Effect of Ta Doping on the Microstructure and Thermoelectric Properties of Bi₂O₂Se

Abstract

In this study, Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics were prepared using a synthesis method combining high-energy ball milling and cold pressing. Furthermore, the effects of tantalum (Ta) doping on the microstructure and thermoelectric properties of Bi₂O₂Se were systematically investigated. The results indicate that Ta doping effectively improves the carrier concentration and mobility, thus increasing the electrical conductivity from 8.75 S cm⁻¹ to 39.03 S cm⁻¹ at 323 K. Consequently, the power factor is improved, reaching a maximum value of 124 μW m⁻¹ K⁻² for the Bi_1.92Ta_0.08O₂Se sample at 773 K. Moreover, the thermal conductivity of Bi_1.96Ta_0.04O₂Se is reduced to 0.50 Wm⁻¹ K⁻¹. Finally, the maximum dimensionless figure of merit (ZT) value of the Bi_1.94Ta_0.06O₂Se sample reached 0.18, which was 64% higher than that of Bi₂O₂Se (0.11). These results indicate that Ta doping and high-energy ball milling can optimize the electrical and thermal properties and thus improve the thermoelectric properties of ceramics.

1. Introduction

Energy shortage and environmental pollution have recently attracted much attention. Fossil fuels, currently the main energy source, generate substantial amounts of waste heat when they are used, which may be able to be used as an important source of additional energy. Lightweight and small-sized thermoelectric (TE) materials, which can be used in harsh environments, and can convert waste heat energy into electricity. TE materials differ from many other materials; they are functional materials with great potential [1,2,3]. The performance of a TE material is evaluated using the dimensionless figure of merit ZT = σS²T/κ, where S is the Seebeck coefficient, σ is the electrical conductivity, T is the absolute temperature, and κ is the thermal conductivity. High-performance TE materials have high S, high σ, and low κ; however, these parameters are strongly coupled [4,5,6]. Thus, developing high-performance TE materials and improving their TE properties has become an important research focus. Based on the concept of a phonon glass electronic crystal, layered materials can separately regulate electrical and thermal conductivities, thereby producing TE materials with excellent TE performance [7,8,9,10,11,12].

The main TE material systems investigated to date include SiGe, PbTe, Bi₂Te₃/Sb₂Te [13,14,15,16,17,18,19], metal silicides [20], and oxide TE materials. Among them, BiCuSeO, which has a natural superlattice structure, has attracted much attention owing to its low thermal conductivity (0.7 W m⁻¹ K⁻¹ at 300 K) [21,22,23,24]. The ZT value of BiCuSeO was increased from 0.5 to 1.5 by doping modification and changing the material preparation process [25,26]. However, despite the remarkable progress in investigating this system, many deficiencies remain in practical applications. Specifically, the poor TE performance of the matched n-type oxide makes its conversion efficiency relatively low for applications. Therefore, finding matching layer-structured n-type oxide TE materials with low thermal conductivities and exploring them extensively is crucial. The Bi₂O₂Se system, which has a similar structure (Figure 1a) and thermal conductivity (~1.1 W m⁻¹ K⁻¹ at 300 K) to that of the BiCuSeO system, has attracted research attention. However, owing to the low carrier concentration of Bi₂O₂Se (~10¹⁵ cm⁻³), its conductivity is not similar to that of high-performance TE materials. Therefore, many studies have sought to improve its electrical conductivity and TE properties. It is reported that introducing Bi vacancies [27] (i.e., partially substituting La³⁺ [28] or Sb³⁺ [29] for Bi³⁺, partially substituting Sn⁴⁺ for Bi³⁺ [30], or partially substituting Nb⁵⁺ for Bi³⁺ [31]), can improve the Seebeck coefficient and electrical conductivity. The reported ZT values have only been increased to 0.3. Therefore, further research is needed to improve the TE performance of the Bi₂O₂Se system.

Herein, Ta⁵⁺ was substituted for Bi³⁺ ions in Bi₂O₂Se, and the effect of tantalum (Ta) doping on the microstructure and TE properties was investigated. The Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics were prepared using a synthesis method combining high-energy ball milling and cold pressing. The results indicate that the maximum ZT value of the Bi_1.94Ta_0.06O₂Se sample is 0.18, which was 64% higher than that of Bi₂O₂Se (0.11).

2. Experimental

2.1. Sample Preparation

To prepare the Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics, Bi (99.99%, Aladdin), Bi₂O₃ (99.999%, Aladdin), Se (99.999%, Aladdin), and Ta₂O₅ (99.9%, Aladdin) powders were employed as the raw materials and mixed in stoichiometric ratios. Furthermore, single-phase Bi_2−xTa_xO₂Se was obtained by grinding the raw materials with a 10 mm carbide ball using a ball:material ratio of 30:1 and a rotational speed of 490 rpm for 2 h in a single-tank planetary high-energy ball mill (MTI Corp., Richmond, CA, USA). Subsequently, the Bi_2−xTa_xO₂Se raw materials were loaded into a strip rubber balloon, which was placed in a manual cold isostatic press container (YLJ-CIP-500 M, Hefei Kejing Materials Technology Co., Hefei, China) and cold-pressed at 300 MPa for 20 min. Finally, the pressed sample was placed in a vacuum quartz tube and calcined at 673 K for 2 h to obtain the Bi_2−xTa_xO₂Se series ceramics.

2.2. Characterization

The phase of Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics was characterized using X-ray powder diffractometry (XRD; Bruker D8 Advance, Karlsruhe, Germany). Furthermore, the diffractometer was operated using Cu–Kα radiation over a 2θ range of 20–60° with a scan rate of 4 °/min at 40 kV and 40 mA. Energy-dispersive spectroscopy was performed using field-emission scanning electron microscopy (Merlin Compact, Carl Zeiss, Oberkochen, Germany) equipped energy-dispersive X-ray spectroscopy (EDS, Oxford AZtec X-Max, Oxford, UK and Energy resolution ≤127 eV (Mn-Kα)) spectrometer. The valence states were determined using X-ray photoelectron spectroscopy with the monochromated Al Kα line as the X-ray source and pass energy of 1486.6 eV, and Au standard sample (XPS, Thermo Scientific K-Alpha, Waltham, MA, USA, energy resolution is 0.5 eV). The sintered bulks were cut into bar-shaped specimens, 3 mm × 3 mm × 15 mm along the radial direction to test the electrical transport properties. The Seebeck coefficients and resistivity were measured by commercial equipment (ZEM-3, Ulvac-Riko, Kanagawa, Japan) using standard DC four-probe technology over the temperature range of 300 K–773 K with a step size of 50 K under He atmosphere. The bulk samples were cut into small wafers with approximate dimensions of ϕ12.5 mm × 2 mm for measuring the thermal transport properties. The coefficient of thermal diffusion (D) was measured on the laser thermal conducting instrument (LFA-457, Netzsch, Selb, Germany) from 300 K to 773 K with a step size of 100 K under a continuous Ar flow. The heat capacity of the samples was measured using a differential scanning thermal analyzer (DSC 8000, PerkinElmer, Waltham, MA, USA) with a heating rate of 10 K min⁻¹, and a flow rate of 20 mL min⁻¹ under argon; α-Al₂O₃ was used as a reference material. The densities of the samples were measured using Archimedes’ method. The thermal conductivity of each sample was calculated using k = ρCpD. Room-temperature Hall coefficients were measured on a physical property measurement system (Quantum Design, San Diego, CA, USA), and the carrier density n and carrier mobility µ were obtained using n = 1/(eRH) and µ = σRH, respectively.

3. Results and Discussion

Figure 1b depicts the XRD patterns for all the Bi_2−xTa_xO₂Se ceramics. The diffraction peaks of the pure Bi₂O₂Se sample shown in the figure are consistent with those of the doped samples. Comparing these results with those from the standard card for Bi₂O₂Se (PDF#73-1316) revealed that the diffraction peaks of all the samples occur at the same locations and that there are no extra peaks, which verifies that all the synthesized samples are single-phase samples. Furthermore, the lattice parameters of the materials were calculated from the following equations, and the results are listed in

Table 1. Lattice constants and crystallite sizes of Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics.

Samples	a (Å)	c (Å)	Grain Size (Å)
x = 0	3.88317	12.20968	164
x = 0.02	3.88791	12.24022	157
x = 0.04	3.89506	12.13847	135
x = 0.06	3.89414	12.21512	145
x = 0.08	3.90810	12.15134	171

. Equation (1) is the Bragg’s law, and Equation (2) provided the interplanar spacing d between adjacent tetragonal (hkl) lattice planes. Since the ionic radius of Ta⁵⁺ (0.74 Å) is smaller than that of Bi³⁺ (0.96 Å), lattice distortion occurs, and the crystal plane spacing of the doped sample decreases.

$$\mathsf{\lambda }=2{\mathrm{d}}_{\mathrm{hkl}}{\sin \mathsf{\theta }}_{\mathrm{hkl}}$$

(1)

$$\frac{1}{{\mathrm{d}}^2}=\frac{{\mathrm{h}}^2+{\mathrm{k}}^2}{{\mathrm{a}}^2}+\frac{{\mathrm{l}}^2}{{\mathrm{c}}^2}$$

(2)

The resulting average crystallite sizes calculated using the Debye–Scheller formula depicted in Table 1. When the Ta doping concentration is less than or equal to 0.06, the average crystallite sizes of the samples are smaller than that of the undoped sample, due to the pinning effect of Ta⁵⁺ dopants in grain boundaries and the dragging effect between Ta⁵⁺ dopants and grain boundaries, thereby restricting the movement of grain boundaries [22]. However, when the doping concentration is 0.08, the crystallite size of the sample is a little greater than that of the pure Bi₂O₂Se sample. This is due to a larger disorder in the layered structure of the Bi₂O₂Se phase containing Zr, which led to an increase in the crystallite size [33]. Figure 2 shows the scanning electron microscopy (SEM) and EDS elemental mapping images of the Bi₂O₂Se and Bi_1.94Ta_0.06O₂Se ceramics. The results show that Bi, O, Se, and Ta are uniformly distributed in the samples, indicating that no impurities are present. These findings are consistent with the XRD results.

Figure 3 illustrates the high-resolution spectrograms of the elements in each Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramic. Figure 3a,b shows the full XPS spectrum of Bi₂O₂Se and Bi_1.96Ta_0.04O₂Se. Ta can be observed in the doped sample. Combining the XRD results (Figure 1) and the element distribution diagram (Figure 2) confirms that Ta was successfully incorporated into the Bi₂O₂Se system using the high-energy ball milling method. The high-resolution spectrum of Bi 4f in Figure 3c shows that the 4f orbital of Bi is split into two peaks, which are Bi 4f_7/2 (158.59 eV) and Bi 4f_5/2 (163.89 eV). Thus, Bi 4f has a spin–orbit splitting of 5.3 eV, indicating its presence in the Bi³⁺ valence state in the sample [34]. Compared with the pure Bi₂O₂Se sample, the binding energy peak of the Bi atom in the doped sample is shifted to a lower critical energy. This phenomenon occurs primarily because the electronegativity of the Ta atom (1.5) is lower than that of the Bi atom (2.02). Thus, Ta doping increases the electron density around the Bi atom. Therefore, the 4f_7/2 and 4f_5/2 binding energies of the 4f orbital of the Bi atom are reduced. Moreover, after Ta doping, the critical energy peak of the Bi atom is similar to that of the Bi₂O₂Se sample, indicating that Ta doping has a limited effect on the valence state of Bi in Bi_2−xTa_xO₂Se. The low binding energy peak of the O atom is shown in Figure 3d and corresponds to lattice oxygen, while the high binding energy peak corresponds to vacancy oxygen. The area of the binding energy peaks of oxygen vacancies in the Bi_1.94Ta_0.06O₂Se sample is more significant than that of the Bi₂O₂Se sample, which occurs because Ta doping induces lattice distortion and increases the density of lattice defects in the Bi_2−xTa_xO₂Se samples. These results are similar to those reported in the literature [35,36]. In Figure 3e, the 3d orbital of the Se atom is split into four peaks. The Se atoms with lower binding energies indicate the existence of Se²⁻, while the higher binding energy peaks of Se 3d_5/2 and 3d_3/2 prove that the Se atoms interact between layers. The binding energy peaks of the O and Se atoms in the same Bi_1.94Ta_0.06O₂Se sample are shifted to lower binding energies, primarily caused by the doping with Ta atoms, which is similar to the results reported in the literature [30]. The 4f orbital of the Ta atom is split into two peaks; the binding energies of Ta 4f_7/2 and Ta 4f_5/2 are 25.08 eV and 28.08 eV, respectively. These results show that the spin splitting of the 4f orbital of the Ta atom is 3.0 eV, indicating that Ta ions exist in the pentavalent (Ta⁵⁺) form in the doped samples.

Figure 4a illustrates the temperature-dependent electrical conductivity σ of Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics. As the temperature increases, the electrical conductivities increase gradually, indicating that these samples have semiconductor properties. Furthermore, after doping a sample with Ta at the Bi site, its electrical conductivity at room temperature is increased significantly, reaching a maximum of 39.03 S cm⁻¹. After Ta doping, the σ of the Bi_2−xTa_xO₂Se sample initially decreases and subsequently increases as the temperature increases, and an inflection point appears at 400 K. From room temperature to 400 K, the σ of the Bi_2−xTa_xO₂Se sample has a slight downward trend with the rise of temperature. When the temperature is higher than 400 K, σ increases further with an increase in temperature, showing the semiconductor properties. Then, σ increases significantly with an increase in doping concentrations, reaching 53.43 S cm^{− 1} at 773 K at x = 0.04, which is about four times that of pure Bi₂O₂Se (13.49 S cm⁻¹). To understand the electrical transport performance of the Bi_2−xTa_xO₂Se samples, their Hall coefficients were measured, and the carrier densities ($n=\frac{1}{e{R}_H}$) were calculated from the results.

Table 2. Carrier densities, Hall mobilities, effective masses m*, and densities ρ of Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics at room temperature.

Samples	n (1018 cm−3)	μ (cm2 V−1 S−1)	m* (m0)	ρ (g/cm3)
x = 0	1.05	30.65	0.1814	6.967
x = 0.02	2.11	43.68	0.2785	7.070
x = 0.04	2.73	89.52	0.3962	7.067
x = 0.06	3.28	54.55	0.4031	7.036
x = 0.08	4.41	35.31	0.4858	7.035

shows the carrier densities n, carrier mobilities μ, effective masses m*, and sample densities ρ for the Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) samples. Table 2 shows that, as the doping concentration increases, the carrier density increases from 1.05 × 10¹⁸ to 4.41 × 10¹⁸ cm⁻³. The mobilities initially increase and then decrease as the doping concentration increases. After Ta⁵⁺ replaces Bi³⁺, each Ta⁵⁺ provides two additional electrons, and the defect reaction is as follows:

$${\mathrm{Ta}}^{5+}\stackrel{{\mathrm{Bi}}^{3+}}{\to }{\mathrm{Ta}}_{{\ddot{\mathrm{B}}}_1}+2 \mathrm{e} \prime$$

(3)

Therefore, doping with Ta induces an increase in the carrier density, which increases the conductivity.

Figure 4b shows S as a function of temperature for the Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics. The S of all the Bi_2−xTa_xO₂Se samples is negative, indicating that they are electron-dominated n-type semiconductors. The absolute value of the Seebeck coefficient of the undoped samples increased from |−152.84 μV K⁻¹| (323 K) to |−243.48 μV K⁻¹| (773 K). Additionally, Ta doping makes the absolute value of the Seebeck coefficient significantly smaller than that of the pure Bi₂O₂Se. Thus, when the doping concentration was 0.04, the absolute value of the Seebeck coefficient decreases to the minimum. With increasing doping concentrations, S decreased from the original |−152.84 μV K⁻¹| to |−71.31 μV K⁻¹| (323 K), while S decreased from |−243.48 μV K⁻¹| to |−142.85 μV K⁻¹| (773 K). According to the Mahan–Sofo theory [31,37], the Seebeck coefficient S can be expressed as $S=\frac{8{\pi }^2{k}_B{}^{2}T}{3e{h}^2}{m}^{*}{\left(\frac{\pi }{3n}\right)}^{2/3}$, where k_B is Boltzmann’s constant, e is the electron charge, ħ is the reduced Planck constant, n is the carrier density, and $m^{*}$ is the effective mass of the carrier. This Equation shows that the Seebeck coefficient is related to the effective mass ($m^{*}$) and the carrier density. The effective masses $m^{*}$ of the Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics calculated using this formula are listed in Table 2. This table shows that the effective masses of the Bi_2−xTa_xO₂Se sample carriers increase with increased Ta doping. The value of $m^{*}$ increases from 0.1814 m₀ for pure Bi₂O₂Se to 0.4858 m₀ for the Bi_1.92Ta_0.08O₂Se sample. However, the effective mass and carrier densities increase with an increase in the doping concentration, ultimately reducing the Seebeck coefficient of the doped samples. The power factor (PF) of the samples increases significantly due to Ta doping at the Bi sites [38]. Although the Seebeck coefficient decreased, the increase in conductivity led to an increased PF. Figure 4c shows the temperature-dependent PF for the Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics. The PF increases with the temperature, and further increases after doping. At 773 K, the PF value is increased from 80 μW m⁻¹ K⁻² to 124 μW m⁻¹ K⁻². Compared with pure Bi₂O₂Se, the PF value is increased by 55%.

The total thermal conductivity is calculated from the Equation $\kappa =\rho C_{p}D$, where ρ is the sample density (Table 2), Cp is the heat capacity, and D is the thermal diffusivity. The temperature -dependent thermal conductivity for Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics is shown in Figure 5a. Over the temperature range, most of the doped samples maintained low thermal conductivities. The thermal conductivity of pure Bi₂O₂Se decreased from 0.99 W m⁻¹ K⁻¹ (323 K) to 0.55 W m⁻¹ K⁻¹ (773 K) owing to the enhancement of phonon scattering and the decrease in the thermal diffusion coefficient due to high temperature [39]. The thermal conductivity of the Bi_2−xTa_xO₂Se sample is reduced when the doping concentrations of the samples are x < 0.06. The thermal conductivity reached a minimum at x = 0.04, decreasing from 0.97 W m⁻¹ K⁻¹ (323 K) to 0.50 W m⁻¹ K⁻¹ (773 K) with increasing temperature. Thus, when the doping concentration exceeds 0.06, the thermal conductivity increases due to the increase in grain size.

Generally, the total thermal conductivity is composed of electronic thermal conductivity ${\kappa }_e$ and lattice thermal conductivity ${\kappa }_{\mathrm{l}}$ ($i.e., \kappa ={\kappa }_e+{ \kappa }_{\mathrm{l}}$). The electronic thermal conductivity is usually calculated from the Wiedemann–Franz relationship (${\kappa }_e=L\sigma T$), where σ is the electrical conductivity, T is the absolute temperature, and L is the Lorentz constant $[L=1.5+exp\left(\frac{-\left|S\right|}{116}\right)]$. The Sommerfeld value was used for the Lorentz constant, L₀= 2.44 × 10⁻⁸ W Ω K⁻² [40]. Figure 5b shows the temperature dependence of the electronic thermal conductivity (${\kappa }_e$) of the Bi_2−xTa_xO₂Se system. After Ta doping at the Bi sites, the electronic thermal conductivity increases compared with that of pure Bi₂O₂Se, which is mainly caused by the increase in carrier density after doping. In addition, the value of the electronic thermal conductivity is much smaller than that of the total thermal conductivity; even at 773 K, the maximum value of the electronic thermal conductivity only reaches ${\kappa }_e$ = 0.1 W m⁻¹ K⁻¹. Thus, the lattice thermal conductivity dominates the total thermal conductivity.

The lattice thermal conductivity κ_l as a function of temperature for Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) is shown in Figure 5c; it is consistent with the total thermal conductivity. After Ta doping, the lattice thermal conductivity of the Bi_2−xTa_xO₂Se (x = 0.02, 0.04, and 0.06) first decreases, mainly because doping with Ta produces lattice defects that cause increased phonon scattering. In addition, when the doping concentration exceeds 0.06, the value of κ_l increases with increased doping concentration. Specifically, when the doping concentration reaches 0.08, both the lattice (κ_l) and total thermal conductivities (κ) are larger than those of Bi₂O₂Se. This is because phonon scattering is reduced when the grain size increases. These results indicate that lattice distortion and phonon scattering induced by grain boundaries jointly affect the thermal conductivity of the Bi_2−xTa_xO₂Se system.

The ZT values of Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.04, 0.06, and 0.08) ceramics depend on the electrical and thermal transport properties, and the temperature dependence of ZT is shown in Figure 5d. The ZT values of all the doped samples are significantly better than those of pure Bi₂O₂Se for the entire range of tested temperatures. The ZT value of Bi_1.94Ta_0.06O₂Se reaches a maximum of 0.18 at 773 K, an increase of 64% compared to the ZT value of Bi₂O₂Se (0.11).

4. Conclusions

Bi_2−xTa_xO₂Se (x = 0, 0.02, 0.06, and 0.08) ceramics were successfully prepared using a facile synthesis method that combines high-energy ball milling and cold pressing. The conductivity increased with Ta doping, reaching a maximum of 34.91 S cm⁻¹ (323 K). A maximum PF of 124 μW m⁻¹ K⁻² was obtained for the Bi_1.94Ta_0.06O₂Se sample, which is approximately 1.55 times that of pure Bi₂O₂Se. The thermal conductivity of the Bi_1.96Ta_0.04O₂Se sample was 0.5 W/mK at 773 K. Bi_1.94Ta_0.06O₂Se reached a maximum value of ZT 0.18 at 773 K, which is a 64% increase compared to the ZT value of pure Bi₂O₂Se (0.11). Therefore, the Ta doping of Bi₂O₂Se was proposed as a new idea for increasing the carrier densities of these TE ceramics.