Ag doped type-III Ba24Ge100 clathrates

Author’s Accepted Manuscript Thermoelectric properties of Cu/Ag doped type-III Ba24Ge100 clathrates Jiefei Fu, Xianli Su, Yonggao Yan, Wei Liu, Zhengkai Zhang, Xiaoyu She, Ctirad Uher, Xinfeng Tang www.elsevier.com/locate/yjssc

PII: DOI: Reference:

S0022-4596(17)30246-3 http://dx.doi.org/10.1016/j.jssc.2017.06.025 YJSSC19843

To appear in: Journal of Solid State Chemistry Received date: 8 May 2017 Revised date: 17 June 2017 Accepted date: 23 June 2017 Cite this article as: Jiefei Fu, Xianli Su, Yonggao Yan, Wei Liu, Zhengkai Zhang, Xiaoyu She, Ctirad Uher and Xinfeng Tang, Thermoelectric properties of Cu/Ag doped type-III Ba24Ge100 clathrates, Journal of Solid State Chemistry, http://dx.doi.org/10.1016/j.jssc.2017.06.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Thermoelectric properties of Cu/Ag doped type-III Ba24Ge100 clathrates Jiefei Fu,a Xianli Su,a* Yonggao Yan,a Wei Liu,a Zhengkai Zhang,a Xiaoyu She,a Ctirad Uher,b Xinfeng Tanga* a

State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China b

Department of Physics, University of Michigan, Ann Arbor, MI 48109, USA *Email:[email protected] or [email protected]

Abstract: Type-III Ba24Ge100 clathrates possess low thermal conductivity and high electrical conductivity at room temperature and, as such, have a great potential as thermoelectric materials for power generation. However, the Seebeck coefficient is very low due to the intrinsically high carrier concentration. In this paper, a series of Ba24CuxGe100-x and Ba24AgyGe100-y specimens were prepared by vacuum melting combined with the subsequent spark plasma sintering (SPS) process. Doping Cu or Ag on the Ge site not only suppresses the concentration of electrons but it also decreases the thermal conductivity. In addition, the carrier mobility and the Seebeck coefficient increase due to the decrease in the carrier concentration. Thus, the power factor is greatly improved, leading to an improvement in the dimensionless figure of merit ZT. Cu-doped Ba24Cu6Ge94 reaches the maximum ZT value of about 0.17 at 873 K, while Ag-doped Ba24Ag6Ge94 attains the dimensionless figure of merit ZT of 0.31 at 873 K, more than 2 times higher value compared to un-doped Ba24Ge100.

Graphical Abstract

1

Key words: Ba24Ge100, Cu, Ag doping, carrier concentration, thermoelectric properties Introduction With the increasing concerns regarding rapidly depleting fossil fuels, there is strong demand for a more efficient use of energy and for improving energy conversion processes. Thermoelectric (TE) conversion technology, which converts the waste industrial heat into electricity or provides spot cooling via the Peltier effect, has attracted worldwide attention as a reliable and environmentally friendly technology [1]. The conversion efficiency of thermoelectric materials is evaluated by the dimensionless thermoelectric figure of merit ZT=2T/, where , , and T are the Seebeck coefficient, the electrical conductivity, the thermal conductivity and the absolute temperature, respectively [2]. Due to strong inter-coupling of the three transport parameters (,

and ), it is a challenge to optimize ZT. The electronic properties (often judged by the power factor PF=2 are closely related to the carrier concentration. Hence, optimizing the carrier concentration n is crucial to guarantee a reasonably high power factor [3-6]. Clathrate compounds are often considered to behave as quintessential “phonon glass-electron crystal (PGEC)” materials [7-11], having polyhedral cages and guest atoms loosely bonded within the cages. Polyhedral cages consist of group III and IV elements, while alkali metal and alkaline earth metal atoms occupy the cages as guest atoms. According to a specific structure of the cage, the clathrate compounds can be divided into type-I [12-19], type-II [20-24], type-III [25-38] and type-VIII [39-42] clathrates. Type-III Ba24Ge100 clathrate compounds are composed of three kinds of cage structures, namely a pentagonal dodecahedron, an open dodecahedron, and a distorted cube [25, 26, 30, 33] constructed by threefold bonded (3b) and fourfold bonded (4b) Ge atoms, as shown in Figure 1. Having a complex crystal structure, the thermal conductivity of these

2

compounds is intrinsically quite low with a room temperature value of 2.8 W/mK [27]. According to

the

Zintl

concept,

the

chemical

structure

of

Ba24Ge100 can

be

written

as

[Ba2+]24[(3b)Ge-]32[(4b)Ge0]68[16e-] [31], i.e., sixteen conduction electrons are expected to contribute to transport in Ba24Ge100. As a result, the type-III Ba24Ge100 clathrate compounds behave as an n-type semiconductor with the carrier concentration of approximately 5×1021 cm-3 at room temperature. Compared to other clathrates, the type-III Ba24Ge100 clathrates possess a high electrical conductivity (about 2×105 Sm-1) at room temperature [31]. The intrinsically low thermal conductivity and high electrical conductivity make these clathrates a potential high temperature thermoelectric material. However, the Seebeck coefficient of Ba24Ge100 is relatively low, leading to a relatively low power factor. To improve the prospect of type-III Ba24Ge100 clathrates as thermoelectric materials, it is essential to increase its Seebeck coefficient. Since the Seebeck coefficient is inversely proportional to the carrier concentration, this means that one must reduce the carrier concentration and approach closely its optimal value. Several investigations have shown that by substituting group IIIA elements, such as Al, Ga, and In, on the Ge site can effectively reduce the carrier concentration and increase the Seebeck coefficient [27-29]. The solid-solubility limit of Ga in Ba24GaxGe100-x compounds is about 15%. Moreover a very high ZT value of 1.25 is obtained at 943 K for Ba24Ga15Ge85 compound due to the reduction of carrier concentration and thermal conductivity, which is the highest ZT value ever reported for type-III clathrate compounds [27]. However, after that, very little work has been done on thermoelectric properties of type-III clathrate. Since elements such as Cu and Ag from group IB have two fewer electrons in the valence shell than elements from group IIIA, there is a chance that substituting them for Ge might lead to a reduction in the carrier concentration. In addition, so far the correlation between the structure and thermoelectric properties of group IB (Ag or Cu) doped type-III clathrates is still unrevealed. In this study, type-III Ba24MxGe100-x clathrates doped with group IB elements (M = Cu or Ag) are synthesized by vacuum melting combined with a subsequent spark plasma sintering (SPS) process and their thermoelectric properties are studied. Due to a large difference in the valence electron count of group IB elements (Cu, Ag) and of Ge, the carrier concentration is suppressed dramatically. Consequently, the carrier mobility and the Seebeck coefficient have increased, and the power factor has been greatly enhanced, leading to an improvement in the dimensionless figure of merit ZT. Thus, Cu-doped Ba24Cu6Ge94 has reached the maximum ZT value of about 0.17 at 873 K, while Ag-doped Ba24Ag6Ge94 has attained the ZT of 0.31 at 873 K, the value a factor of two higher in comparison to the un-doped Ba24Ge100.

3

Figure 1. Structure of Ba24Ge100 clathrates. Green atoms are Ge atoms; blue, red and pink atoms refer to the three crystallographically different Ba atoms. Experimental High-purity Ba (2N), Cu (5N), Ag wire (4N) and Ge bulk (4N) were weighed in the glove-box according to the stoichiometry of Ba24MxGe100-x (M=Cu, Ag; x=0, 2, 4, 6). Carbon crucibles with raw materials were placed into silica tubes that were vacuum-sealed (~10-3 torr) and loaded into a vertical furnace. The ampoules were heated to 1273 K in 5 h, soaked at this temperature for 10 h, and then furnace cooled. The obtained ingots were subsequently hand ground into fine powders and densified by SPS at 923 K under 35 MPa for 10 min using 15 mm diameter graphite die under vacuum. The relative density of all resulting bulk samples are above 95%. The phase composition of the disk-shaped pellets was characterized by powder X-ray diffraction (XRD, Cu KX’Pert PRO-PANalytical). Microstructures of bulk samples were investigated by field-emission scanning electron microscopy (FESEM, Hitachi SU-8020) with energy-dispersive X-ray spectroscopy (EDS). The actual composition and mapping images of all samples were obtained using electron probe microanalysis (EPMA, JXA-8230, JEOL) equipped with wave dispersive x-ray spectroscopy (WDS). High temperature electrical conductivity  and the Seebeck coefficient  (300 K - 973 K) were measured simultaneously by a standard four-probe method in a low pressure He atmosphere (ZEM-3, ULVAC-RIKO). The Hall coefficient RH was measured using the Physical Properties Measurements System (PPMS-9, Quantum Design). The carrier concentration n and the Hall mobility H were calculated using the equations n = 1/e|RH| 4

and H = |RH|, respectively. Thermal conductivities were calculated using the equation = DCp, where D is the thermal diffusivity obtained by the laser flash method (LFA 457, Netzsch), Cp is the specific heat measured by a differential scanning calorimeter (DSC Q20, TA Instrument), and  is the density measured by the Archimedes’ method. Results and Discussion Figures 2(a) and 2(b) show powder X-ray diffraction (XRD) patterns for Ba24CuxGe100-x and Ba24AgyGe100-y clathrate compounds after SPS, respectively. Their single phase type-III Ba24Ge100 structure after the SPS process is indexed by JCPDS 98-041-4380. However, type-I clathrate structure is also detected after doping with Cu or Ag due to the strain in the crystal lattice induced by the radius and charge difference between atoms of Ge and the group IB elements (Cu and Ag). When the doping amount is 2%, the type-I phase appears, and type-I phase is increased with the increase of doping amount. Compared with the Ga doping in the literature, the solid solution of Cu and Ag is small. The inset in Figs. 2(a) and 2 (b) show the expanded view of the respective XRD patterns from 30.5 to 32.5 degree. One notes that the XRD peaks shift toward lower angles, indicating the lattice parameter increases with the increasing content of group IB elements (Cu and Ag). This documents that Cu and Ag substitute on the Ge site in the type-III clathrates. This is further verified by the compositional analysis of the matrix, as shown in Table 1. Furthermore, the BaGe2 phase is also detected as the content of the group IB elements (Cu and Ag) increases. Since the type-I phase is poor in Ba in comparison to the type-III clathrate phase, the BaGe2 phase, rich in Ba, is observed as a minor segregated phase.

5

Figure 2. Powder X-ray diffraction patterns of (a) Ba24CuxGe100-x and (b) Ba24AgyGe100-y after SPS. Insets in 2(a) and 2(b) show expanded views of the XRD pattern from 30.5 to 32.5o. Figure 3 depicts the field emission scanning electron microscopy (FESEM) and back-scattered electron images (BSEI) of Ba24Cu4Ge96 and Ba24Ag4Ge96 after SPS. Figures 3(a) and 3(c) indicate that a fully condensed bulk is obtained after SPS. Figures 3(b) and 3(d) show images of back-scattered electrons from polished surfaces of Ba24Cu4Ge96 and Ba24Ag4Ge96. Letters A and E mark regions of type-I clathrate. Letters B, C, and D indicate type-III clathrates. With the increasing amount of Cu and Ag dopants, the type-I clathrate as well as some segregated BaGe2 (small black regions) are observed, consistent with the XRD result. In the BSEI images, areas that are brighter indicate a higher concentration of Ba, corresponding to the type-III clathrate, while the darker areas indicate phases with a lower concentration of Ba, corresponding to the type-I clathrate. The results are fully confirmed by the EPMA compositional analysis. The composition of area A and area E is Ba19.3Cu3.2Ge96.8 and Ba19.4Ag5.6Ge94.4, respectively, corresponding to the type-I phase. Areas B, C, D are Ba24.5Cu1.4Ge98.6, Ba25.1Cu1.7Ge98.3 and Ba25.5Ag1.8Ge98.2, which are the type-III phase. Based on the compositional analysis, one concludes that Cu and Ag can be doped into the type-III phase and, within the limit of solubility, the amount of Cu and Ag accommodated in the structure increases with their nominal content, as shown in Table 1.

6

Figure 3. FESEM images of (a) Ba24Cu4Ge96 and (c) Ba24Ag4Ge96. Back-scattered electron images (BSEI) of (b) Ba24Cu4Ge96 and (d) Ba24Ag4Ge96. The letters A through E indicate various secondary phases present in the structure. The actual chemical composition is described in the text. According to the charge compensation theory, there are 16 additional electrons per formula unit of type-III Ba24Ge100 clathrates. Doping group IB elements Cu and Ag into the frame on Ge sites leads to a significantly reduced electron concentration. Table 1 shows the actual composition, the Hall coefficient RH, the carrier concentration n, the Hall mobility H, the Seebeck coefficient , and the electrical conductivity for Ba24CuxGe100-x and Ba24AgyGe100-y compounds at 300K. The Hall coefficient is negative, which indicates the n-type conduction character of the two compounds. The carrier concentration of undoped Ba24Ge100 is 5.5 × 1021 cm-3 at room temperature, which is consistent with the literature value of 5.0 × 1021 cm-3[31]. Although doping Cu and Ag into the structure induced a small amount of the type-I phase, Cu and Ag also entered into the type-III matrix, where they significantly reduced the carrier concentration.

7

Table 1. Nominal composition, actual composition, Hall coefficient RH, carrier concentration n, Hall mobility H, Seebeck coefficient , and electrical conductivity for Ba24CuxGe100-x and Ba24AgyGe100-y compounds at 300K.

 H



RH 10-3cm3/C

n 1021cm-3

cm2V-1s-1

Ba24.6Ge100

-1.14

5.50

2.39

-21.60

21

Ba24Cu2Ge98

Ba24.5Cu1.4Ge98.1

-1.29

4.83

2.32

-19.52

18

x=4

Ba24Cu4Ge96

Ba24.7Cu2.0Ge98.0

-2.58

2.42

4.12

-22.13

16

x=6

Ba24Cu6Ge94

Ba24.4Cu1.9 Ge98.1

-5.63

1.11

5.06

-19.74

8.98

y=2

Ba24Ag2Ge98

Ba24.7Ag1.4Ge98.6

-1.34

4.65

2.96

-18.45

22.0

y=4

Ba24Ag4Ge96

Ba24.4Ag2.3Ge97.7

-3.12

2.03

6.24

-22.10

20

y=6

Ba24Ag6Ge94

Ba24.8Ag2.9Ge97.1

-5.58

1.10

7.84

-42.52

15.97

xy

Nominal composition

Actual composition ( matrix)

x=0

Ba24Ge100

x=2



V/K

104Sm-1

Figures 4(a)-4(d) show the temperature dependence of the electrical conductivity and the Seebeck coefficient of Ba24CuxGe100-x and Ba24AgyGe100-y compounds. As the doping amount increases, both the carrier concentration and the electrical conductivity decrease. As a function of temperature, the electrical conductivity decreases, displaying a metallic nature of conduction. Although the carrier mobility distinctly increases upon doping with Cu or Ag, the conductivity decreases because the reduced carrier concentration plays a more prominent role. The carrier concentration in the Ag-doped specimens is slightly lower than in the Cu-doped specimens for the 8

same nominal doping content. This is because the actually accommodated amount of Ag in the type-III clathrate matrix phase is slightly higher than that of Cu. As the data in Figure 4 (c) indicate, the Seebeck coefficient is negative and in accord with the negative Hall coefficient. The single-phase Ba24Ge100 compound has the Seebeck coefficient of about -20 V/K at room temperature and approximately -51 V/K at 873 K. As the Cu and Ag doping content increases, the absolute value of the Seebeck coefficient increases. The magnitude of the Seebeck coefficient of Ba24Cu6Ge94 is 80.5 V/K at 873 K while Ba24Ag6Ge94 has reached its maximum value of the Seebeck coefficient of 87 V/K at 820 K, both Cu and Ag doped structures significantly exceeding the magnitude of the Seebeck coefficient of the undoped sample.

Figure 4. Temperature dependence of the electrical conductivity of (a) Ba24CuxGe100-x and (b) Ba24AgyGe100-y. Temperature dependence of the Seebeck coefficient of (c) Ba24CuxGe100-x and (d) Ba24AgyGe100-y. Figure 5(a) shows the Hall mobility H of Ba24CuxGe100-x and Ba24AgyGe100-y compounds at 300 K as a function of the carrier concentration. Open symbols in Figure 5(a) refers to 9

Ba24GazGe100-z compounds from the literature [38]. As shown in Figure 5(a), the carrier mobility of Ga doped type-III clathrate is smaller than that of the Ag doped type-III clathrate with the same carrier concentration. As the doping content of Cu and Ag increases, the carrier concentration decreases and the carrier mobility increases. From the trend in Figure 5(a) and for the same carrier concentration, the mobility of Ag-doped compounds is higher than the mobility of Cu-doped compounds. According to the compositional analysis of the matrix presented in Table 1, the solubility limit of Ag is slightly higher than that of Cu in type-III clathrate, so the amount of the second phase (type-I clathrate and BaGe2) in Ag-doped type-III clathrates is less than in the Cu-doped clathrate compounds. This might have an influence on the intensity of interface scattering in the two phase region of the structure. Assuming that the average mean free path of the charge carriers is constant, the Seebeck coefficient of degenerate semiconductors is given by the following equation:

8 2 k B2T *     m   3eh 2  3n 

2

3

(1)

where n is the carrier concentration, T is the temperature, m* is the effective mass, kB is the Boltzmann constant and h is the Planck constant. Although the single parabolic model might be an over-simplified model, it should adequately describe the trend in the transport properties of Cuand Ag-doped clathrate compounds. Figure 5(b) shows Pisarenko plots of the measured Seebeck coefficient at 300K for Ba24CuxGe100-x and Ba24AgyGe100-y as a function of the carrier concentration. The fitted data suggest that the effective mass of the respected compounds falls in the range from m0 to 3.6 m0, where m0 is the free electron mass. The effective mass increases with the increasing carrier concentration, indicating that the conduction band of the compounds is likely non-parabolic.

10

Figure 5. (a) Hall mobility H, (b) Pisarenko plots showing the measured Seebeck coefficients at 300K of Ba24CuxGe100-x and Ba24AgyGe100-y compounds as a function of the carrier concentration. Open symbols in (a) refers to Ba24GazGe100-z compounds from reference 38. We estimated the optimum carrier concentration for the Ba24Ge100 compound by the following equations [2]: 2m*kBT 2 ) exp  h2 3

n  2(

  s

1 2

(2)

(3)

where  is the reduced Fermi level, the scattering parameter s is -1/2 for acoustic phonons scattering, and the electron effective mass m* is 3.6 m0 according to the Pisarenko plot in Fig. 5(b). The optimal carrier concentration of the Ba24Ge100 compound is calculated as 1.7×1020cm-3, which is about one order of magnitude lower than the lowest carrier concentration observed in Ba24Ag6Ge94. Figures 6(a) and 6(b) show the temperature dependent power factor for Ba24CuxGe100-x and Ba24AgyGe100-y clathrate compounds computed using experimental values of the electrical conductivity and the Seebeck coefficient. As plotted in Figure 6(a), below about 500 K, the power factor of undoped Ba24Ge100 exceeds all Cu-doped compounds. However, above 500 K, the power 11

factor of Cu-doped Ba24Cu2Ge98 and Ba24Cu4Ge96 is higher than that of the undoped sample and reaches values close to 6.0×10-4 Wm-1K-2 at 873 K. The power factor of the highest Cu-doped compound, Ba24Cu6Ge94, is distinctly uncompetitive and reflects its poor electrical conductivity. In contrast, the power factor of all Ag-doped clathrate compounds is higher than that of the undoped compound and higher than the power factor of the corresponding Cu-doped compounds. Moreover, the power factor increases with the increasing Ag doping content even for the highest doped Ba24Ag6Ge94, which, in fact, shows the best power factor at all temperatures and reaches 7.8×10-4Wm-1K-2 at about 873 K, nearly twice the value of undoped Ba24Ge100.

Figure 6. Temperature dependent power factor of (a) Ba24CuxGe100-x and (b) Ba24AgyGe100-y clathrate compounds. Temperature dependence of the thermal conductivity is plotted in Figure 7. For comparison, the temperature dependent thermal conductivity of the type-I Ba8Ge43 clathrate is also plotted in Figures 7(a) and 7(b) [17]. Thermal conductivities of both Ba8Ge43 (type-I clathrate) and Ba24Ge100 (type-III clathrate) are quite high. Ba24Ge100 has the thermal conductivity of 2.8 W/mK at room temperature and an even higher thermal conductivity of 4.0 W/mK at 873 K. As the content of Cu and Ag increases, the thermal conductivity shows a markedly decreasing trend. Although the thermal conductivity of Cu and Ag-doped compounds increases with temperature, the trend distinctly weakens in compounds with higher Cu and Ag content and, in Ba24Cu6Ge94, the thermal conductivity clearly decreases with the increasing temperature, at least up to close to 750 K. The total thermal conductivity = L+e, consists of the lattice component L and the electronic thermal conductivity e, which can be estimated using the Wiedemann-Franz law, e = L0T, where the Lorenz number L0 can be calculated using Eqs. 4 and 5, based on the single parabolic band model with the relaxation time approximation and assuming the dominance of acoustic 12

phonon scattering [43-45]. Figure 7(c) displays the electronic parts of the thermal conductivity. e decreases upon Cu and Ag doping in accord with the decreasing electrical conductivity. The data also show that the lattice thermal conductivity decreases quite dramatically with Ag doping. The lattice thermal conductivity of Ba24Ag6Ge94 is 0.71 W/mK at room temperature and only 0.42 W/mK at 873 K.

L(

 kB 2  (r  7 / 2) Fr  5 / 2 ( F ) )    2 ( F ) e  (r  3 / 2) Fr 1 / 2 ( F ) 

  F  

r  5 2Fr 3 2  F  r  3 2Fr 1 2  F 

(4)

(5)

Figure 7. (a) Temperature dependence of the thermal conductivity of Ba24CuxGe100-x. (b) Temperature dependence of the thermal conductivity of Ba24AgyGe100-y. (c) The lattice L and the electronic e parts of the thermal conductivity of Ba24CuxGe100-x and Ba24AgyGe100-y clathrate compounds. The dimensionless thermoelectric figure of merit ZT is calculated by ZT = 2T/, using experimental values for , , and . Figures 8(a) and 8(b) show the temperature dependence of ZT for Ba24CuxGe100-x and Ba24AgyGe100-y compounds. With the increasing content of Cu and Ag, the ZT value clearly increases. This is especially the case for Ag-doped compounds where their reduced carrier concentration, low thermal conductivity, and marginally higher Seebeck coefficients yield ZT values as high as 0.31 at 850 K for Ba24Ag6Ge94. This is to compare to the ZT value of 0.17 at 873 K for Ba24Cu6Ge94 and a meager ZT = 0.1 at 873 K for undoped Ba24Ge100.

13

Figure 8. (a) Temperature dependence of ZT for Ba24CuxGe100-x, (b) Temperature dependence of ZT for Ba24AgyGe100-y. Conclusions In this paper, group IB elements Cu and Ag were used as dopants to synthesize the dominantly type-III clathrate compounds Ba24CuxGe100-x, Ba24AgyGe100-y by a vacuum melting combined with the subsequent spark plasma sintering (SPS) process. The effects of Cu and Ag doping on the microstructure and thermoelectric properties were systematically studied. Cu and Ag are effective dopants in the type-III clathrate matrix and reduce the carrier concentration and the thermal conductivity. Ag is more effective than Cu doping. The decreased carrier concentration, in turn, enhances the Seebeck coefficient and the carrier mobility, leading to improved thermoelectric properties. Ag doping appears to be particularly effective with the ZT value in Ba24Ag6Ge94 reaching 0.31 at 873 K. This is to be compared with a similarly Cu-doped Ba24Cu6Ge94 that attains only about 0.17 at 873 K, and a nearly twice lower value of ZT in un-doped Ba24Ge100. Eexploring a possibility of doping on the site of Ba with the group IA elements (Na, K) might be an additional avenue how to reduce the carrier concentration and bring it closer to the optimum one. Acknowledgement This work is financially supported by the National Basic Research Program of China (973 program) under project 2013CB632502, the Natural Science Foundation of China (Grant No.51402222, 51172174, 51521001 and 51632006) and the 111 Project of China (Grant No. B07040). Reference: 14

[1] G. J. Snyder and E. S. Toberer, Nat. Mater. 7 (2008) 105. [2] D.M. Rowe, in CRC Handbook of Thermoelectrics, Taylor & Francis, Boca Raton, FL (1995). [3] X. Su, P. Wei, H. Li, W. Liu, Y. Yan, P. Li, C. Su, C. Xie, W. Zhao and P. Zhai, Adv. Mater. (2017) 1602013. [4] W. Liu, K. Yin, Q. Zhang, C. Uher and X. Tang, Natl. Sci. Rev.0 (2017) 1. [5] H. Xie, X. Su, G. Zheng, T. Zhu, K. Yin, Y. Yan, C. Uher, M. G. Kanatzidis and X. Tang, Adv. Energy Mater. 7 (2016) 1601299. [6] Y. Zheng, S. Wang, W. Liu, Z. Yin, H. Li, X. Tang and C. Uher, J. Phys. D: Appl. Phys. 47 (2014) 115303. [7] D. Kahn and J. P. Lu, Phys. Rev. B 56 (1997) 13898. [8] D. G. Cahill, S. K. Watson and R. O. Pohl, Phys. Rev. B 46 (1992) 6131. [9] Y. Guyot, L. Grosvalet, B. Champagnon, E. Reny, C. Cros and M. Pouchard, Phys. Rev. B 60 (1999) 14507. [10] F. Tournus, B. Masenelli, P. Mélinon, D. Connétable, X. Blase, A.M. Flank, P. Lagarde, C. Cros and M. Pouchard, Phys. Rev. B 69 (2004) 035208. [11] S. L. Fang, L. Grigorian, P. C. Eklund, G. Dresselhaus, M. S. Dresselhaus, H. Kawaji and S. Yamanaka, Phys. Rev. B 57 (1998) 7686. [12] Y. Takasu, T. Hasegawa, N. Ogita, M. Udagawa, M.A. Avila, K. Suekuni, I. Ishii, T. Suzuki and T. Takabatake, Phys. Rev. B 74 (2006) 1743031. [13] S. Deng, X. Tang and Q. Zhang, J. Appl. Phys. 102 (2007) 043702. [14] C. H. Lee, H. Yoshizawa, M. A. Avila, I. Hase, K. Kihou and T. Takabatake, J. Phys: Conf. Ser. 92 (2007) 012169. [15] A. Saramat, G. Svensson, A. E. C. Palmqvist, C. Stiewe, E. Mueller, D. Platzek, S. G. K. Williams, D. M. Rowe, J. D. Bryan and G. D. Stucky, J. Appl. Phys. 99 (2006) 023708. [16] X. Shi, J. Yang, S. Bai, J. Yang, H. Wang, M. Chi, J. R. Salvador, W. Zhang, L. Chen and W. Wong-Ng, Adv. Funct. Mater. 20 (2010) 755.

15

[17] U. Aydemir, C. Candolfi, H. Borrmann, M. Baitinger, A. Ormeci, W. Carrillo-Cabrera, C. Chubilleau, B. Lenoir, A. Dauscher and N. Oeschler, Dalton Trans. 39 (2010) 1078. [18] I. Ishii, Y. Suetomi, T.K. Fujita, K. Suekuni, T. Tanaka, T. Takabatake, T. Suzuki and M. A. Avila, Phys. Rev. B 85 (2012) 085101. [19] W. Wong-Ng, Q. Huang, I. Levin, J. C. Woicik, X. Shi, J. Yang and J. A. Kaduk, Powder Diffr. 27 (2012) 25. [20] G. S. Nolas, D. G. Vanderveer, A. P. Wikinson and J. L. Cohn, J. Appl. Phys. 91 (2002) 8970. [21] G. S. Nolas, C. A. Kendziora, J. Gryko, J. Dong, C. W. Myles, A. Poddar and O. F. Sankey, J. Appl. Phys. 92 (2002) 7225. [22] C. W. Myles, J. Dong and O. F. Sankey, Phys. Stat. Sol. B 239 (2003) 26. [23] M. Beekman, W. Wong-Ng, J. A. Kaduk, A. Shapiro and G. S. Nolas, J. Solid State Chem. 180 (2007) 1076. [24] Y. Dong and G. S. Nolas, CrystEngComm 17 (2015) 2242. [25] W. Carrillo-Cabrera, J. Curda, H. G. von Schnering, S. Paschen and Y. Grin, Ζ. Kristallogr. NCS 215 (2000) 207. [26] W. Carrillo-Cabrera, J. Curda, K. Peters, S. Paschen, Y. Grin and H. G. von Schnering, Ζ. Kristallogr. NCS 216 (2001) 183. [27] J. H. Kim, N. L. Okamoto, K. Kishida, K. Tanaka and H. Inui, Acta Mater. 54 (2006) 2057. [28] J. H. Kim, N. L. Okamoto, K. Kishida, K. Tanaka and H. Inui, J. Appl. Phys. 102 (2007) 034510. [29] J. H. Kim, N. L. Okamoto, K. Kishida, K. Tanaka and H. Inui, J. Appl. Phys. 102 (2007) 094506. [30] H. Fukuoka, K. Iwai, S. Yamanaka, H. Abe, K. Yoza and L. Häming, J. Solid State Chem. 151 (2000) 117. [31] S. Paschen, V. H. Tran, M. Baenitz, W. Carrillo-Cabrera, Y. Grin and F. Steglich, Phys. Rev. B 65 (2002) 134435.

16

[32] W. Carrillo-Cabrera, R. Cardoso Gil, S. Paschen and Y. Grin, Ζ. Kristallogr. NCS 218 (2003) 397. [33] S. J. Kim, S. Hu, C. Uher, T. Hogan, B. Huang, J. D. Corbett and M. G. Kanatzidis, J. Solid State Chem. 153 (2000) 321. [34] H. Shimizu, T. Fukushima, T. Kume, S. Sasaki, H. Fukuoka and S. Yamanaka, J. Appl. Phys. 101 (2007) 113531. [35] T. Rachi, H. Yoshino, R. Kumashiro, M. Kitajima, K. Kobayashi, K. Yokogawa, K. Murata, N. Kimura, H. Aoki and H. Fukuoka, Phys. Rev. B 72 (2005) 144504. [36] N. L. Okamoto, J. H. Kim, K. Tanaka and H. Inui, Acta Mater. 54 (2006) 5519. [37] K. Akai, K. Koga and M. Matsuura, Proceedings of the 2006 International Conference on Thermoelectric (2006) 614. [38] J. Fu, X. Su, Y. Zheng, H. Xie, Y. Yan, X. Tang and C. Uher, ACS Appl. Mater. Inter. 7 (2015) 19172. [39] S. Deng, Y. Saiga, K. Suekuni and T. Takabatake, J. Appl. Phys. 108 (2010) 073705. [40] Y. Saiga, K. Suekuni, S. K. Deng, T. Yamamoto, Y. Kono, N. Ohya and T. Takabatake, J. Alloy. Compd. 507 (2010) 1. [41] Y. Saiga, B. Du, S. K. Deng, K. Kajisa and T. Takabatake, J. Alloy. Compd. 537 (2012) 303. [42] B. Du, Y. Saiga, K. Kajisa and T. Takabatake, J. Appl. Phys. 111 (2012) 013707. [43] G. A. Slack and M. A. Hussain, J. Appl. Phys. 70 (1991) 2694. [44] W. Liu, H. Chi, H. Sun, Q. Zhang, K. Yin, X. Tang, Q. Zhang and C. Uher, Phys. Chem. Chem. Phys. 16 (2014) 6893. [45] Y. Zheng, Q. Zhang, X. Su, H. Xie, S. Shu, T. Chen, G. Tan, Y. Yan, X. Tang and C. Uher, Adv. Energy Mater. 5 (2015) 1401391.

17