peaks (PDF#15-863), indicating that the as-prepared samples have a rhombohedral lattice structure with the space group of R3̄m. The (00l) peaks, due to the effect of synthesis pressure, are increased strongly and preferentially, which is consistent with our previous work.^17,19 In addition, the peaks of In₂Te₃ compounds may be identified in Fig. 1 due to the In doping, which means that the In atoms have entered into the matrix and substituted for the local lattice sites of Bi atoms in the crystals. The results above suggest that (Bi,In)₂Te₃ with varying In content has been successfully synthesized in 30 min by HPHT.

Fig. 1 XRD patterns of the standard Bi₂Te₃ data card (PDF#15-863) and In-doped Bi₂Te₃-based alloys with In-1, 1.4, 2, 4 at% prepared by the HPHT method.

3.2 FESEM and HRTEM micrographs

The FESEM images in Fig. 2 exhibit the universal morphologies for In-1, 1.4, 2 and 4 at%. The as-prepared samples have the typical lamellar structures of Bi₂Te₃-based alloys. However, with the doped In from 1 to 4 at% in the matrix, it can be observed that the lamellar textures are blocked, thereby being randomly divided into small plates, just like the crushed crystal planes. This means that the In atoms with weak electronegativity compared to Bi atoms gradually substitute for local Bi sublattices in crystals, affecting the Te¹–Te¹ bonding (van der Waals bonding) and interatomic binding, and then forming an inhomogeneous morphology. As is well known, the textures, which have abundant boundaries and plate distributions, are generally regarded as a positive effect for decreasing the lattice thermal conductivity. In addition, the random regions in Fig. 2 for different samples are chosen, and conduct EDS measurement in order to verify the atomic ratios of compositions and the segregation of In. The results indicate that the In elementary substance is inexistence, and the atomic ratios for each of the chemical elements are approximate to the nominal composition (Table S1†). Moreover, the A and B regions in Fig. 2(c) are also measured via EDS. The results suggest that the In atomic ratios are between 2 at% and 4 at% (Fig. S1†), which may be the transition states for the lamellar turning into inhomogeneous textures. As mentioned above, the traces of In can enter the Bi₂Te₃ matrix to substitute for local Bi atoms and observably transform the texture morphology, in addition to indicating that the solubility limit of In is higher than 4 at% in Bi₂Te₃ matrix.

Fig. 2 FESEM micrographs of as-prepared samples with In-1 at% (a), 1.4 at% (b), 2 at% (c) and 4 at% (d). The inserts exhibit the EDS patterns of the blue rectangular regions, respectively.

The HRTEM image of the as-prepared sample with In-2 at% exhibits periodic lattice fringes in Fig. 3(a). The adjacent fringes are measured to show a distance of 0.99 nm, which can be indexed to the typical quintuple layered structure of Bi₂Te₃-based samples,²⁰ further confirming the synthesis of the expected samples by HPHT. In Fig. 3(b), it can be seen that copious lattice orientations present a disorderly growth pattern, indicating that the as-prepared samples are polycrystalline, which is beneficial to the phonon scattering. Besides, the lattice mismatch originating from pressure and In doping induces abundant lattice defects including dislocations and lattice distortions (Fig. 3(c)). In order to clearly observe the defects, the inverse fast Fourier transform (IFFT) image corresponding to Fig. 3(c) is applied, which can readily identify abundant dislocations and lattice distortions in Fig. 3(d). The phonon mean free path will be impacted by abundant lattice defects to be decreased, effectively scattering the phonons to reduce the thermal conduction.

Fig. 3 HRTEM micrographs of as-prepared Bi_1.9In_0.1Te₃ (In-2 at%). (a) Representative quintuple layered structure of Bi₂Te₃-based sample. (b) Lattice orientations. (c) Lattice defects: abundant lattice distortions and dislocations. (d) IFFT image corresponding to (c).

3.3 Thermoelectric transport properties

The temperature dependence of the Seebeck coefficient (S) of the as-prepared (Bi,In)₂Te₃ with In-1, 1.4, 2 and 4 at% has been shown in Fig. 4(a). In the measured temperature range, the S exhibits a negative value, indicating a typical n-type semiconductor behavior. The absolute value of S with temperature presents an enhancement and then a reduction, and reaches a maximum absolute value of 135.9 μV K⁻¹ at 385 K from the as-prepared sample with In-2 at%. The absolute S value of the sample with In-4 at% doping exhibits a different behavior comparing with that of In-1, 1.4 and 2 at% doping, which may be ascribed to the increased carriers. As in the report of Drasar et al., an In-4 at% content is heavy doping for Bi₂Te₃-based alloys,¹⁸ and thus the impurity ionization readily excites vast carriers that reveal stronger effects on S than carrier scattering, leading to a reduction in the absolute S of In-4 at% doped samples. The electrical resistivity (ρ) of the as-prepared samples is plotted in Fig. 4(b), and increases monotonously with rising temperature, implying a metal-like conduction or a degenerate semiconductor behavior.^21–23 The minimum value of ρ is obtained to be 7.47 × 10⁻⁶ Ω m at 305 K, and then the thermal excitation with elevated temperatures gives rise to lattice vibrations, restraining the transport of low-energy electrons to achieve a similar behavior of metallic conduction.²⁴ In addition, the ρ almost reveals a reduction as the enhancement of In content, this situation is mainly dependent on In atoms serving as electron donors substitute for the Bi sublattices in lattices, improving the metallicity of as-prepared samples.

Fig. 4 Temperature dependence of the Seebeck coefficient (a), electrical resistivity (b), total thermal conductivity (c) and lattice thermal conductivity (d) of as-prepared samples with In-1, 1.4, 2 and 4 at%.

The total thermal conductivity (κ) of the as-prepared (Bi,In)₂Te₃ with varying In content as a function of temperature is displayed in Fig. 4(c). The κ exhibits a parabola-like relationship with temperature. In addition, we obviously observe that the κ realizes a dramatically monotonous decrease with doped In content from 1 at% to 4 at%. To further understand the mechanisms, the lattice thermal conductivity is calculated by subtracting the electron thermal conductivity (κ_e) from the total thermal conductivity (κ), where κ_e can be estimated by the Wiedemann–Franz law (κ_e = LT/ρ) using the Lorentz number (L) of 2.0 × 10⁻⁸ V² K⁻² for degenerate semiconductors.²⁵ The κ − κ_e is thus obtained and plotted in Fig. 4(d). Surprisingly, note that the κ − κ_e exhibits a decrease of 64–54% at 305–585 K with increasing In content. The huge reduction in lattice thermal conductivity is derived from the effects of In-induced multiple textures, lattice distortions and dislocations on the phonon transport, which noticeably induce the scattering of mid- and long-wavelength phonons across a broad wavelength spectrum to decrease the κ − κ_e. In short, the multi-scale hierarchical textures and lattice defects produced by In doping effectively restrain different frequency phonons’ transport, decreasing the lattice thermal conductivity contribution to κ.

The temperature dependence of the figure of merit (ZT) values of the as-prepared (Bi,In)₂Te₃ with In-1, 1.4, 2 and 4 at% is calculated using the measured electrical and thermal properties, and is plotted in Fig. 5. The ZT values exhibit an increase and then a decrease with temperature rise. It is well known that S, ρ and κ_e are associated with the carriers, which will be influenced by two stages of lattice vibration and intrinsic excitation with temperature, impacting on thermoelectric performance to reveal a parabola-like behavior. This situation can be observed generally in all kinds of thermoelectric materials with appropriate temperature ranges. The maximum ZT of 0.65 is achieved at 385 K from the as-prepared sample with In-2 at%. The acceptable ZT value is comparable with that of the state-of-the-art In-doped Bi₂Te₃-based alloys,¹⁸ which is ascribed to the reductions in electrical resistivity and lattice thermal conductivity. Based on the positive effects of In doping, we believe that the ZT value of the as-prepared samples via the HPHT method is hopeful to be further improved through incorporating with the optimization of chemical composition, ultrahigh pressure, multi-elements doping, etc.

Fig. 5 Temperature dependence of the figure of merit (ZT) values of as-prepared samples with In-1, 1.4, 2 and 4 at%.

4. Conclusions

In summary, In-doped Bi₂Te₃-based alloys with In-1, 1.4, 2 and 4 at% are successfully synthesized in 30 min by a HPHT method for the first time. The traces of In in the Bi₂Te₃ matrix noticeably transform the texture morphology to form multi-scale hierarchical textures. In addition, abundant lattice defects are formed in the microstructure due to lattice mismatch resulting from In/Bi substitution. The structural features above are beneficial to phonon scattering, thus inducing a decrease of 64–54% at 305–585 K of the lattice thermal conductivity. It’s worth mentioning that the reduction of lattice thermal conductivity is not at the cost of the increase of resistivity due to doped In as electron donors and high-pressure inhibition. Ultimately, the acceptable maximum ZT value of 0.65 is obtained at 385 K from as-prepared Bi_1.9In_0.1Te₃ (In-2 at%) via the HPHT method.

Acknowledgements

This work was financially supported by the National Science Foundation of China (51171070, 51071074, 51301024 and 11504319) and the Open Project (Key Laboratory of Functional Materials Physics and Chemistry (Jilin Normal University), Ministry of Education, China (No. 201610)).

Notes and references

1. R. Venkatasubramanian, E. Siivola, T. Colpitts and B. O’Quinn, Nature, 2001, 413, 597.
2. B. Poudel, Q. Hao, Y. Ma, Y. C. Lan, A. Minnich, B. Yu, X. Yan, D. Z. Wang, A. Muto, D. Vashaee, X. Y. Chen, J. M. Liu, M. S. Dresselhaus, G. Chen and Z. F. Ren, Science, 2008, 320, 634.
3. Y. Gelbstein, Z. Dashevsky and M. P. Dariel, Phys. B, 2007, 396, 16.
4. J. Davidow and Y. Gelbstein, J. Electron. Mater., 2013, 42, 1542.
5. K. Kirievsky, Y. Gelbstein and D. Fuks, J. Solid State Chem., 2013, 203, 247.
6. Y. Gelbstein, N. Tal, A. Yarmek, Y. Rosenberg and M. P. Dariel, J. Mater. Res., 2011, 26, 1919.
7. Y. Sadia, L. Dinnerman and Y. Gelbstein, J. Electron. Mater., 2013, 42, 1926.
8. Q. H. Zhang, X. Ai, L. J. Wang, Y. X. Chang, W. Luo, W. Jiang and L. D. Chen, Adv. Funct. Mater., 2015, 25, 966.
9. B. Qiu and X. L. Ruan, Appl. Phys. Lett., 2010, 97, 183107.
10. C. J. Liu, H. C. Lai, Y. L. Liu and L. R. Chen, J. Mater. Chem., 2012, 22, 4825.
11. X. Yan, B. Poudel, Y. Ma, W. S. Liu, G. Joshi, H. Wang, Y. C. Lan, D. Z. Wang, G. Chen and Z. F. Ren, Nano Lett., 2010, 10, 3373.
12. D. Li, X. Y. Qin, J. Zhang, C. J. Song, Y. F. Liu, L. Wang, H. X. Xin and Z. M. Wang, RSC Adv., 2015, 5, 43717.
13. B. Sun, X. P. Jia, D. X. Huo, X. Guo, H. R. Sun, Y. W. Zhang, B. W. Liu and H. A. Ma, Mod. Phys. Lett. B, 2015, 29, 1550214.
14. T. C. Su, C. Y. He, H. T. Li, X. Guo, S. S. Li, H. A. Ma and X. P. Jia, J. Electron. Mater., 2013, 42, 109.
15. P. W. Zhu, X. Jia, H. Y. Chen, W. L. Guo, L. X. Chen, D. M. Li and G. T. Zou, Solid State Commun., 2002, 123, 43.
16. Y. Dong, A. S. Malik and F. J. DiSalvo, J. Solid State Chem., 2010, 183, 1817.
17. X. Guo, X. P. Jia, T. C. Su, K. K. Jie, H. R. Sun and H. A. Ma, Chem. Phys. Lett., 2012, 550, 170.
18. C. Drasar, A. Hovorkova, P. Lostak, H. Kong, C. P. Li and C. Uher, J. Appl. Phys., 2008, 104, 023701.
19. X. Guo, X. P. Jia, K. K. Jie, H. R. Sun, Y. W. Zhang, B. Sun and H. A. Ma, CrystEngComm, 2013, 15, 7236.
20. Z. Aabdin, N. Peranio, O. Eibl, W. Tollner, K. Nielsch, D. Bessas, R. P. Hermann, M. Winkler, J. Konig, H. Bottner, V. Pacheco, J. Schmidt, A. Hashibon and C. Elsasser, J. Electron. Mater., 2012, 41, 1792.
21. W. J. Xie, X. F. Tang, Y. G. Yan, Q. J. Zhang and T. M. Tritt, J. Appl. Phys., 2009, 105, 113713.
22. D. Li, X. Y. Qin, Y. C. Dou, X. Y. Li, R. R. Sun, Q. Q. Wang, L. L. Li, H. X. Xin, N. Wang, N. N. Wang, C. J. Song, Y. F. Liu and J. Zhang, Scr. Mater., 2012, 67, 161.
23. T. J. Zhu, H. L. Gao, Y. Chen and X. B. Zhao, J. Mater. Chem. A, 2014, 2, 3251.
24. G. Y. Xu, S. T. Niu and X. F. Wu, J. Appl. Phys., 2012, 112, 073708.
25. T. J. Zhu, Z. J. Xu, J. He, J. J. Shen, S. Zhu, L. P. Hu, T. M. Tritt and X. B. Zhao, J. Mater. Chem. A, 2013, 1, 11589.

---

Footnote

† Electronic supplementary information (ESI) available: Additional EDS patterns and table with elemental composition data. See DOI: 10.1039/c6ra09767g

---