Enhanced thermoelectric performance in GeTe-Sb₂Te₃ pseudo-binary via lattice symmetry regulation and microstructure stabilization

Highlights

• In the aspect of thermoelectric performance, a maximal ZT_max ∼ 2.4 at 773 K was realized in the sample (Sb₂Te₃)_0.5(Ge_0.91Pb_0.09Te)_17.5 (annealed for 4 days), together with an average ZT_avg ∼ 1.5 at the temperature difference 323–773 K.

Abstract

Pseudo-binary GeTe-rich Sb₂Te₃(GeTe)_n materials recently exhibited promising thermoelectric performance at intermediate temperatures (500–800 K), largely due to the intrinsically low lattice thermal conductivity coming from the discrete van der Waals gaps dispersed in a rhombohedral matrix. In this work, by alloying Ge with Pb and adjusting the molar ratio of GeTe/Sb₂Te₃ in the binary, we successfully modulated the crystal structure from rhombohedral Sb₂Te₃(GeTe)₁₇ to pseudo-cubic (Sb₂Te₃)_0.5(Ge_0.91Pb_0.09Te)_17.5 at room temperature, thus achieved higher electronic band degeneracy and electrical performance. High-resolution scanning transmission electron microscope (STEM) characterizations revealed the existence of high-density discrete van der Waals gaps (length ∼ 10–40 nm) along {111} equivalent planes in GeTe matrix; surprisingly, these planar defects appear quite stable in following annealing processes at 873 K unlike what literatures reported. Further elemental mapping suggests that the enrichment of Pb element around van der Waals gaps are possibly responsible to the formation and stabilization of these planar defects. Eventually, a figure of merit ZT_max ∼2.4 at 773 K and average ZT_avg ∼1.5 at 323–773 K were simultaneously realized in the (Sb₂Te₃)_0.5(Ge_0.91Pb_0.09Te)_17.5 sample after 4 days annealing at 873 K.

Introduction

In past decades, many efforts have been done to exploring high performance thermoelectric materials, which can directly and reversibly convert heat to electricity [1,2]. The performance of a thermoelectric material is evaluated by a dimensionless figure of merit ZT = S²σT/(k_lat + k_ele), where S, σ, k_lat, k_ele, and T are the Seebeck coefficient, electrical conductivity, lattice thermal conductivity, electronic thermal conductivity, and absolute temperature, respectively [3,4]. To obtain a high ZT, a large power factor (PF=S²σ) and a low lattice thermal conductivity are necessary. The power factor can be improved by optimizing the carrier concentration [[5], [6], [7], [8]] and modulating the electronic bands [[9], [10], [11], [12], [13], [14]], whereas the lattice thermal conductivity can be effectively reduced via introducing point defects [15,16], nano-structure [[17], [18], [19], [20]] and/or hierarchical architectures [[21], [22], [23]].

The aforementioned methods were already successfully applied in many typical intermediate temperature IV–VI semiconductors, such as PbTe [5,19,[24], [25], [26], [27]], SnTe [17,[28], [29], [30]], SnSe, [[31], [32], [33]] and indeed induced large enhancement of their thermoelectric performance. GeTe, which is also a IV–VI narrow-gap semiconductor, experiences a ferroelectric phase transition at ∼700 K from low-temperature rhombohedral to high-temperature cubic phase through lattice elongation and Ge off-centering along the [111] direction. This material has recently been reported as a promising thermoelectric material [12,34,35] comparable to state-of-art PbTe. The high-temperature cubic GeTe (C-GeTe) are believed to have high band degeneracy thus superior electrical performance, while the low-temperature rhombohedral GeTe (R-GeTe) exhibits low symmetry and inferior electrical properties [4]. More recently, enhanced thermoelectric figure of merit ZTs have been realized in Ge_0.95Pb_0.01Sb_0.05Te, [36] Ge_0.76Sb_0.08Pb_0.12Te, [37] Ge_0.85Pb_0.1Bi_0.04Te [38], Ge_0.89Sb_0.1In_0.01Te, [39] Ge_0.90V_0.02Bi_0.08Te, [40] Ge_0.89Ti_0.03Sb_0.08Te, [41] Ge_0.9Mg_0.04Bi_0.06Te, [42] by refining the carrier concentration and reducing lattice thermal conductivity. In some Sb³⁺/Bi³⁺ doped GeTe samples, researchers reported extremely low lattice thermal conductivity and attributed it to the occasionally found “stacking faults” [43,44] which localized in the GeTe matrix. Most recently, Xu et al. [45] reported that these stacking faults actually the intrinsic microstructure when GeTe is alloyed with Sb₂Te₃, and shall be named van der Waals gaps; moreover, they found these van der Waals gaps feature could be further optimized by annealing processes, and eventually achieved a ZT_max ∼2.4 at 773 K in Sb₂Te₃(GeTe)₁₇ (GST17) after annealed 7 days. Almost at the same time, Wu et al. [46] demonstrated that Bi₂Te₃ alloying in GeTe lattice can introduce vast Ge neutral vacancies, which can further evolve into van der Waals gaps upon proper heat treatment; together with hierarchical ferroelectric domain structure, they reported a striking ZT_max ∼2.4 at 773 K in the composition ∼ Bi₂Te₃(GeTe)_26.6 (GBT26.6). Both GST17 and GBT26.6 are featured with extremely low lattice thermal conductivity ∼0.3 W/mK at proper temperatures, suggesting the efficacy of discrete van der Waals gaps on strengthening phonon scattering. Beyond the GeTe-based materials, many efforts have been done for realizing a high efficient thermoelectric application of GeTe, such as Ag/SnTe/(Ge_0.98Cu_0.04Te)_0.88(PbSe)_0.12 [47]; Ni/Ti/Ge_0.85Mg_0.05Sb_0.1Te [48]; Ge_0.92Sb_0.04Bi_0.04Te_0.95Se_0.05/Yb_0.3Co₄Sb₁₂ [49].

Despite of great advances achieved in GeTe-based materials, two problems still cannot be ignored. First, although the phase transition temperature of GeTe can be modulated via proper alloying/doping [39,50,51], considerable range of working temperatures still locates at the low-symmetry R-phase. Second, although striking figure of merit ZT can be achieved in GeTe-Sb₂Te₃ and GeTe-Bi₂Te₃ binaries, the best performance can only be realized by optimizing these van der Waals gap structure via proper and tedious heat treatments [45,46]. In this work, by partially substituting Ge with Pb and adjusting the molar ratio of GeTe/Sb₂Te₃, we successfully modulated the lattice symmetry at room temperature from rhombohedral Sb₂Te₃(GeTe)₁₇ (GST17) to pseudo-cubic (Sb₂Te₃)_0.5(Ge_0.91Pb_0.09Te)_17.5 (LGST17.5), with the peak ZT_max improved from 1.6 to 2.1 and average ZT_avg from 0.9 to 1.3, as shown in Fig. 1. Further annealing at 873 K for 4 days resulted in little change of these discrete van der Waals gap structure, and produced subtle enhancement of thermoelectric performance, i.e., ZT_max and ZT_avg slightly increase to 2.4 and 1.5 for the (Sb₂Te₃)_0.5(Ge_0.91Pb_0.09Te)_17.5 sample annealed for 4 days (LGST17.5-AN4). Our findings offer an alternative strategy for pursuing high and stable thermoelectric performance in GeTe-Sb₂Te₃ pseudo-binary systems.