Open Submitted: 31 January 2015 Accepted: 17 March 2015 Published Online: 30 March 2015
APL Materials 3, 041516 (2015); https://doi.org/10.1063/1.4916526
more...View Affiliations
View Contributors
• Elisabeth Rausch
• Benjamin Balke
• Torben Deschauer
• Siham Ouardi
• Claudia Felser

The carrier concentration in the p-type half-Heusler compound Ti0.3Zr0.35Hf0.35CoSb1−xSnx was optimized, which is a fundamental approach to enhance the performance of thermoelectric materials. The optimum carrier concentration is reached with a substitution level x = 0.15 of Sn, which yields the maximum power factor, 2.69 × 10−3 W m−1 K−2, and the maximum ZT = 0.8. This is an enhancement of about 40% in the power factor and the figure of merit compared to samples with x = 0.2. To achieve low thermal conductivities in half-Heusler compounds, intrinsic phase separation is an important key point. The present work addresses the influence of different preparation procedures on the quality and reproducibility of the samples, leading to the development of a reliable fabrication method.
The efficiency of a thermoelectric material depends on its electronic and lattice properties, summarized in its figure of merit ZT. Desirable are high electrical conductivity σ and Seebeck coefficients S, and low thermal conductivity κ, which consists of a lattice component κlat and an electronic component κel. All these parameters are interdependent; therefore, their adjustment is challenging. One basic approach to enhance the electronic properties and consequently ZT is to optimize the carrier concentration n. Excellent graphical illustrations of the interdependence of those parameters are found in review articles.1,21. G. J. Snyder and E. S. Toberer, Nat. Mater. 7, 105 (2008). https://doi.org/10.1038/nmat20902. A. J. Minnich, M. S. Dresselhaus, Z. F. Ren, and G. Chen, Energy Environ. Sci. 2, 466 (2009). https://doi.org/10.1039/b822664b The maximum of ZT typically occurs at carrier concentrations between 1019 and 1021 carriers × cm−3 (depending on the material system), which is found in heavily doped semiconductors.11. G. J. Snyder and E. S. Toberer, Nat. Mater. 7, 105 (2008). https://doi.org/10.1038/nmat2090 In this work, we focus on the optimization of the charge carrier concentration in the p-type (Ti/Zr/Hf)CoSb1−xSnx half-Heusler compounds.
Half-Heusler compounds with the general formula XYZ (where X and Y are transition metals, and Z is a main group element) crystallizing in the space group $F4̄3m$ have become a focus of research for thermoelectric materials in recent years. Apart from significant improvements in their thermoelectric properties, their advantages are their low toxicity, elemental abundance, and high mechanical stability.3–53. G. B. Jan-Willem and A. D. Ruth, J. Phys.: Condens. Matter 26, 433201 (2014). https://doi.org/10.1088/0953-8984/26/43/4332014. W. J. Xie, A. Weidenkaff, X. F. Tang, Q. J. Zhang, J. Poon, and T. M. Tritt, Nanomaterials 2, 379 (2012). https://doi.org/10.3390/nano20403795. S. Chen and Z. Ren, Mater. Today 16, 387 (2013). https://doi.org/10.1016/j.mattod.2013.09.015
The best n-type half-Heusler materials with ZT values > 1 are found in the (Ti/Zr/Hf)NiSn system. Key to their thermoelectric high efficiency is an intrinsic phase separation66. M. Schwall and B. Balke, Phys. Chem. Chem. Phys. 15, 1868 (2013). https://doi.org/10.1039/C2CP43946H leading to a submicrostructuring, which effectively reduces the lattice thermal conductivity. This concept is well-established for the n-type half-Heusler compounds7–97. M. Schwall, B. Balke, and M. Köhne, U.S. patent application 20140127070 (08 May 2014).8. S. Sakurada, N. Shutoh, and S. Hirono, U.S. patent application US2005/0217715 A1 (06 October 2005).9. N. Shutoh, S. Sakurada, N. Kondo, and N. Takezawa, U.S. patent application US20050172994 A1 (11 August 2005). and the intrinsic phase separation is even retained under thermal cycling conditions.1010. J. Krez, B. Balke, C. Felser, W. Hermes, and M. Schwind, “Long-term stability of phase-separated Half-Heusler compounds” (2015); e-print arXiv:1502.01828 [cond-mat.mtrl-sci].
In contrast, state-of-the-art p-type materials are found in the MCoSb0.8Sn0.2 (M = Ti, Zr, Hf) system. A breakthrough was achieved by a nanostructuring approach via ball milling of Zr0.5Hf0.5CoSb0.8Sn0.2.1111. X. Yan, G. Joshi, W. Liu, Y. Lan, H. Wang, S. Lee, J. W. Simonson, S. J. Poon, T. M. Tritt, G. Chen, and Z. F. Ren, Nano Lett. 11, 556 (2011). https://doi.org/10.1021/nl104138t All further enhancements in this material system were achieved by variation of the composition M.12,1312. X. Yan, W. S. Liu, H. Wang, S. Chen, J. Shiomi, K. Esfarjani, H. Z. Wang, D. Z. Wang, G. Chen, and Z. F. Ren, Energy Environ. Sci. 5, 7543 (2012). https://doi.org/10.1039/c2ee21554c13. X. Yan, W. Liu, S. Chen, H. Wang, Q. Zhang, G. Chen, and Z. F. Ren, Adv. Energy Mater. 3, 1195 (2013). https://doi.org/10.1002/aenm.201200973 Very recently, we transferred the well-established concept of intrinsic phase separation to the p-type MCoSb0.8Sn0.2 system.1414. E. Rausch, B. Balke, S. Ouardi, and C. Felser, Phys. Chem. Chem. Phys. 16, 25258 (2014). https://doi.org/10.1039/C4CP02561J Nevertheless, only a few systematic investigations of the optimum doping level in the MCoSb1−xSnx system have been reported in the literature. In particular, samples with 15% Sn substitution are rare.15–1715. T. Sekimoto, K. Kurosaki, H. Muta, and S. Yamanaka, Jpn. J. Appl. Phys., Part 2 46, L673 (2007). https://doi.org/10.1143/JJAP.46.L67316. V. Ponnambalam, P. N. Alboni, J. Edwards, T. M. Tritt, S. R. Culp, and S. J. Poon, J. Appl. Phys. 103, 063716 (2008). https://doi.org/10.1063/1.289659117. P. Maji, J. P. A. Makongo, X. Zhou, H. Chi, C. Uher, and P. F. P. Poudeu, J. Solid State Chem. 202, 70 (2013). https://doi.org/10.1016/j.jssc.2013.03.024 A common statement is that the most effective p-doping for MCoSb is a substitution of 20%–30% Sn for Sb.55. S. Chen and Z. Ren, Mater. Today 16, 387 (2013). https://doi.org/10.1016/j.mattod.2013.09.015
In this work, first, we verify the reproducibility of the structural and thermoelectric properties. We investigate the effect of different synthesis methods for Ti0.3Zr0.35Hf0.35CoSb0.8Sn0.2 as an example and ensure the repeatability by having two different persons perform the experiments independently of each other. Subsequently, we optimize the thermoelectric properties of the system by adjusting the carrier concentration within the series Ti0.3Zr0.35Hf0.35CoSb1−xSnx with x = 0.1, 0.15, 0.2, 0.3, 0.4, and 0.5 using partial substitution of Sb by Sn. This wide range of substitution was chosen to demonstrate the dependency of the thermoelectric properties on the carrier concentration (compare Figure 4).
For the standard arc melting procedure, stoichiometric amounts of the elements were weighted in and melted three times under an Ar atmosphere (10−4 millibars) with a current just large enough to provide rapid melting of the elements (approximately 50–75 A). The ingots were flipped and remelted three times on each side. Next, they were crushed, and the weight loss due to the evaporation of Sb during the arc melting process was compensated by adding Sb. The melting steps were then performed two more times with intermediate crushing and addition of Sb.
For the reproducibility study, we also prepared a sample using the same procedure but without Sb addition and another one in which an averaged amount of Sb was added to analyze the sensitivity of the thermoelectric properties to the correct stoichiometry. The experiments were performed by two different experimenters (persons A and B) independently of one another.
For comparison, ingots were prepared by inductive melting in an argon atmosphere at a low pressure of approximately 1.0 millibar. Special care was taken to prevent the evaporation of Sb. We therefore used a specially designed lidded boron nitride (BN) crucible coated with Y 2O3 and controlled the temperature carefully with a pyrometer. The temperature can be regulated by controlling the power output of the high-frequency generator. The best results were obtained using a quick melting process (within 3 min) and a power of 1.0–4.3 kW and then holding the temperature close to the melting point at 1400–1500 °C for 30 min.
Ingots for the carrier concentration optimization of Ti0.3Zr0.35Hf0.35CoSbxSn1−x (x = 0.1 – 0.5) were all prepared by the standard arc melting procedure.
Subsequently, all the samples were annealed in quartz ampules under vacuum at 900 °C for 7 days and then quenched in ice water.
The crystal structure of all the samples was analyzed by X-ray powder diffraction (XPD) with Mo Kα radiation (Bruker AXS D8 Advance in reflection mode). XPD with synchrotron radiation (λ = 1.653 07 Å) was performed at bending magnet D10 of the XPD beamline at the Brazilian Synchrotron Light Laboratory. Detailed beamline characteristics are given by Ferreira et al.1818. F. F. Ferreira, E. Granado, W. Carvalho, Jr., S. W. Kycia, D. Bruno, and R. Droppa, Jr., J. Synchrotron Radiat. 13, 46 (2006). https://doi.org/10.1107/S0909049505039208 Scanning electron microscopy (SEM), in combination with energy-dispersive X-ray (EDX) detection (FEI Nova nanoSEM 630), was applied to determine the composition of the phases and display the microstructure of the samples.
The ingots were cut into discs and bars for characterization of the thermoelectric properties. For the reproducibility studies, the transport properties were determined using the thermal transport option of a physical properties measurement system (PPMS, Quantum Design, San Diego, CA). At high temperatures (100–600 °C), the electrical conductivity σ and Seebeck coefficient S were measured simultaneously by an LSR 3 (Linseis) in He atmosphere. The thermal conductivity κ was calculated from diffusivity measurements (LFA 457, Netzsch) as κ = αρCp. The density ρ was determined from the mass and volume of the cut bars, and the heat capacity Cp was estimated by the Dulong–Petit law. The figure of merit ZT was calculated according to
 $ZT=S2σκT.$ (1)
The samples’ carrier concentration n was determined at room temperature by Hall measurements using the IPM-HT-Hall-900K system developed and constructed by Fraunhofer IPM in Freiburg, Germany.
The reproducibility of the sample preparation for p-type (Ti/Zr/Hf)CoSb0.8Sn0.2 was investigated by employing two synthesis methods, arc melting and inductive melting. Three different approaches were applied to handle the weight loss due to the high vapor pressure of Sb during the arc melting process: In the first, the missing amount was added (samples AM-A1, AM-A2, and AM-B2); in the second, the weight loss was not compensated (AM-A3), and in the third, an average value of Sb obtained from the previous experiments was added (AM-B2). For comparison, samples were fabricated by inductive melting, one with the correct nominal composition (IM-B1) and a second with 5 at. % less Sb, Ti0.3Zr0.35Hf0.35CoSb0.75Sn0.2 (IM-B2). Two different persons (A and B) performed the experiments to eliminate the impact of the experimenter on the results.
The XPD patterns obtained using a Mo Kα lab source show very little difference between samples fabricated by the two methods. Figure 1(a) compares arc melted samples with and without the addition of Sb to samples fabricated by inductive melting. The reflections of samples with an Sb deficiency are shifted to slightly higher scattering angles 2θ compared to those of samples with the correct stoichiometry. In samples prepared by inductive melting, β-Sn impurities were detectable. Nevertheless, the thermoelectric properties should be very sensitive to the correct stoichiometry. Evaporation of Sb produces unoccupied lattice sites, which should increase the resistivity. At the same time, the Seebeck coefficient is increased. This agrees well with our measurements using the PPMS (compare Table I). The choice of production method has a very obvious effect on the figure of merit ZT (see Figure 1(b)). Samples produced by arc melting with the addition of the missing amount of Sb all exhibit the same values regardless of which person fabricated them. Further, the sample with the addition of an average amount of Sb shows the same characteristics. An off-stoichiometric amount of Sb—either produced on purpose by inductive melting or by not adding Sb during the arc melting process—yielded the lowest ZT values. The sample with the correct stoichiometry produced by inductive melting exhibits ZT values between those two extremes. This could be the result of Sn impurities, which are metallic. Further, the bars for the measurements were cut from the middle of the sample, and owing to the long melting time in the induction furnace, the Sb could recrystallize on the surface of the sample, causing an Sb deficiency in the middle. Additionally, those ingots contained a certain amount of voids, which made cutting very difficult. Nevertheless, EDX analysis confirmed the homogeneity of the samples. From these results, it is obvious that to obtain homogeneous and dense ingots, the quick arc melting process in combination with crushing of the samples between the individual melting steps and the addition of evaporated Sb must be used. By following this procedure, reproducible thermoelectric properties can be obtained.
TABLE I. Reproducibility of Ti0.3Zr0.35Hf0.35CoSb0.8Sn0.2 using different fabrication methods. Listed are sample code, fabrication procedure, and thermoelectric properties at 300 K.
Sample codeFabrication methodCompensation for weight lossPersonκ(W K−1m−1)S(μV K−1)σ(10−4 S m−1)S2σ(10−3 W m−1K−1)ZT
AM-B2Arc meltingaverage mass of Sb addedB2.68139.46.361.230.14
IM-B1Inductive meltingcorrect stoichiometryB2.89174.22.670.810.09
IM-B2Inductive melting5 at. % less SbB2.56120.61.670.240.03
Phase analysis by XPD using a Mo Kα lab source confirmed the C1b structure. All samples show a broadening of the reflections due to the presence of two or sometimes even three half-Heusler phases. Samples with x ≥ 0.3 contain a small amount of impurities (compare Table II). As previously discussed,1919. E. Rausch, S. Ouardi, U. Burkhardt, C. Felser, J. M. Stahlhofen, and B. Balke, “High thermoelectric figure of merit in p-type Half-Heuslers by intrinsic phase separation” (2015); e-print arXiv:1502.03336 [cond-mat.mtrl-sci]. the resolution of an XPD lab machine gives a hint of the presence of different half-Heusler phases, but it is not sufficient to clearly distinguish them. As an example, we investigated the sample with x = 0.15 using synchrotron radiation (see Figure 2). A very small amount of β-Sn was detected, and the characteristic splitting of the reflections by phase separation into half-Heusler phases with very similar lattice parameters was observed.
TABLE II. EDX analysis of the Ti0.3Zr0.35Hf0.35CoSb1−xSnx samples. Indicated are overall composition, composition of the matrix (HH1), additional half-Heusler (HH2, HH3) phases, and observed inclusions of impurity phases.
TiZrHfCoSbSn
x = 0.1Overall0.270.330.430.970.920.09
HH10.210.360.480.960.920.06
HH20.450.280.310.990.860.11
ImpuritiesHf
x = 0.15Overall0.310.340.390.940.880.15
HH10.190.390.450.930.950.08
HH20.650.170.151.100.350.59
ImpuritiesSn
x = 0.2Overall0.300.340.390.940.830.21
HH10.180.400.450.930.950.09
HH20.600.150.241.090.290.65
HH30.170.280.581.000.950.04
ImpuritiesSn, Hf
x = 0.3Overall0.270.360.390.930.780.28
HH10.270.360.410.930.810.24
Impurities(Zr,Hf)Sn2, (Zr,Hf)2Sn3
x = 0.4Overall0.290.310.440.980.590.38
HH10.160.380.520.980.810.16
HH20.410.170.381.160.260.62
Impurities(Ti,Zr)Sn, (Zr,Hf)2Sn3
x = 0.5Overall0.380.350.280.840.250.90
HH10.380.260.351.050.360.60
HH20.450.180.301.120.230.71
Impurities(Ti,Zr)3Sn2, ZrSn2
EDX analysis confirmed the correct overall stoichiometry of all the compounds (see Table II). Secondary electron images of samples with x = 0.1, 0.15, and 0.2 show a microstructure consisting of an Sb- and Zr/Hf-rich matrix (HH1) and an additional half-Heusler phase rich in Ti and Sn (HH2) compared to the nominal composition. The HH2 phase typically exhibits a slight excess of Co, which presumably occupies the vacant lattice positions like in the related L21 structure Co2TiSn.19–2119. E. Rausch, S. Ouardi, U. Burkhardt, C. Felser, J. M. Stahlhofen, and B. Balke, “High thermoelectric figure of merit in p-type Half-Heuslers by intrinsic phase separation” (2015); e-print arXiv:1502.03336 [cond-mat.mtrl-sci].20. H. C. Kandpal, V. Ksenofontov, M. Wojcik, R. Seshadri, and C. Felser, J. Phys. D: Appl. Phys. 40, 1587 (2007). https://doi.org/10.1088/0022-3727/40/6/S1321. B. Balke, S. Ouardi, T. Graf, J. Barth, C. G. F. Blum, G. H. Fecher, A. Shkabko, A. Weidenkaff, and C. Felser, Solid State Commun. 150, 529 (2010). https://doi.org/10.1016/j.ssc.2009.10.044 In the sample with x = 0.2, three different half-Heusler phases can be identified. The typical microstructure is displayed in Figure 2(c). As impurity phases, we detected very small inclusions of Sn and Hf. Samples with x ≥ 0.3 exhibit additional binary (Ti/Zr/Hf)xSny phases. This explains the additional reflection at 2θ = 15.75° in the corresponding XPD patterns. The network of the second half-Heusler phase is formed around these islands (compare Figure 2(d)). In summary, our structural investigations proved that all the samples underwent an intrinsic phase separation, but to achieve a phase separation into solely half-Heusler phases, the substitution level of Sn cannot exceed 20%. More Sn leads to the inclusion of additional binary phases on a micrometer scale.
Substitution of Sn for Sb in p-type Ti0.3Zr0.35Hf0.35CoSb1−xSnx introduces holes in the system; therefore, the concentration of the main charge carriers is increased. Even though EDX analysis revealed small amounts of binary impurity phases for samples with x ≥ 0.3, the carrier concentration increases systematically with increasing substitution level x as shown in Figure 4(a). However, these do not affect the net charge carrier concentration. For comparison, we calculated the nominal carrier concentration for each composition assuming that the doping of one Sn atom introduces one hole to the system starting from the experimentally determined n of the sample with x = 0.1. Since the experimentally estimated carrier concentration is in good agreement with the calculation, it reflects the intrinsic electronic state of the main phase.
As one expects from the increased carrier concentration, the electrical conductivity σ increases with increasing substitution level x (see Figure 3). The temperature dependence shows a decrease of σ with increasing temperature, indicating metal-like behavior of samples with x = 0.1–0.2. The sample with x = 0.3 shows a maximum around 200 °C. Whereas the sample with x = 0.5 shows a semiconducting behavior. Samples with x ≥ 0.3 contain additional binary impurity phases (compare Table II) indicating that a maximum substitution level is reached with 20% Sn. The presence of these inclusions reduces the mobility by scattering of the main charge carriers leading to different electronic properties as in the samples with solely half-Heusler phases.
The Seebeck coefficients (see Figure 3(b)) are highest for x = 0.1 and decrease with increasing Sn content, as one can expect from the interrelation of the electronic transport properties.
The thermal conductivity increases with increasing Sn content (see Figure 3(c)), and the temperature dependence is also effected. Samples with x = 0.1 − 0.2 show a decrease with increasing temperature, whereas samples with x = 0.3 show nearly temperature-independent behavior, and the thermal conductivity of samples with x = 0.4 and 0.5 increases. To explain this behavior, we calculated the lattice thermal conductivity κlat by subtracting the electronic contribution κel = LσT estimated by assuming the Wiedemann–Franz law from the total thermal conductivity. The Lorentz number L was approximated by
 $L=1.5+exp−|S|116×10−8 W Ω K−2,$ (2)
with the Seebeck coefficient S in μV K−1, as recently proposed by Kim et al.2222. H.-S. Kim, Z. M. Gibbs, Y. Tang, H. Wang, and G. J. Snyder, APL Mater. 3, 041506 (2015). https://doi.org/10.1063/1.4908244 as a much better approximation than the constant value L = 2.44 × 10−8 W Ω K−2. All samples showed a decrease in κlat with increasing temperature, suggesting that they are in the range where the lattice thermal conductivity is dominated by the phonon-phonon scattering. The temperature dependence of κlat for samples with x ≥ 0.3 is much weaker than for samples with x ≤ 0.2, leading to a nearly glass-like behavior. Nevertheless, the increase in the total thermal conductivity κ is mainly due to the increasing electronic contribution for x ≥ 0.3.
To visualize the effect of the substitution level x, we plotted the thermoelectric properties at 610 °C versus the carrier concentration n in Figure 4(b). The figure illustrates the interdependence of the thermoelectric properties in the system Ti0.3Zr0.35Hf0.35CoSb1−xSnx via the carrier concentration, like the figures found in fundamental review articles on thermoelectric materials.1,21. G. J. Snyder and E. S. Toberer, Nat. Mater. 7, 105 (2008). https://doi.org/10.1038/nmat20902. A. J. Minnich, M. S. Dresselhaus, Z. F. Ren, and G. Chen, Energy Environ. Sci. 2, 466 (2009). https://doi.org/10.1039/b822664b It proves that the best compromise for high power factors is reached at x = 0.15, which yields the maximum figure of merit ZT = 0.79. This is an enhancement of 40% compared to 0.57 for x = 0.1 or x = 0.2. We improved the thermoelectric properties significantly simply by filling the gap between those two substitution levels. The conformity of the experimental data with the theoretical predictions is picture-book like, suggesting that unlike the more complicated MNiSn system,2323. J. Krez, J. Schmitt, G. J. Snyder, C. Felser, W. Hermes, and M. Schwind, J. Mater. Chem. A 2, 13513 (2014). https://doi.org/10.1039/C4TA03000A a single parabolic band-model can be assumed to tune the carrier concentration. It is a powerful tool to further enhance the thermoelectric performance in this system with a given composition M.
To summarize, we investigated the influence of different fabrication procedures on the thermoelectric performance of p-type half-Heusler compounds. The detailed study was performed on Ti0.3Zr0.35Hf0.35CoSb0.8Sn0.2 proving that our standard arc melting fabrication delivers reproducible results. The average of the standard deviation of the ZT values measured at temperatures of 100–380 K is only 6.5%, which lies in the range of the measurement errors; therefore, the results can be considered accurate.
This standard fabrication process was then applied to synthesize a substitution series of Ti0.3Zr0.35Hf0.35CoSb1−xSnx for carrier concentration optimization. Within this system, the maximum power factor of 2.69 × 10−3 W m−1 K−2 and maximum figure of merit ZT of 0.79 are reached at x = 0.15, which correspond to a carrier concentration of 1.4 × 1021 cm−3. To verify the intrinsic phase separation, this sample was analyzed in detail by SEM, EDX analysis, and high resolution XPD with synchrotron radiation. By adjusting the carrier concentration, we achieved an enhancement of about 40% in the power factor and figure of merit ZT compared to samples with x = 0.1 or 0.2 simply by choosing a substitution level between those of these two samples. Thus, we suggest a Sn content of 15% as a starting point for subsequent enhancements by varying the composition M, instead of using the previous Zr0.5Hf0.5CoSb0.8Sn0.2, which has a ZT of 0.5 at 1000 K.2424. S. R. Culp, J. W. Simonson, S. J. Poon, V. Ponnambalam, J. Edwards, and T. M. Tritt, Appl. Phys. Lett. 93, 022105 (2008). https://doi.org/10.1063/1.2959103 This further fine-tuning of the Ti-to-Hf ratio for optimal phonon scattering has already demonstrated the effectiveness of this approach, reaching a record ZT for p-type half-Heusler compounds of 1.15 at 710 °C.1919. E. Rausch, S. Ouardi, U. Burkhardt, C. Felser, J. M. Stahlhofen, and B. Balke, “High thermoelectric figure of merit in p-type Half-Heuslers by intrinsic phase separation” (2015); e-print arXiv:1502.03336 [cond-mat.mtrl-sci]. In conclusion, we demonstrated that the microstructuring by an intrinsic phase separation in the (Ti/Zr/Hf)CoSb1−xSnx system can be reproduced reliably by a conventional arc melting sample preparation procedure. We optimized the carrier concentration in this system which is reached by a substitution level of 15% Sn for Sb. This leads to enhanced thermoelectric properties compared to the previous state-of-the-art p-type half-Heusler compounds with a substitution level of 20% Sn. The experimentally very simple approach of plotting the transport properties against the carrier concentration is a powerful tool to further enhance the thermoelectric performance of MCoSb1−xSnx in combination with a optimum composition M.
The authors gratefully acknowledge the financial support of the DFG Priority Program 1386 Nanostructured Thermoelectric Materials under proposal BA 4171/2-2 and the thermoHEUSLER Project (Project No. 0327876D) of the German Federal Ministry of Economics and Technology (BMWi). Further support of this work was provided by the Brazilian Synchrotron Light Laboratory (LNLS) under proposal XPD - 17015 and by the DAAD (Project No. 57060637). The authors thank the staff of the LNLS (Campinas) for support as well as Dean H. Barrett (LNLS, Campinas) for help with the XPD experiments. The authors give special thanks to Dr. Michael Schwall and the workshop of the Institute for Inorganic and Analytical Chemistry of the Johannes Gutenberg-University Mainz for the design and setup of the induction furnace.
1. 1. G. J. Snyder and E. S. Toberer, Nat. Mater. 7, 105 (2008). https://doi.org/10.1038/nmat2090, Google ScholarCrossref, ISI
2. 2. A. J. Minnich, M. S. Dresselhaus, Z. F. Ren, and G. Chen, Energy Environ. Sci. 2, 466 (2009). https://doi.org/10.1039/b822664b, Google ScholarCrossref, ISI
3. 3. G. B. Jan-Willem and A. D. Ruth, J. Phys.: Condens. Matter 26, 433201 (2014). https://doi.org/10.1088/0953-8984/26/43/433201, Google ScholarCrossref, ISI
4. 4. W. J. Xie, A. Weidenkaff, X. F. Tang, Q. J. Zhang, J. Poon, and T. M. Tritt, Nanomaterials 2, 379 (2012). https://doi.org/10.3390/nano2040379, Google ScholarCrossref, ISI
5. 5. S. Chen and Z. Ren, Mater. Today 16, 387 (2013). https://doi.org/10.1016/j.mattod.2013.09.015, Google ScholarCrossref, ISI
6. 6. M. Schwall and B. Balke, Phys. Chem. Chem. Phys. 15, 1868 (2013). https://doi.org/10.1039/C2CP43946H, Google ScholarCrossref, ISI
7. 7. M. Schwall, B. Balke, and M. Köhne, U.S. patent application 20140127070 (08 May 2014). Google Scholar
8. 8. S. Sakurada, N. Shutoh, and S. Hirono, U.S. patent application US2005/0217715 A1 (06 October 2005). Google Scholar
9. 9. N. Shutoh, S. Sakurada, N. Kondo, and N. Takezawa, U.S. patent application US20050172994 A1 (11 August 2005). Google Scholar
10. 10. J. Krez, B. Balke, C. Felser, W. Hermes, and M. Schwind, “Long-term stability of phase-separated Half-Heusler compounds” (2015); e-print arXiv:1502.01828 [cond-mat.mtrl-sci]. Google Scholar
11. 11. X. Yan, G. Joshi, W. Liu, Y. Lan, H. Wang, S. Lee, J. W. Simonson, S. J. Poon, T. M. Tritt, G. Chen, and Z. F. Ren, Nano Lett. 11, 556 (2011). https://doi.org/10.1021/nl104138t, Google ScholarCrossref, ISI
12. 12. X. Yan, W. S. Liu, H. Wang, S. Chen, J. Shiomi, K. Esfarjani, H. Z. Wang, D. Z. Wang, G. Chen, and Z. F. Ren, Energy Environ. Sci. 5, 7543 (2012). https://doi.org/10.1039/c2ee21554c, Google ScholarCrossref, ISI
13. 13. X. Yan, W. Liu, S. Chen, H. Wang, Q. Zhang, G. Chen, and Z. F. Ren, Adv. Energy Mater. 3, 1195 (2013). https://doi.org/10.1002/aenm.201200973, Google ScholarCrossref, ISI
14. 14. E. Rausch, B. Balke, S. Ouardi, and C. Felser, Phys. Chem. Chem. Phys. 16, 25258 (2014). https://doi.org/10.1039/C4CP02561J, Google ScholarCrossref, ISI
15. 15. T. Sekimoto, K. Kurosaki, H. Muta, and S. Yamanaka, Jpn. J. Appl. Phys., Part 2 46, L673 (2007). https://doi.org/10.1143/JJAP.46.L673, Google ScholarCrossref, ISI
16. 16. V. Ponnambalam, P. N. Alboni, J. Edwards, T. M. Tritt, S. R. Culp, and S. J. Poon, J. Appl. Phys. 103, 063716 (2008). https://doi.org/10.1063/1.2896591, Google ScholarScitation, ISI
17. 17. P. Maji, J. P. A. Makongo, X. Zhou, H. Chi, C. Uher, and P. F. P. Poudeu, J. Solid State Chem. 202, 70 (2013). https://doi.org/10.1016/j.jssc.2013.03.024, Google ScholarCrossref, ISI
18. 18. F. F. Ferreira, E. Granado, W. Carvalho, Jr., S. W. Kycia, D. Bruno, and R. Droppa, Jr., J. Synchrotron Radiat. 13, 46 (2006). https://doi.org/10.1107/S0909049505039208, Google ScholarCrossref, ISI
19. 19. E. Rausch, S. Ouardi, U. Burkhardt, C. Felser, J. M. Stahlhofen, and B. Balke, “High thermoelectric figure of merit in p-type Half-Heuslers by intrinsic phase separation” (2015); e-print arXiv:1502.03336 [cond-mat.mtrl-sci]. Google Scholar
20. 20. H. C. Kandpal, V. Ksenofontov, M. Wojcik, R. Seshadri, and C. Felser, J. Phys. D: Appl. Phys. 40, 1587 (2007). https://doi.org/10.1088/0022-3727/40/6/S13, Google ScholarCrossref, ISI
21. 21. B. Balke, S. Ouardi, T. Graf, J. Barth, C. G. F. Blum, G. H. Fecher, A. Shkabko, A. Weidenkaff, and C. Felser, Solid State Commun. 150, 529 (2010). https://doi.org/10.1016/j.ssc.2009.10.044, Google ScholarCrossref, ISI
22. 22. H.-S. Kim, Z. M. Gibbs, Y. Tang, H. Wang, and G. J. Snyder, APL Mater. 3, 041506 (2015). https://doi.org/10.1063/1.4908244, Google ScholarScitation, ISI
23. 23. J. Krez, J. Schmitt, G. J. Snyder, C. Felser, W. Hermes, and M. Schwind, J. Mater. Chem. A 2, 13513 (2014). https://doi.org/10.1039/C4TA03000A, Google ScholarCrossref, ISI
24. 24. S. R. Culp, J. W. Simonson, S. J. Poon, V. Ponnambalam, J. Edwards, and T. M. Tritt, Appl. Phys. Lett. 93, 022105 (2008). https://doi.org/10.1063/1.2959103, Google ScholarScitation, ISI