1 Introduction

Thermoelectric materials have attracted extensive attention in the past decades due to their potential applications in direct thermal-to-electrical energy conversion without hazardous liquids, moving parts or greenhouse emissions [1,2,3]. The thermoelectric conversion efficiency is mainly determined by the dimensionless figure of merit ZT. ZT is defined as \(~ZT=\sigma {\alpha ^2}T/~\kappa ,\) where α is the Seebeck coefficient, σ is the electrical conductivity, T is the absolute temperature, and κ is the total thermal conductivity containing both the carrier contribution κe and phonon contribution κl. Increasing the power factor \(\left( {\sigma {\alpha ^2}} \right)\) by energy band engineering and reducing κ by introducing additional phonon scattering are two common methods to improve the conversion efficiency of thermoelectric [4,5,6,7,8,9,10].

Bismuth telluride and its alloys (BiSbTe) are one of the most important thermoelectric materials near room temperature as so far. In recent years, several methods and techniques can be adopted to raise the ZT values of bismuth telluride based alloys such as microwave assisted method [11], hot pressing [12], mechanical alloying [13], spark plasma sintering [14, 15], magnetron sputtering [16] or by doping them with certain materials which can tune its properties [17]. Nanocomposite has been considered as a promising way to improve ZT by synergistically tuning the electrical and the thermal properties at the same time. ZnAlO, graphene, WSe2 and silver have been reported to form effective nanoinclusions in Bi2Te3-based alloys, and improve their TE performance [8, 9, 18, 19]. For example, Zhao et al. reported that the addition of SiC nanopowders resulted in a remarkable decrease in thermal conductivity and improvement of mechanical properties of Bi2Te3-based alloys [20].

Lead telluride (PbTe) is one of the best materials used in thermoelectric generators operating at intermediate temperatures (450–800 K) [21]. This compound shows a larger band gap (0.3 eV) than bismuth telluride (0.13 eV) [22, 23], and low thermal conductivity. Better thermoelectric performance may be expected in bismuth telluride combined with PbTe. In this work, we prepare the BiSbTe alloys with additional PbTe using the zone melting (ZM) method. As we will see, the PbTe addition improves the figure of merit in the medium temperature range.

2 Experimental section

Elements of bismuth (Bi, 99.999%), antimony (Sb, 99.999%), tellurium (Te, 99.999%), and PbTe (prepared by a melting method) were used for the preparation of the samples without further purifying. PbTe was directly prepared by melting Pb and Te. Then Bi0.48Sb1.52Te3 + x wt% PbTe samples (x = 0, 0.05, 0.1 and 0.15) were prepared by ZM. The starting materials were weighted and sealed into evacuated quartz tubes, and then heated at 1173 K for 30 min in a rocking furnace to ensure homogeneity. After naturally cooling to room temperature, the alloys were grown via the ZM method at temperatures of 993 K with a growth speed of 25 mm h−1. Bars of 2 mm × 2 mm × 11 mm and discs of Ф 10 mm × 1.5 mm were cut from the samples along their growth direction to measure their electrical and thermal transport properties, respectively.

Phase structure of the samples was characterized by X-ray diffraction (XRD, Bruker AXS) using Cu Kα radiation (λ = 1.5406 Å). Microstructure was characterized by the field emission scanning electron microscopy (FESEM, Hitachi S4800). Electrical conductivity and Seebeck coefficient were measured by using a four-point probe method (ZEM-3). Thermal conductivity was calculated from the specific heat Cp, the thermal diffusivity λ, and the density ρ, using the equation \(~\kappa =~{C_p}\lambda \rho .\) Thermal diffusivity was measured by a laser flash method (NETZSCH LFA-457), and specific heat was a differential scanning calorimeter (Shimdzu DSC-50, Japan). The Hall coefficient RH was measured by a physical property measurement system (Quantum Design, PPMS-9). Carrier concentration n and mobility \(\mu\) were calculated by the relations \(~n=1/e{R_H}\) and \(\mu =\sigma {R_H}.\)

3 Results and discussion

The powder XRD patterns of Bi0.48Sb1.52Te3 + x wt% PbTe composites (x = 0, 0.05, 0.1 and 0.15) are shown in Fig. 1a. All patterns are indexed to the BiSbTe structure with R-3m space group (JCPDS Card No.65-3674). No secondary peak of PbTe phase was observed in all compounds. This may be due to the tiny amount of PbTe. We use the Rietveld method to refine the lattice parameters of these Bi0.48Sb1.52Te3 + x wt% PbTe samples. As shown in Fig. 1b, with the increase of PbTe content, both a and c decrease monotonously, indicating that PbTe atoms may enter the sites of the BiSbTe lattice [24].

Fig. 1
figure 1

a XRD patterns and b lattice parameters of Bi0.48Sb1.52Te3 + x wt% PbTe samples (x = 0, 0.05, 0.1 and 0.15)

Full size image

Figure 2 displays the FESEM image of the fractured surfaces perpendicular to the ZM direction for BiSbTe + 0.05 wt% PbTe. The lamellar structure on the micron scale can be clearly observed. The grains show a size of several tens of micrometer and are preferentially oriented. It is well-known that the ZM method can fabricate polycrystalline samples with good textured degree.

Fig. 2
figure 2

SEM image of the fractured surfaces perpendicular to ZM direction for Bi0.48Sb1.52Te3 + 0.05 wt% PbTe

Full size image

Figure 3 presents the temperature dependence of the (a) electrical conductivity and (b) Seebeck coefficients of the Bi0.48Sb1.52Te3 + x wt% PbTe samples (x = 0, 0.05, 0.1 and 0.15). In Fig. 3a, the electrical conductivity σ decreases with increasing the temperature from 300 to 500 K for all the samples, showing a typical metallic behavior [25,26,27]. The conductivity σ increases monotonously with x and reaches its maximum value at x = 0.15. Especially at 300 K, the σ value significantly enhances from 1.29 × 103 Scm−1 for x = 0 to 2.09 × 103 Scm−1 for x = 0.1. These values are distinctly higher than 5.0 × 102 Scm−1 of PbTe–BiSbTe samples prepared by the solvothermal method [28]. The high value of electrical conductivity in the BiSbTe–PbTe composites should be attributed to the presence of Pb, which can introduce more hole carriers. To clarify the behavior of σ, we evaluated n and µ at room temperature. As shown in Table 1, the room temperature carrier concentration n is in the range of 3.37–5.56 × 1019 cm−3. Clearly, it is observed that PbTe addition increases the carrier concentration. On the other hand, µ decreases moderately with increasing PbTe, from 238 cm2 V−1 s−1 for x = 0 to 191 cm2 V−1 s−1 for x = 0.1, which is consistent with previous reports. It can be attributed to the enhancement of defect scattering [29,30,31].

Fig. 3
figure 3

Temperature dependence of the a electrical conductivity and b Seebeck coefficient for the Bi0.48Sb1.52Te3 + x wt% PbTe samples

Full size image
Table 1 Hall coefficient RH, carrier concentration n, mobility µ, Seebeck coefficient α and effective mass m* of all the samples at room temperature
Full size table

Figure 3b gives the Seebeck coefficient α of all the samples as a function of temperature. The positive values of the Seebeck coefficients indicate that all the samples are p-type conductive. The Seebeck coefficients increase with increasing the temperature initially and then decreases, which is consistent with the previous reports [18, 19, 31]. The Goldsmid–Sharp band gap \({E_g}=2e{\alpha _{\hbox{max} }}T\left( {{\alpha _{\hbox{max} }}} \right)\) can be estimated from the maximum value of the Seebeck coefficient \({\alpha _{\hbox{max} }}\) and the corresponding temperature \(T\left( {{\alpha _{\hbox{max} }}} \right).\) The estimated values are about 0.17 eV for all samples which is similar to the literature report [25, 32] indicating that the introduction of PbTe has little impact on the band gap for the BiSbTe matrix. In contrast to the electrical conductivity σ, with the increasing PbTe content, the value of the Seebeck coefficients decreases. The highest Seebeck coefficient of 217 µV K−1 is obtained at 405 K in the BiSbTe matrix. This value is similar to the results of the commercial zone melted BiSbTe [12, 19, 31], but smaller than those prepared from the chemical or spark plasma sintering methods [8, 28]. The peak of the Seebeck coefficient shifts to a higher temperature, which may be due to the suppression of the intrinsic excitation induced by the introduction of PbTe. The Seebeck coefficient can be calculated by

$$\alpha =\frac{{8{\pi ^2}{k_B}^{2}}}{{3e{h^2}}}{m^*}T{\left( {\frac{\pi }{{3n}}} \right)^{{\raise0.7ex\hbox{$2$} \!\mathord{\left/ {\vphantom {2 3}}\right.\kern-0pt}\!\lower0.7ex\hbox{$3$}}}},$$

where kB, h, m*, and n are Boltzmann constant, Planck constant, effective mass of carrier, and carrier concentration, respectively [8]. The effective mass of each sample is estimated and listed in Table 1. Such results are similar to those reported in previous work [33,34,35]. It can be seen that PbTe addition enhances the effective mass of carrier in the composites, compared to that in BiSbTe matrix.

Based on the measured σ and α results, the temperature dependence of the power factors \(\left( {{\text{PF}}=\sigma {\alpha ^2}} \right)\) is summarized in Fig. 4. With the rising temperature, power factors for all the samples exhibit decrement. The maximum power factors of BiSbTe matrix is about 49 µW cm−1 K−2. The power factors of all samples with PbTe incorporation are higher than the BiSbTe matrix and surpass 50 µW cm−1 K−2 at room temperature, which is attributed to the greatly improved electrical conductivity. When the addition is 0.05 wt%, the power factor reaches a maximum of 55 µW cm−1 K−2 at room temperature, which is 12% higher than that of BiSbTe. But with further increasing the content of PbTe, the power factor begins to decrease, due to the decrease of Seebeck coefficients (see Fig. 3b).

Fig. 4
figure 4

Temperature dependence of the power factor for the Bi0.48Sb1.52Te3 + x wt% PbTe samples

Full size image

As shown in Fig. 5a, the temperature dependence of thermal conductivity for all samples first reduces due to the increasing phonon–phonon scattering, and then increases rapidly when further increasing testing temperature, which can be attributed to the bipolar thermal conductivity. The increasing n due to intrinsic excitation leads to a significant increase of the bipolar thermal conductivity. The minimum total thermal conductivity of BiSbTe matrix is about 1.54 W m−1 K−1. The total thermal conductivity κ increases monotonously with increasing the content of PbTe over the entire temperature range. It is partly ascribed to the increase in the electronic thermal conductivity \(~{\kappa _e}.\) The electronic thermal conductivity can be calculated from the formula \(~{\kappa _e}=L\sigma T,\) where L is the Lorenz number. Here the Lorenz number L is roughly obtained by fitting the S to the reduced chemical potentials, which results in an L with a deviation of less than 10% as compared with a more rigorous single nonparabolic band and multiple band model calculation. The lattice thermal conductivity is then calculated by subtracting the electronic thermal conductivity from the total thermal conductivity, as shown in Fig. 5b. The reduction of lattice thermal conductivity can be observed clearly in the composites, indicating strong point defect phonon scattering by PbTe addition. When the content is 0.15 wt% the minimum lattice thermal conductivity is 0.83 W m−1 K−1 at 300 K.

Fig. 5
figure 5

Temperature dependence of the a total thermal conductivity, b lattice thermal conductivity, c figures of merit and d average ZT for the Bi0.48Sb1.52Te3 + x wt% PbTe samples

Full size image

The dimensionless thermoelectric figures of merit ZT of all bulk samples are calculated according to \(ZT=\sigma {\alpha ^2}T/~\kappa ~.~\) Figure 5c shows the temperature dependent ZT values of the BiSbTe + x wt% PbTe samples. It may seem that ZT for all samples increases with increasing temperature, reaches a maximum, and then decreases with further increasing temperature. The highest ZT value is about 1.0 at 350 K for the sample with x = 0.05. Comparing with the BiSbTe matrix, the composites have comparable ZT values before 375 K. Above this temperature, the ZT values of BiSbTe + x wt% PbTe composites are higher than the BiSbTe matrix, which are attributed to the improvement of power factors.

As known, the thermoelectric applications not only require a high peak ZT (ZTmax) but also require high ZTs in a wide temperature range. Thus, improving the average ZT (ZTave) is very beneficial for thermoelectric device applications. We calculated the average ZTs in the temperature range of 300–500 K and presented them in Fig. 5d. The PbTe-incorporated compounds show higher average ZT values, and the maximum ZTave is about 0.81 for x = 0.05.

4 Conclusions

Bi0.48Sb1.52Te3 + x wt% PbTe (x = 0, 0.05, 0.1, and 0.15) composites were prepared by a ZM method. The results demonstrate that PbTe addition increases the electrical conductivity, and consequently the largest power factor is increased to 55.5 µW cm−1 K−2. The improved electrical performance is mainly due to the increased carrier concentration. Compared with the Bi0.48Sb1.52Te3 sample, the PbTe-incorporated samples have slightly higher peak ZT values, and a substantial enhancement of the average ZT values. The optimization of thermoelectric performance for Bi0.48Sb1.52Te3 has been achieved by PbTe incorporating.