Structural Characterization and Thermoelectric Properties of Br-Doped AgSnm[Sb0.8Bi0.2]Te2+m Systems

Herein, we report the synthesis, structural and microstructural characterization, and thermoelectric properties of AgSnm[Sb0.8Bi0.2]Te2+m and Br-doped telluride systems. These compounds were prepared by solid-state reaction at high temperature. Powder X-ray diffraction data reveal that these samples exhibit crystal structures related to the NaCl-type lattice. The microstructures and morphologies are investigated by scanning electron microscopy, energy-dispersive X-ray spectroscopy (EDS), and high-resolution transmission electron microscopy (HRTEM). Positive values of the Seebeck coefficient (S) indicate that the transport properties are dominated by holes. The S of undoped AgSnm[Sb0.8Bi0.2]Te2+m ranges from +40 to 57 μV·K−1. Br-doped samples with m = 2 show S values of +74 μV·K−1 at RT, and the Seebeck coefficient increases almost linearly with increasing temperature. The total thermal conductivity (κtot) monotonically increases with increasing temperature (10–300 K). The κtot values of undoped AgSnm[Sb0.8Bi0.2]Te2+m are ~1.8 W m−1 K−1 (m = 4) and ~1.0 W m−1 K−1 (m = 2) at 300 K. The electrical conductivity (σ) decreases almost linearly with increasing temperature, indicating metal-like behavior. The ZT value increases as a function of temperature. A maximum ZT value of ~0.07 is achieved at room temperature for the Br-doped phase with m = 4.


Introduction
In recent years, the challenge of finding new sources of renewable energy that can generate power from waste heat has attracted considerable interest. It has been estimated that only one-third of produced energy is used efficiently, while the remaining two-thirds are discarded, mainly as waste heat. Therefore, taking advantage of this form of energy would result in an increase in energy efficiency [1,2]. Thermoelectric materials can be used for this purpose due to their ability to generate a potential difference (∆V) from a temperature gradient (∆T). The efficiency of these materials is determined by the dimensionless figure of merit (ZT) defined as ZT = (σS 2 /κ tot )T, where T is the temperature, S is the Seebeck coefficient, σ is the electrical conductivity, σS 2 is the thermopower, and κ is the thermal conductivity given by κ tot = κ e + κ latt (electronic thermal conductivity and lattice thermal conductivity, respectively) [3][4][5][6][7][8].
One of the most common and well-studied thermoelectric materials is PbTe due to its high efficiency; however, the presence of Pb has limited its applications. There has been interest in SnTe as a similar alternative, but spontaneously formed Sn vacancies induce a high carrier concentration, which leads to a low Seebeck coefficient and a high electric contribution to the thermal conductivity [9][10][11]. To increase the efficiency of this material by decreasing the carrier concentration, alloys with AgSbTe 2 and AgBiTe 2 have been previously reported, which form quaternary compounds with the general formula AgSn m MTe m+2 (M = Sb, Bi) [12][13][14][15][16].
To enhance the efficiency of materials that contain Sb in their composition, different types of doping have been reported in recent years, including the use of Bi to replace Sb. Bi has an atomic weight greater than that of Sb, which can increase phonon dispersion, leading to a decrease in the lattice thermal conductivity and, therefore, the total thermal conductivity [15,16]. In 2013, Mohanraman et al. [17] studied the effect of this substitution on p-type Ag(Sb 1−x Bi x )Te 2 material (x = 0; 0.03; 0.05; 0.07; 0.1) and reported a decrease in κ latt , with a minimum value of 0.38 Wm −1 K −1 for x = 0.05 at 510 K compared to the value obtained for AgSbTe 2 (0.52 Wm −1 K −1 ). In 2015, Guin et al. [18] obtained an increase in electrical conductivity from 5 to 51 S cm −1 when AgSnSe 2 was doped with 2% Bi, which led to an increase in ZT compared to the pristine sample. Tan et al. studied the effects of replacing all Sb in AgSn m SbTe m+2 with Bi and reported that Bi is more efficient at neutralizing the Sn vacancies in SnTe than Sb, which leads to a higher Seebeck coefficient [15].
On the other hand, studies with the aim of improving the thermoelectric properties of PbTe have reported the use of halogen anions such as Cl − , I − , or Br − . In 2011, Lalonde et al. reported an improvement in the electrical properties of PbTe when doped with iodine (n-type) at a temperatures between 700 and 800 K, obtaining a decrease in resistivity [19]. In 2018, Li et al. reported a study of SnSe doped with bromine, which caused a fourfold increase in the thermopower for the composition SnSe 0.9 Br 0.147 compared to the pristine sample (4.5 W cm −1 K −2 and 1.1 W cm −1 K −2 , respectively) [20]. In addition, doping generates a change in semiconductor behavior, from p-type to n-type, which is reflected by Hall-effect measurements. The ZT values obtained for samples doped with Br were larger than those for polycrystalline SnSe and can be compared to the values obtained for SnSe single crystals [20]. Guin et al. reported improvements in the room-temperature electrical conductivity of AgBiSe 2 (n-type) by doping with chlorine, bromine, or iodine. This increase was mainly due to an increase in the carrier concentration, from 5.85 × 10 18 (AgBiSe 2 ) to 3.72 × 10 19 carriers per cm 3 (AgBiSe 1.98 Cl 0.02 ), while the mobility decreased slightly for samples doped with 2% halogen compared to the pristine material [18].
In this work, we report the structural characterization and thermoelectric properties of lead-free systems with the general formula of AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m and doped with Br (m = 2 and 4). These compounds were synthesized at 1223 K using solid-state reactions. Powder X-ray diffraction patterns fitted using the Rietveld method are consistent with phases related to the cubic NaCl-type lattice. Scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM) were used to investigate the microstructures and morphologies of these systems. The electrical and thermal transport properties of the samples at low temperature were characterized by measurements of the Seebeck effect, thermal conductivity, and electrical conductivity. The figure of merit (ZT) for temperatures from 10 K to 300 K was evaluated.

Experimental Methods
AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m and Br-doped samples were synthesized under a dry and oxygen-free argon atmosphere using silver powder (99.99% purity, Aldrich, Saint Louis, MO, USA), antimony powder, (99.99% purity, Aldrich), tin powder (99.9% purity, Aldrich), tellurium powder (99.99% purity, Aldrich), bismuth powder (99.99% purity, Aldrich), and tin (II) bromide (Aldrich). Phases with the nominal compositions listed above were prepared via the solid-state reaction of powders of Ag, Sn, Bi, Sb, and Te (as well as SnBr 2 for the Br-doped phases) mixed in stoichiometric proportions, then placed inside evacuated quartz ampoules. The reaction mixture was gradually heated to 1223 K at a rate of +423 K/h and maintained at this temperature for~16 h. Then, the furnace was cooled to room temperature. The AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m (m = 2, 4,) system shows congruent melting points at~680 and~720 • C. A comparison of the PXRD patterns before and after DSC/TG analyses showed no significant changes. The melting point decreased from 700 • C to 680 • C when Bi was substituted by Sb in the pristine phase (m = 4). The same trend was shown when doped with bromine, where the melting point decreased from 720 • C to 660 • C (m = 4). For electrical measurements, the samples obtained via the solid-state reactions were crushed into powders and placed into a quartz cell with a parallelepiped shape. This quartz cell was placed into a Schlenk tube under an argon atmosphere to prevent oxidation of the bulk by air. This tube was then placed into a furnace at 1223 K for 20 min and quickly removed. The obtained ingots were cut and polished for measurements of their electrical transport and thermal properties, with approximate dimensions of 3 × 3 × 8 mm 3 . The density was calculated from the sample's geometry and mass. Table S1 (Supplementary Materials) shows the percentage density of the parallelepiped-shaped samples (>92% theoretical density for all samples).
XRD patterns were obtained at RT using a Bruker D8 Advance powder diffractometer (Bruker, Billerica, MA, USA) with CuKα radiation over the 2θ range of 5-80 • at a step size of 0.01. The collected data were analyzed by Fullprof Rietveld refinement software (https://www.ill.eu/sites/fullprof/php/downloads.html) [21]. A standard LaB 6 sample was used to determine the instrumental profiles. The chemical compositions of the samples were determined by scanning electron microscopy (SEM, JEOL 5400 system, Tokyo, Japan) and energy-dispersive X-ray spectroscopy (EDS, Oxford LinK ISIS microanalyzer, Oxford Instruments, Abingdon, UK). High-resolution transmission electron microscopy (HRTEM) and electron diffraction (ED) patterns were obtained using a JEOL JEM 2100 operating at an accelerating voltage of 300 kV. Samples were prepared by crushing the powders under n-butanol and dispersing them over copper grids covered with a porous carbon film. Semiquantitative chemical analyses were carried out using EDS. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TG) were performed on a Rheometric Scientific STA 1500H/625 thermal analysis system. DSC/TG curves were acquired simultaneously for each sample over a temperature range from room temperature to 1273 K; the samples were heated at 10 K min −1 under flowing argon. Low-temperature thermoelectric properties were obtained in a helium-cooled cryostat using a PPMS system (Quantum Design) for temperatures from 10 K to 300 K. Hall-effect measurements were performed using an ECOPIA HMS 2000 system. Electrical contacts of gold were deposited by sputtering on pellets used in the thermoelectric measurements. The Hall coefficient at a field of ±0.556 T was obtained from linear fits of the Hall resistivity using the van der Pauw method at RT.

X-ray Powder Diffraction (XRD) and Electron Microscopic Characterization (SEM-EDS and HRTEM)
The XRD patterns of all samples were fully indexed in the Pm-3m space group, with the exception of two very weak impurity peaks, within the detection limits of the technique. The shape and intensity of the XRD peaks indicate the high crystallinity of all the telluride samples, as shown in Figure 1. The experimental XRD patterns were compared with those of previously reported pristine AgSn m SbTe 2+m samples, indicating that they are isostructural, and the measured d-spacings are in good agreement with the calculated values. Increasing Sn content in AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m leads to an increase in the cell parameters (Table S1, Supplementary Materials). For the Br-doped compounds, the cell parameters increase gradually as tellurium is replaced by bromine. Figure 1b   The inset in Figure 1b shows the a-lattice parameter as a function of Ag[Sb0.8Bi0.2]Te2% plotted in the form of the 100/(1 + m) molar ratio. This linear dependence has previously been observed for AgSnxBiTex+2 and AgSnmSbSe2Tem systems [15,23].
SEM-EDS analyses of the powder samples indicate that the chemical composition of all the phases is uniform throughout the scanned region, within the detection limits of the technique, as represented for AgSn2[Sb0.8Bi0.2]Te4 ( Figure 2 and Table 1) and AgSn4[Sb0.8Bi0.2]Te5.97Br0.03 ( Figure S1, Supplementary Materials). EDS microprobe chemical analysis performed on several areas of the AgSn2[Sb0.8Bi0.2]Te4 sample revealed an average chemical composition consistent with the nominal composition (as shown in Figure  2, yellow squares). The chemical distributions of bismuth, antimony, silver, tellurium, tin, and bromine in both samples are homogeneous. The inset in Figure 1b shows the a-lattice parameter as a function of Ag[Sb 0.8 Bi 0.2 ]Te 2 % plotted in the form of the 100/(1 + m) molar ratio. This linear dependence has previously been observed for AgSn x BiTe x+2 and AgSn m SbSe 2 Te m systems [15,23].
SEM-EDS analyses of the powder samples indicate that the chemical composition of all the phases is uniform throughout the scanned region, within the detection limits of the technique, as represented for AgSn 2 [Sb 0.8 Bi 0.2 ]Te 4 ( Figure 2 and Table 1) Figure 2). The chemical formula averaged over these areas is shown.    Figure 2). The chemical formula averaged over these areas is shown.

Spectrum
Mass
The temperature dependence of the Seebeck coefficient typical of metallic or degenerate semiconductors is expressed by the following formula: where S is the Seebeck coefficient, m* is the effective mass, k B is Boltzmann's constant, e is the charge of an electron, h is Planck's constant, and n is the carrier concentration [3,7]. The Seebeck coefficient was fitted in the temperature range from 10 K to 300 K for AgSn 2 Figure 4b. Hall-effect measurements revealed that the carrier concentration of the telluride samples was in the range of +2.5-1.7 × 10 19 cm −3 at room temperature. These measurements imply an m* of~0.91·m 0 assuming an acoustic phonon scattering mechanism (r = −1/2) and an m* of 0.32·m 0 assuming an ionized impurity scattering mechanism (r = 3/2). Density-of-states effective masses, m*~0.61−0.99·m 0 and m*~0.32·m 0 at room temperature, were previously reported for Sn 0.85 Sb 0.15 Te and PbTe, respectively [27,28]. Figure 5a shows plots of the temperature dependence of the electrical conductivity (σ) for AgSn 4 where S is the Seebeck coefficient, m* is the effective mass, kB is Boltzmann's constant, e is the charge of an electron, h is Planck's constant, and n is the carrier concentration [3,7]. The Seebeck coefficient was fitted in the temperature range from 10 K to 300 K for AgSn2[Sb0.8Bi0.2]Te4 and AgSn2[Sb0.8Bi0.2]Te3.97Br0.03, as shown in Figure 4b. Hall-effect measurements revealed that the carrier concentration of the telluride samples was in the range of +2.5-1.7 × 10 19 cm −3 at room temperature. These measurements imply an m* of ~0.91·m0 assuming an acoustic phonon scattering mechanism (r = −1/2) and an m* of ~0.32·m0 assuming an ionized impurity scattering mechanism (r = 3/2). Density-of-states effective masses, m* ~ 0.61−0.99·m0 and m* ~ 0.32·m0 at room temperature, were previously reported for Sn0.85Sb0.15Te and PbTe, respectively [27,28]. Figure 5a shows plots of the temperature dependence of the electrical conductivity (σ) for AgSn4[Sb0.8Bi0.2]Te6 and the Br-doped phases. σ decreases almost linearly with increasing temperature, indicating that the samples show metal-like behavior. AgSn4[Sb0.8Bi0.2]Te6 has similar electrical conductivity to the Br-doped phases and lower values than the pristine phases (Table 2). For example, our measured value of σ is 1232 S cm −1 for AgSnm[Sb0.8Bi0.2]Te2+m (m = 4), and the reported σ value of AgSn2SbTe4 is ~1800 S cm −1 at 300 K [13].

Conclusions
In summary, polycrystalline AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m and Br-doped phases were successfully prepared at high temperatures using solid-state reactions. EDS mapping analysis using SEM and HRTEM indicates that the chemical substitution of bismuth and bromine in the pristine AgSn m SbTe 2+m samples is homogeneous within experimental error. HRTEM images and electron diffraction patterns reveal the existence of nanoregions with different orientations or symmetries, whether cubic or tetragonal. X-ray diffraction (XRD) and electron diffraction (ED) data are consistent with a cubic NaCl-type superstructure. The Seebeck coefficient (S) of AgSn m [Sb 0.8 Bi 0.2 ]Te 2+m ranges from +40 to 74 µV·K −1 . The total thermal conductivity (κ tot ) is decreased by bromine doping (~1.2 W m −1 K −1 at RT). S and κ tot increase over the studied temperature range, whereas the electrical conductivity (σ) decreases with increasing temperature. It is worth mentioning that all samples show metallike behavior. Finally, a maximum ZT value of~0.07 was obtained at room temperature for the Br-doped AgSn 4 [Sb 0.8 Bi 0.2 ]Te 6 phase. This study deepens the understanding of rocksalt-type telluride phases and provides a new approach for optimizing TE performance by introducing chemical substitutions.