Abstract
SnSe2, a layered posttransition metal chalcogenide, has attracted attention as a high-efficiency thermoelectric material owing to the intrinsic low thermal conductivity. Herein, a series of (, 0.025, 0.0375, 0.075, 0.1, and 0.125) samples was synthesized to examine the influence of Te doping on electrical, thermal, and thermoelectric properties of -type SnSe2 alloys. Interestingly, carrier concentration and mobility were simultaneously increased for and 0.0375. Therefore, electrical conductivity is increased for and 0.0375 compared to that for the pristine sample, resulting in power factor increase to 0.25 mW/mK2 for by 12% at 790 K. However, reductions in the electrical conductivity were observed for the samples with due to the decrease in carrier mobility for , resulting in the decrease of power factor. The lattice thermal conductivity slightly reduced for the doped samples owing to point defects of Te and vacancies originating from Te doping. Consequently, the thermoelectric figure of merit () was increased to 0.45 and 0.49 for Sn(Se1.975Te0.025)2 () and Sn(Se1.9625Te0.0375)2 () samples at 790 K, respectively, which was enhanced by 40% and 53% compared to that for undoped SnSe2. The enhanced electrical transport properties were validated by weighted mobility, density-of-state effective mass, and quality factor, and the reduction of the lattice thermal conductivity is analyzed by the Debye-Callaway model.
1. Introduction
Thermoelectric materials can improve the efficiency of energy usage because they can convert waste heat into electricity. The performance of thermoelectric materials depends on dimensionless thermoelectric figure of merit, , where , , , , and denote the electrical conductivity, Seebeck coefficient, absolute temperature, electronic thermal conductivity, and lattice thermal conductivity, respectively [1, 2]. However, the complex interdependence between , , and hinders the realization of high for enhanced thermoelectric performance [3, 4]. Nevertheless, typical approaches adopted to achieve a high include improving the power factor () by carrier concentration optimization or reducing by inducing additional phonon scattering [5, 6]. For example, Liu et al. achieved zT of 0.30 for Bi1.92Ge0.08O2Se at 823 K, which demonstrated an improvement of 233% compared to that of pristine Bi2O2Se, by adjusting the carrier concentration to 1015–1019 cm-3 [7]. Orabi et al. synthesized Ga-doped Sn1.03Te samples, which exhibited a significantly reduced value, resulting in a significant enhancement in approaching ~1 at 873 K for Sn0.96Ga0.07Te [8].
Layered metal chalcogenides including posttransition metals, such as InSe [9, 10], In4Se3 [11], SnSe [12], SnSe2 [13–16], and SnTe [17], exhibit low intrinsic values and have thus attracted attention as potential thermoelectric materials. SnSe2, a CdI2-type semiconductor with 0.97 eV [18], can be considered an -type thermoelectric material with a low cost, nontoxic nature, and earth abundance [19, 20]. However, it has a low carrier concentration of 1017–1018 cm-3, resulting in a low power factor and [21, 22]. Thus, increasing the carrier concentration is a straightforward approach to improve the power factor and . Based on band structure calculations, Sun et al. calculated that the value of -type SnSe2 along the direction could be controlled at 0.06, 0.24, 0.82, and 0.46 at 790 K for 1017, 1018, 1019, and 1020 cm-3 of carrier concentration, respectively [23]. Xu et al. reported that SnSe2-based compound exhibited value of 0.4 at 673 K owing to substantial Cl doping [24]. Wu el al. reported of 0.3 at 300°C in Cl-doped SnSe2 [25]. Zhou et al. reported of 0.67 for SnSe2 by Cu intercalation and Br doping [26].
Herein, Te element is employed as a dopant to increase the carrier concentration of SnSe2. A series of (, 0.025, 0.0375, 0.075, 0.1, and 0.125) samples was synthesized to investigate the influence of Te doping on the electrical and thermoelectric properties of SnSe2. Indeed, Cho et al. reported the increased carrier concentration from to in SeSe via 15 at.% Te substitutional doping [27], and Li et al. also show enhanced thermoelectric performance in Te-doped SnSe [28]. For SnSe2, the electrical conductivity increased by the small Te doping of and 0.0375 compared to that for the pristine sample, resulting in power factor increase to 0.25 mW/mK2 for by 12% at 790 K. The lattice thermal conductivity slightly reduced by Te doping and vacancy formation originating from Te doping. For a better understanding of the electronic transport properties of these materials, the weighted mobility, density-of-state effective mass, and quality factor were calculated and analyzed. In addition, reduced lattice thermal conductivity is analyzed by the Debye-Callaway model.
2. Experimental Methods
Polycrystalline (, 0.025, 0.0375, 0.075, 0.1, and 0.125) samples were synthesized in a vacuum-sealed quartz tube via a conventional solid state reaction. Stoichiometric amounts of Sn (99.99%), Te (99.999%), and Se (99.999%) were weighed and heat-treated at 1073 K for 8 h, followed by water quenching and annealing at 850 K for 72 h [13, 14]. The obtained ingots were pulverized, and the crystalline phases of the powders were characterized through X-ray diffraction (XRD, D8 Discover, Bruker, Germany) using Cu Kα1 radiation. Powders were sintered using spark plasma sintering (SPS-1030, Japan) at 773 K for 5 min under 65 MPa. Further, the composition-dependent vibration modes of the samples were analyzed using confocal Raman spectroscopy (LabRAM HR Evolution, HORIBA, Japan) with excitation wavelengths of 50–400 cm-1 supplied with a 532 nm laser. The and were measured using a thermoelectric property measurement system (ZEM-3 M8, Advance Riko, Japan) along the direction parallel to the pressing axis in the range of 300 to 790 K. The Hall carrier concentrations () and mobilities () were measured by using the Hall measurement system (HMS-5300, Ecopia, Korea). The density-of-state mass was obtained based on the measured and using the Pisarenko plot. The total thermal conductivity was calculated using , where , , and denote the thermal diffusivity, density, and specific heat capacity, respectively. The value of each sample was measured for 300-790 K using a laser flash analysis (LFA457, Netzsch, Germany) across the perpendicular planes of the pressing direction to ensure appropriate determination of . Further, was considered as the theoretical density of hexagonal SnSe2 with a value of 5.950 g/cm3 [29]. was estimated based on the following empirical relationship [30]: . The changes in and resulting by small Te doping were considered negligible.
3. Results and Discussion
Figure 1(a) shows the XRD patterns of the (, 0.025, 0.0375, 0.075, 0.1, and 0.125) powders. Based on our process, hexagonal SnSe2 samples (Joint Committee on Powder Diffraction Standards (JCPDS) #01-089-2939) were successfully synthesized. However, in the XRD patterns of all Te-doped samples, unexpected peaks at ~27.56°, labeled by purple star symbols, were observed; these could be indexed to the (011) peaks of Te (JCPDS #03-065-3370). The intensities of the (001) peak for the (, 0.025, 0.0375, 0.075, 0.1, and 0.125) samples denoted by , intensities of the (011) peak for Te denoted by , and their ratios (Imax/Iimpurity) are listed in Table 1. Even though there is a variation on due to the different amounts of powders between 2958 and 8865, the relative ratios gradually increased with the Te content, indicating that the amount of Te impurities in the samples increased with . The phase diagram of the Sn–Te system indicates the presence of a single chemical compound SnTe [31], suggesting that the stoichiometric addition of Te in the SnSe2 compound could result in the Te deficiency, leading to the creation of anion vacancies in the compound and Te impurities. The calculated lattice parameters along the in-plane and out-of-plane ( and , respectively) directions are indicated with error bars in Figure 1(b).

(a)

(b)
Figure 2(a) presents the Raman spectra of the series of samples. The major peaks observed around 106 and 181 cm-1 can be assigned to the in-plane () and out-of-plane () vibration modes of SnSe2, respectively [32, 33]. As illustrated in Figure 2(b), the Raman peaks of the and modes shifted to high frequencies with the increase in Te content. Notably, redshifts and blueshifts in a Raman spectrum reflect tensile and compressive stresses, respectively [34, 35]. Thus, the gradual blueshifts of the and modes indicate the action of compressive stress along the and axes, respectively, which provides another experimental evidence for the presence of vacancies.

(a)

(b)
Figure 3(a) presents the results of and for with respect to . The values of were , , , , , and for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. Here, increased exponentially with the increase in Te content. The sample of exhibited the highest . This can be explained based on the formation of anion vacancies. In particular, an anion vacancy generated by a Te deficiency can act as an electron donor, providing two electron carriers [20]. As shown in the inset of Figure 3(a), the values were 3.21, 7.90, 6.58, 3.06, 3.56, and 2.19 cm2/Vs for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. Thus, significantly increased for and 0.0375; however, it decreased at higher doping of , 0.1, and 0.125. Notably, Te doping in SnSe2 generates defects such as vacancies and Te impurities, which disrupt the transfer of electrons. Thus, it can be anticipated that this abnormal behavior of may have originated from such defects in the SnSe2 crystal structure.

(a)

(b)

(c)

(d)

(e)
The and values were measured along the pressing direction (Figures 3(b) and 3(c)). The values increased with , exhibiting an intrinsic semiconducting behavior (Figure 3(b)). As Te doping increased, initially increased at ; however, it then decreased with a further Te doping. For example, the values at 690 K for , 0.025, 0.0375, 0.075, 0.1, and 0.125 were 19.6, 21.8, 19.6, 15.4, 12.2, and 10.2 S/cm, respectively. For the samples with high doping (, 0.1, and 0.125), the values decreased compared to those for the undoped sample, primarily owing to the significant decrease in . The significant decrease in could be due to presence of vacancy and Te impurities (Figure 1(a) and Table 1). The inset of Figure 3(b) presents logarithmic . At 300 K, the value for increased by approximately 2.5 times compared to that for undoped SnSe2; however, the difference between the two gradually decreased with . Similarly, the samples with and 0.125 presented higher values than SnSe2 at 300 K; however, at a high , they demonstrated lower values than SnSe2.
Figure 3(c) presents the values for the series of samples. All samples had negative values, exhibiting an -type conduction behavior. The magnitude of for undoped SnSe2 decreased over the complete range of measured , which is consistent with our previous results [13]. As Te doping increased, the magnitude of initially increased at , after which it decreased. The values at 300 K for , 0.025, 0.0375, 0.075, 0.1, and 0.125 were −557, −574, −536, −359, −317, and −223 μV/K, respectively, whereas those at 790 K were −307, −342, −331, −319, −286, and −251 μV/K, respectively. Generally, as the carrier concentration increases, a decrease in the magnitude of is expected [36]: where and denote the density-of-state effective mass and carrier concentration, respectively. Therefore, the reductions in for the doped samples (, 0.1, and 0.125) could be primarily attributed to increments in . The increase in for the samples with and 0.0375 can be explained based on the increase in .
Figure 3(d) presents the calculated based on the and values measured at 300 K. The inset of Figure 3(d) presents the calculated value as a function of in (, 0.025, 0.0375, 0.075, 0.1, and 0.125). The values were 1.46, 1.77, 1.58, 0.46, 0.41, and 0.23 for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. Here, the sample with exhibited the largest , primarily owing to the increase in and . However, the value decreased at higher doping levels. Figure 3(e) presents the weighted mobility () 300 K. The values of were 5.89, 18.55, 11.18, 0.48, 0.63, and 0.27 cm2/Vs for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. Note that is proportional to , whereis the nondegenerate mobility. Thus, the general trend followed by is in accordance with the trend followed by the composition-dependent [37].
Figures 4(a) and 4(b) present the temperature dependences of and , respectively. The value was calculated using the Wiedemann-Franz law [38]: , where denotes the Lorenz number. is calculated with the following equation: (where is in 10−8 WΩK−2 and in μV/K) [39]. Then, the was obtained by subtracting from . However, and presented almost identical values, since the electronic contribution to was very small. In particular, the values of the samples were 1.34, 1.40, 1.28, 1.16, 1.19, and 1.20 W/mK at 300 K and 0.51, 0.41, 0.32, 0.46, 0.40, and 0.44 W/mK at 790 K for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. The values of the doped samples were slightly lower than the values of undoped SnSe2.

(a)

(b)
The Debye-Callaway model was used to confirm that the reduction for at is attributed to additional point defect phonon scattering (lines in Figure 4(b); see Supplementary Information for details (available here)) [40]. The slight reductions in values for the doped samples could be attributed to point defect phonon scattering resulting from Te substitutional doping at Se sites and the vacancies generated from Te doping (the atomic masses of Te and Se are 127.60 and 78.96 amu, respectively). A higher led to a lower calculated , owing to intensified point defect phonon scattering (see Supplementary Information for details). Thus, the calculated reduction in values for alloy compositions may be attributed to both phonon scatterings from substitutional (Te at Se sites) and vacancy defects.
The power factor calculated using the values of and is presented in Figure 5(a). The power factors were 0.007, 0.018, 0.015, 0.002, 0.004, and 0.001 mW/mK2 at 300 K and 0.22, 0.25, 0.21, 0.15, 0.10, and 0.07 mW/mK2 at 790 K for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. The power factor at was enhanced the most compared to that of undoped SnSe2; this can be attributed to the increase in and . The power factor at almost presented a general improvement compared to that of the SnSe2 sample. However, the power factors of the other doped samples (, 0.1, and 0.125) decreased because the and values for the samples decreased simultaneously with increasing doping content. Hence, the sample with exhibited the power factor of 0.25 mW/mK2 at 740 K, which demonstrated an improvement of 22% compared to the power factor of pristine SnSe2.

(a)

(b)

(c)

(d)
Figure 5(b) shows the values of the samples. Here, was enhanced for and 0.0375 compared to that for pristine SnSe2 at 790 K; however, its value gradually decreased with the increase in Te content to approximately 0.12 for . The enhanced for and 0.0375 originated primarily from the small decrease in and increase in the power factor, owing to the increase in and . In contrast, at higher doping levels of , the power factors and are decreased. The was slightly decreased for the doped samples. Consequently, the value of 0.49 was obtained for at 790 K. The is decreased beyond compared to SnSe2 sample, due to the significant decrease in (inset of Figure 3(a)).
To further understand the improved electrical transport properties of the Te-doped samples, the temperature dependence of for each sample was inferred based on the and values, as presented in Figure 5(c). Note that is related to the theoretical maximum electronic performance of a thermoelectric material. Thus, can be obtained based on a simple analytic form that approximates the exact Drude-Sommerfeld free-electron model for [37]: where and denote Planck’s constant and electron mass, respectively. is relevant to the maximum power factor when is optimally tuned [41]. In our analysis, the values of were found to be 8.60, 11.69, 9.39, 6.36, 3.44, and 2.16 cm2/Vs at 790 K for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. Thus, the values for and 0.0375 were significantly improved compared to the value for undoped SnSe2; this could be attributed to the increase in the power factors for and 0.0375 compared to that of undoped SnSe2. Thus, by appropriately tuning the carrier concentration at , the power factor can be further increased.
Figure 5(d) presents the dimensionless thermoelectric quality factor () calculated using and . Here, was calculated as [37] follows:
Note that is proportional to the maximum value that can be achieved for an optimized value of [42, 43]. The values of at 790 K were 0.13, 0.22, 0.22, 0.11, 0.07, and 0.04 for , 0.025, 0.0375, 0.075, 0.1, and 0.125, respectively. At 790 K, the value of improved at and 0.0375, after which it gradually decreased with the increase in ; this behavior is in agreement with the trend. Because the gap between the values for the and SnSe2 samples was largest at 790 K, the improvement is expected to be maximized at 790 K if is optimized. Therefore, herein, we successfully optimized for the Sn(Se1.9625Te0.0375)2 () composition via Te doping.
4. Conclusion
A series of Te doping SnSe2, (, 0.025, 0.0375, 0.075, 0.1, and 0.125), polycrystalline samples was synthesized, and the electrical, thermal, and thermoelectric properties were investigated. The Te impurity in Te-doped SnSe2 was identified to create anion vacancies owing to its structural instability upon high doping. The gradual reductions in the lattice parameters and blueshifts of the and vibration modes confirmed the presence of these vacancies. Simultaneous increase in electron concentration and mobility was observed for and 0.0375. As a result, the electrical conductivity and Seebeck coefficient simultaneously increased for and 0.0375. The increase of density-of-state effective mass and the weighted mobility was also seen for and 0.0375. The power factor improved to 0.25 mW/mK2 for the sample with from 0.22 mW/mK2 for pristine SnSe2 at 740 K. In addition, the lattice thermal conductivity slightly is reduced for the doped samples owing to the substitution of Te atoms at Se sites and vacancies generated by Te doping. As a result, the value of 0.49 at 790 K was achieved for the Sn(Se1.9625Te0.0375)2 () sample, which presented a 53% improvement compared to that of undoped SnSe2. The trends followed by weighted mobility and quality factor were compared to those followed by the power factor and to verify the optimization of carrier concentrations for the doped samples.
Data Availability
Data are available on request.
Conflicts of Interest
The authors declare that they have no competing interests.
Authors’ Contributions
Se Woong Lee and Okmin Park contributed equally to this work.
Acknowledgments
This research was supported by the Nano·Material Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (2022M3H4A1A04076667).
Supplementary Materials
The Debye-Callaway model was used to confirm that lattice thermal conductivity reduction for at is attributed to the additional point defect phonon scattering. Parameters used to estimate in the Debye-Callaway model of the samples are provided in Table S1 in the Supplementary Information. Table S1: parameters used to estimate of the samples. (Supplementary Materials)