The strength and thermoelectric properties of PbTe and Sn_0.9Pb_0.1Te medium-temperature polycrystalline specimens with p and n conductivity types, respectively, have been studied. The specimens have been produced using extrusion and spark plasma sintering. The strength parameters of the materials were studied using uniaxial compression at 20 to 500 °C. The structure of the materials was studied using X-ray diffraction and electron microscopy. The electrical conductivity and the Seebeck coefficient were measured simultaneously using the four-probe and differential methods. The temperature conductivity and the specific heat capacity were measured using the laser flash and differential scanning calorimetry methods.

The PbTe and Sn_0.9Pb_0.1Te materials produced using extrusion and spark plasma sintering prove to be single-phase and have homogeneous compositions. For comparable synthesis methods, the dislocation density in the Sn_0.9Pb_0.1Te specimens is by an order of magnitude lower than in the PbTe ones.

Study of the mechanical properties of n and p conductivity type specimens over a wide temperature range from 20 to 500 °C has shown that their deformation is plastic and has no traces of brittle fracture. For these plastic materials, the strength criterion has been accepted to be the arbitrary yield stress corresponding to the stress at a 0.2% deformation. The 20 °C yield stress of PbTe and Sn_0.9Pb_0.1Te is far higher for the specimens produced by extrusion. For all the test temperatures and synthesis methods the Sn_0.9Pb_0.1Te specimens have a higher strength than the PbTe ones.

The PbTe and Sn_0.9Pb_0.1Te specimens produced by extrusion have better thermoelectric properties than the spark plasma sintered ones. The heat conductivity of the PbTe and Sn_0.9Pb_0.1Te specimens is almost the same regardless of compaction method.

thermoelectric materials, lead telluride, tin telluride, dynamic compaction, heat conductivity, thermoelectric efficiency

The abundance of both manmade and natural exhaust heat sources offers a wide range of opportunities for thermoelectric heat conversion to electric power [1–4]. However, wide application of thermoelectric energy converters is currently limited by the insufficient efficiency of thermoelectric devices. The dependence of thermoelectric device efficiency on material properties is expressed through the dimensionless thermoelectric efficiency

$ZT=\frac{{\alpha }^2\sigma }{\kappa }$,

where T is the working temperature, Z is the thermoelectric efficiency, α is the Seebeck coefficient, σ is the electrical conductivity and κ is the heat conductivity.

Thermoelectric materials exhibit the maximum efficiency value ZT_max in a relatively narrow temperature range, whereas the actual operation temperature range of thermoelectric materials is far wider. Depending on the temperature T_max, the thermoelectric materials are divided into low-, medium- and high-temperature ones [5–7]. The low-temperature thermoelectric materials are those having the maximum thermoelectric efficiency at below 300 °C. The working range of the medium-temperature thermoelectric materials is 300 to 600 °C. The high-temperature thermoelectric materials are used at above 600 °C. No thermoelectric materials can provide sufficiently high ZT in the entire temperature range. Therefore for achieving a wide operation temperature range of high-efficiency thermoelectric generators, their design comprises segmented thermoelectric cells. The cell segments are made from different thermoelectric materials each of which has the highest efficiency in its specific temperature range [8–11]. Thermoelectric energy converters find demand in renewable sources (hybrid photo-thermoelectric solar cells, thermoelectric generators for low-grade heat recovery from geothermal sources, thermoelectric generators driven by temperature difference between water or ice layers at different depths of seas and oceans etc.). However, the most promising application domain of thermoelectric generators is currently the electric power conversion of waste heat from units of cars and other vehicles, as well as electric power plants and industrial installations. The most promising operation temperature range for thermoelectric generators, primarily, those recovering exhaust heat, is 300 to 600 °C [12]. The main thermoelectric material for this range is lead telluride (PbTe) and solid solutions on its basis [16–18].

Thermoelectric materials, especially those used in electric power generators, may undergo mechanical stresses produced at sufficiently large temperature differences between the hot and cold ends of the thermoelectric cell [19]. In that case, a significant additional source of thermal stresses is the difference in the thermal expansion coefficients of materials of directly contacting adjacent segments [20]. It is therefore important to know the strength parameters of thermoelectric materials having relatively high thermoelectric efficiencies in their respective temperature ranges. There are only a quite limited number of studies dealing with the mechanical properties of medium-temperature thermoelectric materials [21]. The available literary data on the strength of thermoelectric materials mainly refer to low-temperature thermoelectric materials based on bismuth and antimony telluride solid solutions [22, 23].

This work presents dynamic uniaxial compression test data on the temperature dependences of the mechanical properties of PbTe and Sn_0.9Pb_0.1Te medium-temperature thermoelectric materials having n and p conductivity types, respectively.

PbTe and Sn_0.9Pb_0.1Te medium-temperature thermoelectric materials were synthesized by direct smelting of raw components taken in the stoichiometric ratio at 1000 °C for 4 h. The synthesized material was crushed to a powder size of 500±40 mm. The specimens were produced by powder compression using the hot extrusion and spark plasma sintering (SPS) methods. Hot extrusion was carried out under a 100-ton hydraulic press at 400–420 °C, 400–500 MPa and an extrusion coefficient of 9. SPS was conducted at 450 °C, 80 MPa and 5 min exposure at the temperature plateau. The specimens for uniaxial compression tests had dimensions of 5 × 5 × 6 mm³.

The mechanical tests were conducted on an Instron 5982 universal automatic tester controlled by the Bluehill Materials Testing Software. The specimen load measuring error was within 0.4% of current readings. The compression tests were carried out at a constant traverse speed of 0.05 mm/min, the path measuring error being ±0.001 mm. The specimens were heated with a hanging split electric oven, the specimen temperature being measured with a Chromel-Alumel thermocouple.

The phase compositions of the specimens were studied using X-ray diffraction on a Bruker D8 diffractometer in CuK_α radiation. The structural element sizes of the PbTe and Sn_0.9Pb_0.1Te specimens were studied on cleaves using scanning electron microscopy under a JSM-6480LV microscope. The cleaves were made at room temperature. The secondary electron images were produced at a 30 kV acceleration voltage.

The electrical conductivities and Seebeck coefficients of the specimens were measured simultaneously using the four-probe and differential methods with an Ulvac ZEM-3 instrument. The heat conductivity was calculated using the formula

κ = D_tC_pd,

where D_t is the temperature conductivity, С_р is the specific heat capacity and d is the density. The temperature conductivity and the specific heat capacity were measured using laser flash and differential scanning calorimetry methods on LFA 457 (Netzsch) and DSC-404C (Netzsch) plants, respectively. The density of the specimens was measured using Archimedes’ method.

According to X-ray diffraction data, the PbTe polycrystalline specimens produced by SPS and extrusion were single-phase. The diffraction maximum intensities suggest that the test specimens had no preferential grain orientations (Fig. 1). The coherent region sizes (CRS) and microdeformations were assessed from diffraction peak broadening. Regardless of production method, diffraction peak broadening for the PbTe specimens was only caused by microdeformations originating from randomly distributed dislocations. The CRS were greater than 200 nm and did not contribute to diffraction peak broadening. The dislocation density in the PbTe specimens produced by SPS was 4 · 10⁵ cm^-2. The dislocation density in the PbTe specimens produced by extrusion was 2 · 10⁷ cm^-2.

Figure 1.

Diffraction patterns of (a) PbTe and (b) Sn_1-xPb_xTe specimens produced using SPS

The p conductivity type materials produced by SPS and extrusion also were single-phase and had no texture. The diffraction patterns of the p conductivity type specimens contained only peaks of the Sn_1-xPb_xTe solid solution. Solid solution stratification or component segregation were not observed in the compacted specimens. The composition of the Sn_1-xPb_xTe solid solution as determined from lattice parameters was stoichiometric, i.e., Sn_0.9Pb_0.1Te. The solid solution composition was homogeneous. The dislocation densities assessed from dislocation peak broadening were 5 · 10⁴ and 3 · 10⁶ cm^-2 for the SPS and extrusion Sn_0.9Pb_0.1Te specimens, respectively.

Thus the initial polycrystalline sintered and hot extruded PbTe and Sn_0.9Pb_0.1Te specimens were single-phase and had no preferential grain orientations. The dislocation density in the extruded specimens was higher than in the SPS specimens. Comparison between the structures of the PbTe and Sn_0.9Pb_0.1Te specimens produced using the same method showed that the dislocation density in the Sn_0.9Pb_0.1Te specimens is by an order of magnitude lower than in the PbTe specimens.

The less defective structure of the Sn_0.9Pb_0.1Te specimens as compared with PbTe for the comparable synthesis method can originate from the fact that the homological temperature at which the specimens were compacted is higher for the SnTe solid solution (T/T_melt = 0.67) than for PbTe (T/T_melt = 0.60). Therefore the diffusion mobility in Sn_0.9Pb_0.1Te is higher than in PbTe and hence the defects will be annealed faster and the crystalline structure will be more perfect in the Sn_0.9Pb_0.1Te specimens than in the PbTe ones regardless of production method.

Figure 2 shows cleavage surface images for the SPS and extruded PbTe and Sn_0.9Pb_0.1Te specimens. There was no significant difference between the cleavage surface patterns of the specimens produced using different methods. The cleavage textures of the PbTe and Sn_0.9Pb_0.1Te specimens reflect the grain structure of the materials because secondary electron imaging occurs by electron reflection from different crystal surfaces having different grain cleavage surface orientations. The cleavage surface images of the PbTe and Sn_0.9Pb_0.1Te specimens clearly show grain facets. The grain shapes are close to isotropic. The images show individual pores located in grain bulk and at grain boundaries. The pore sizes are ~200–400 nm for the PbTe and Sn_0.9Pb_0.1Te specimens produced using different methods, but the extruded PbTe and Sn_0.9Pb_0.1Te specimens contain more pores than the SPS specimens. Grain sizes as assessed from scanning electron microscopy images for the extruded and SPS PbTe specimens are ~20 mm and 25 mm, respectively. The grains in the Sn_0.9Pb_0.1Te specimens are larger than in PbTe produced using the same method. The average pore size in the extruded Sn_0.9Pb_0.1Te specimen was ~35 mm and that in the SPS specimen, ~40 mm.

Figure 2.

Cleavage surface images of (a and b) PbTe and (c and d) Sn_0.9Pb_0.1Te specimens produced by (a and c) extrusion and (b and d) SPS

Figure 3 shows the deformation curves of the extruded and SPS PbTe specimens. The curves suggest that at 20 to 500 °C the deformation to as high as ε = 12 % remains plastic: the linear stress vs deformation curve in the elastic region changes to plastic flow at greater deformations. This is the fundamental difference between these specimens and the low-temperature Bi₂Te₃ based solid solutions in which a brittle-ductile transition occurs at below 200 °C. Deformation of the extruded and SPS PbTe specimens at below 300 °C has a deformation hardening region without fracture, this region being more expressed for the SPS specimens. In this case, estimation of the materials’ yield stress would be incorrect. However, assessing the strength of the specimens from their maximum stress during deformation, one can conclude that it is almost insensitive to material production method. Mechanical stresses in the PbTe specimens reach their maximum values of ~149 MPa at room temperature. With an increase in the test temperature to 200 °C, σ_max decreases by two times to 70 MPa. With a further increase in temperature the evolution pattern of σ_max remains the same.

Figure 3.

Diffraction patterns of (a) extruded and (b) SPS PbTe specimens at temperatures: (1) 20 °С; (2) 200 °С; (3) 300 °С; (4) 500 °С

Figure 4 shows the deformation curves of the extruded and SPS Sn_0.9Pb_0.1Te specimens. It can be seen from Fig. 4 that deformation of Sn_0.9Pb_0.1Te is also plastic over a wide temperature range. The only exclusion is room temperature at which the deformation curve has an elastic/plastic pattern: the stress is the highest in a region of relatively low deformations which is at 2–4% for the extruded specimens and 4–6% for the SPS ones, following which the specimens undergo softening which however does not cause complete fracture. This is most probably because microctacks or twins form during 20 °C deformation. The highest mechanical stresses in the Sn_0.9Pb_0.1Te specimens are achieved, by analogy with PbTe, at room temperature. However, unlike the PnTe specimens for which the production method had no significant effect on σ_max, the hot extruded Sn_0.9Pb_0.1Te specimens have almost twice as high room temperature maximum mechanical stresses as those of the SPS Sn_0.9Pb_0.1Te specimens. The room temperature σ_max values for the extruded and SPS Sn_0.9Pb_0.1Te specimens are 240 and 135 MPa, respectively.

Thus, the extruded specimens have the highest room temperature σ_max for both the PbTe and Sn_0.9Pb_0.1Te materials. This difference decreases with an increase in temperature and then becomes almost zero (at 200 °C for PbTe and at 400 °C for Sn_0.9Pb_0.1Te). This is because an increase in temperature leads to recrystallization making the structure of extruded materials more similar to the structure of SPS materials. σ_max decreases with an increase in temperature more intensely for the extruded specimens. The σ_max vs T curves for PbTe intersect and at above 200 °C the sintered material has but slightly higher strength than the extruded one.

Figure 4.

Diffraction patterns of (a) extruded and (b) SPS Sn_0.9Pb_0.1Te specimens at temperatures: (1) 20 °С; (2) 200 °С; (3) 300 °С; (4) 500 °С

It should be noted that the Sn_0.9Pb_0.1Te solid solution specimens have a higher strength than the PbTe specimens at any temperature and for all production methods.

Figures 5 and 6 illustrate the thermoelectric properties of the extruded and SPS PbTe and Sn_0.9Pb_0.1Te specimens.

Figure 5.

Temperature dependences of (a) electrical conductivity, (b) Seebeck coefficient, (c) heat conductivity and (d) thermoelectric efficiency for (1) extruded and (2) SPS PbTe specimens

It can be seen from the data shown in Fig. 5 that the electrical conductivity of the n conductivity type PbTe specimens decreases with an increase in temperature, the electrical conductivity of the extruded material being higher in the entire temperature range than that of the SPS specimens. The Seebeck coefficients of all the specimens increase with temperature, the SPS PbTe specimens having higher Seebeck coefficients. The heat conductivities of the PbTe specimens decrease with an increase in temperature and are close for the extruded and SPS specimens in the entire temperature range. The thermoelectric efficiency of SPS PbTe is higher than that of the extruded specimens in the 25 to 200 °C range. However, at 400–600 °C range which is the working one for medium-temperature thermoelectric materials the hot extruded PbTe specimens have higher thermoelectric efficiencies. At 600 °C the thermoelectric efficiencies of the extruded and SPS specimens are ~1.4 and ~1.3, respectively. Thus the electrical conductivity of the PbTe specimens is almost insensitive to compaction method. The electrical conductivity of the extruded specimens is higher, possibly due to the formation of point defects during extrusion which increase the concentration of majority carriers.

The temperature dependences of the electrical conductivity and the Seebeck coefficient of the p conductivity type Sn_0.9Pb_0.1Te specimens have a similar pattern to those of PbTe (Fig. 6). However, in contrast to the n conductivity type, the electrical conductivity of the SPS Sn_0.9Pb_0.1Te specimens is higher than that of the extruded specimens in almost the whole temperature range. Accordingly, the Seebeck coefficients of the SPS Sn_0.9Pb_0.1Te specimens are lower than those of the hot extruded Sn_0.9Pb_0.1Te specimens. This difference in the behavior of the electrophysical properties between the specimens produced by extrusion and SPS is caused by the formation of point defects during extrusion which have different effects on the concentration of majority carriers in n and p conductivity type thermoelectric materials. The heat conductivities of all the Sn_0.9Pb_0.1Te specimens are close and, same as for the PbTe specimens, are insensitive to the compaction method used. The thermoelectric efficiencies of the extruded Sn_0.9Pb_0.1Te specimens are higher than those of the SPS specimens in the entire temperature range. The 600 °C thermoelectric efficiencies of the extruded and SPS specimens are 1.0 and 0.9, respectively.

Figure 6.

Temperature dependences of (a) electrical conductivity, (b) Seebeck coefficient, (c) heat conductivity and (d) thermoelectric efficiency for (1) extruded and (2) SPS Sn_0.9Pb_0.1Te specimens

The mechanical properties of medium-temperature PbTe and Sn_0.9Pb_0.1Te thermoelectric materials produced by extrusion and spark plasma sintering were studied. Regardless of production method, the PbTe and Sn_0.9Pb_0.1Te specimens are single-phase and have no preferential grain orientations. The dislocation density in the extruded specimens is noticeably higher than in the spark plasma sintered specimens. Comparison of the structural parameters between the n and p conductivity type specimens showed that the structure of the Sn_0.9Pb_0.1Te specimens is less defective than that of PbTe.

Study of the strength parameters of the specimens as a function of dynamic uniaxial compression temperature showed that deformation in a wide temperature range from 20 to 500 °C is plastic and has no traces of brittle fracture. We showed that the room temperature yield stresses are higher for the extruded specimens and that a decrease in the maximum stress with an increase in temperature is more intense in the extruded specimens; as a result, at above 200 °C the spark plasma sintered material acquires a higher strength than the extruded one.

Regardless of production method the strength parameters of the n conductivity type specimens are higher than those of the p conductivity type PbTe specimens in the entire temperature range.

The extruded PbTe and Sn_0.9Pb_0.1Te specimens have better thermoelectric properties than the spark plasma sintered specimens. The heat conductivities of the PbTe and Sn_0.9Pb_0.1Te specimens are almost insensitive to compaction method. The 600 °C thermoelectric efficiencies ZT of PbTe and Sn_0.9Pb_0.1Te were ~1.4 and 1.0, respectively.

The work was conducted within State Assignment of the Kurchatov Institute Research Center.