Promising bulk nanostructured Cu₂Se thermoelectrics via high throughput and rapid chemical synthesis

Abstract

A facile and high yield synthesis route was developed for the fabrication of bulk nanostructured copper selenide (Cu₂Se) with high thermoelectric efficiency. Starting from readily available precursor materials and by means of rapid and energy-efficient microwave-assisted thermolysis, nanopowders of Cu₂Se were synthesized. Powder samples and compacted pellets have been characterized in detail for their structural, microstructural and transport properties. α to β phase transition of Cu₂Se was confirmed using temperature dependent X-ray powder diffraction and differential scanning calorimetry analyses. Scanning electron microscopy analysis reveals the presence of secondary globular nanostructures in the order of 200 nm consisting of <50 nm primary particles. High resolution transmission electron microscopy analysis confirmed the highly crystalline nature of the primary particles with irregular truncated morphology. Through a detailed investigation of different parameters in the compaction process, such as applied load, heating rate, and cooling profiles, pellets with preserved nanostructured grains were obtained. An applied load during the controlled cooling profile was demonstrated to have a big impact on the final thermoelectric efficiency of the consolidated pellets. A very high thermoelectric figure of merit (ZT) above 2 was obtained at 900 K for SPS-compacted Cu₂Se nanopowders in the absence of the applied load during the controlled cooling step. The obtained ZT exceeds the state of the art in the temperature ranges above phase transition, approaching up to 25% improvement at 900 K. The results demonstrate the prominent improvement in ZT attributed both to the low thermal conductivity, as low as 0.38 W m⁻¹ K⁻¹ at 900 K, and the enhancement in the power factor of nanostructured Cu₂Se. The proposed synthesis scheme as well as the consolidation could lead to reliable production of large scale thermoelectric nanopowders for niche applications.

---

1. Introduction

Thermoelectric (TE) materials are solid-state energy-converters, which can convert heat into electricity or serve as Peltier heat pumps. They have a wide range of applications in thermal management and power generation.¹ The materials are classified with respect to their intended operational range as low (from room temperature up to 250 °C), medium (from 200 °C up to 600 °C) and high temperature (above 600 °C)¹ TEs. Several families of materials have been introduced for each respective temperature range. Inorganic solids with low thermal conductivity are of great interest for TE applications. The formation of nanostructures is one of the important strategies for thermal conductivity reduction through phonon scattering.^2–9 Recently, copper selenide (Cu_2−xSe) has gained a renewed interest for the medium temperature range operation. Copper selenide structures have previously been used in photovoltaic cells and devices^10–12 and are being considered even in biological systems.¹³ This renewed interest can be attributed to the low thermal conductivity of the material as well as to its phase changing nature, which is an interesting aspect of the material.^14–20 The constituent elements are abundant in nature, and are considered as more benign in comparison to many other TE structures, especially those containing Pb, Sb, As, etc. Additionally, Cu_2−xSe exhibits an electronic–ionic conduction mechanism that has a great influence on different properties, such as the specific heat capacity (C_p), which approaches the theoretical limit of solid crystalline materials in contrast to liquids.^{14,15,17,21,22} Different methods have been used for the fabrication of these materials, including melt alloying,^{10,14,16,17,20,22–24} mechanical alloying (ball milling),^15,25,26 self-propagating high-temperature synthesis (SHS),¹⁹ and chemical routes using aqueous solutions either directly^18,27–31 or solvothermally precipitated using autoclave^32–37 or organometallic precursors.^13,38–40 Some of these processes require high temperatures (above 1000 K), long processing times, which consumes a lot of energy, or inert atmosphere glove boxes for powder handling, which are expensive.

Our motivation for this study is to develop an alternative, fast, energy and resource effective synthesis method for Cu_2−xSe-based compounds with improved performance intended for TE applications. Our technique is based on as-received materials and the synthesis is performed via microwave (MW) assisted heating at temperature (250 °C) much lower than those used in conventional methods (above 1000 °C). Fabricated nanomaterials have been carefully consolidated to preserve the nanostructure and characterized for their temperature-dependent TE properties depending on the sintering temperature and pressure.

2. Experimental

Copper acetate [Cu(CO₂CH₃)₂·H₂O], selenium powder (Se, >99.5%), and the organic solvents oleic acid (C₁₈H₃₄O₂), 1-octadecene (C₁₈H₃₆, ODE), and trioctylphosphine [P(C₈H₁₇)₃, TOP] methanol and hexane (C₆H₁₄) were all purchased from Sigma Aldrich and used as received. The precursor mixture was prepared by adding a stoichiometric amount of copper acetate and Se powder in a MW glass vial with 4 ml of ODE and 8 ml of oleic acid. The vial was then sealed with silicon septa caps and degassed with several purges of nitrogen followed by vacuum, into which 2 ml of TOP was added. The degassing process is performed to prevent TOP oxidation. The vial was then loaded into the laboratory MW synthesizer (Biotage® Initiator+), pre-stirred for 2 minutes and then heated with the following parameters; a MW power of 400 W, a synthesis temperature of 250 °C, and a synthesis holding time of 5 min. The reaction process on the MW synthesizer is mechanized using a sample loader therefore the human influence is minimized. This results in highly reproducible results. Due to the limitation of the volume on the particular instrument a full batch results in about 12 grams of nanopowder. The powders were then washed several times with hexane and methanol and left in vacuum oven to dry at 60 °C for 3 hours. Then the dried powder underwent several characterization processes, including Scanning Electron Microscopy coupled with EDX (SEM, Zeiss Ultra 55), Transmission Electron Microscopy (TEM, Jeol JEM 2100), and X-Ray Powder Diffraction (XRPD) (PANalytical Empyrean, PANalytica X'pert PRO with Anton Paar XRK 900 High Temperature Chamber). Afterwards, the powder batches were compacted using Spark Plasma Sintering (SPS, Dr Sinter 825). The optimized sintering parameters were identified as a heating rate of 75 °C min⁻¹, maximum temperature of 400 °C, and no holding time. The cooling of the pellets had to be controlled to minimize the effect of the phase transition on the structural integrity of the pellet. Inductively coupled plasma optical emission spectrometry (iCAP 6500, Thermo Scientific) is used to measure the concentration of Cu and Se ions in powder and compacted samples. All high temperature transport measurements were carried out under helium (He) atmosphere in the range of 300–900 K. The Seebeck coefficient (S) and the electrical conductivity measurements were simultaneously measured using an ULVAC-RIKO ZEM3 system in a low pressure He atmosphere. The total thermal conductivity κ_tot was calculated using the equation κ_tot = D·ρ·C_p, where D is the thermal diffusivity, C_p the specific heat capacity and ρ the bulk density of the pellet. The thermal diffusivity was measured using a laser flash analysis system (LFA 457, Netzsch). The specific heat capacity, C_p, was measured by Differential Scanning Calorimetry (Netzsch-DSC 404 Pegasus214 Polyma), and the density was obtained from the Archimedes' method. High temperature Seebeck coefficient and electrical conductivity measurements were repeated using an in-house developed transport property measurement system under Ar atmosphere at the University of Michigan to confirm the electronic transport properties of these compounds.

3. Results and discussions

By a careful exploration of the MW synthesis parameters, the optimum synthesis temperature was identified as 250 °C and the holding time of 5 min. In this work, several batches underwent the same procedure and the XRD results were identical; all these batches are mixed together to form a large batch of nanopowder. Therefore, the study performed on the compaction and optimization of the compaction parameters were performed on the same big batch of nanopowder. Fig. 1a illustrates the XRPD pattern for samples at different stages of processing: as-prepared, SPS-compacted, and after measurements of the TE transport properties.

Fig. 1 Room temperature XRPD pattern of (a) powders at different processing steps; as-prepared, compacted powder and the pellet after the TE measurements and (b) temperature-dependent XRPD of the copper selenide powder (indexed with ICDD card number 29-0575 for α-Cu₂Se (*), 01-079-1841 for β-Cu₂Se (▾) and 98-002-9287 for Si (☐)).

Depending on the temperature, Cu_2−xSe has two main phases: a monoclinic α phase below 140 °C and a cubic β phase above 140 °C. The XRPD 2θ scans of the compacted sample show a slight shift with respect to the as-prepared sample from higher angles to lower angles at 26 and 39.7 degrees, corresponding to the (2 2 2) and (5 3 0) planes, respectively, which can be attributed to the recrystallization of some of the defects and/or amorphous parts and the formation of the proper phase. Additionally, the room temperature XRPD pattern taken after TE transport measurements shows the transformation of most of the Cu_2−xSe from the β phase to the α phase. Moreover, Fig. 1b illustrates the temperature-dependent XRPD of the powder (a mixture of the as-prepared Cu_2−xSe nanopowder and Si powder used as the internal reference) at elevated temperatures and confirms the α to β phase transition around 100 °C, a much lower temperature than that reported in the literature. According to Chakrabarti et al.⁴¹ and Gahtori et al.,²⁶ the transformation takes place at above 120 °C. In the material under investigation, the phase transformation commences at temperatures lower than 100 °C and is completed at about 110 °C. This transformation is also confirmed by the DSC measurements, which were performed in 2 cycles, as displayed in Fig. 2. The data reveal that the phase transformation process is reversible, the traces being identical upon heating and cooling. The process parameter was set for 2 °C min⁻¹ increments, and the phase transition takes place at about 105 °C, where we observe a sharp peak in the derivative of the heat flow.

Fig. 2 DSC thermogram of Cu_2−xSe nanopowder with 2 heating–cooling cycles.

Fig. 3 displays SEM micrographs of the as-prepared powder. The particle size ranges from 50 nm up to several hundreds of nanometers, with an average size of 150 nm ± 20 nm (counting about 300 particles/grain from several micrographs). Additionally, secondary particles consisting of much smaller domains, as primary particles, are visible in Fig. 3b. These primary particles have sizes in the range of 15–40 nm, with an average size of 25 ± 5 nm. This is seen from the TEM image presented in the inset of Fig. 3b, which confirms the presence of the primary smaller particles within the observed clusters.

Fig. 3 (a) and (b) SEM micrographs of the as-prepared Cu_2−xSe nanopowder at different magnifications (inset in (b) is the TEM micrograph of as-prepared Cu_2−xSe nanopowder).

HRTEM was performed on the as-prepared powder samples to more closely observe the crystal structure and geometry of the primary particles in the cluster. Fig. 4 represents the TEM micrograph of the as-prepared powder at different magnifications. The HRTEM image depicts the primary particles with a high crystallinity. The interplanar distance was calculated as 0.33 nm, corresponding to the (2 2 2) plane of the α-Cu_2−xSe phase matched with the ICDD card number 29-0575. The FFT performed on the designated area also displays the periodicity of the interplanar distance.

Fig. 4 (a) Low, and (b) high resolution TEM micrographs of as-prepared powder with the corresponding FFT of the interplanar distances.

Table 1 lists the SPS processing conditions for the samples used for transport characterization. The heating rate during the SPS process was selected in such a way as to prevent any accidental surge of current that could have melted the powder. Additionally, the densification curve provided us with the plateau limit for the compaction (data not shown). The optimum parameters were identified as a 400 °C compaction temperature with a heating rate of 75 °C min⁻¹ and no holding time. The most important difference in the SPS process parameters was the fact that, in one compaction processes, the load/pressure on the sample was released to its default pressure condition during cooling (Cu₂Se_Nano_nl), while in the other process, the load/pressure was maintained (Cu₂Se_Nano_l).

Table 1 SPS processing conditions and the resulting sample density

Sample tag	Chemical process parameters	Mass density (g cm^−3)	Heating rate (°C min^−1)	Compaction temperature (°C)	Holding time	Load (MPa)	Cooling load
Cu2Se_Nano_nl	250 °C 5 mins holding time	5.97 (≈90%)	75	400	0	75	No
Cu2Se_Nano_l	250 °C 5 mins holding time	6.02 (≈90%)	75	400	0	75	Yes

---

The density of the compacted Cu₂Se nanopowder samples was around 90% of the theoretical density, a relatively smaller density compared to the density of nano-structured and bulk Cu₂Se reported by other groups.^17,27,42,43 This might be due to differences in the grain size.⁴⁴ Different particle sizes have different surface energy, which may cause resistance to further compaction.⁴⁵ SEM micrographs of two compacted samples with different compaction parameters are presented in Fig. 5a and b. Although both samples have nearly the same density, the sample Cu₂Se_Nano_nl contains a mixture of nano- and micron-sized grains, while in the other sample the grains have grown quite a lot, a feature attributed to maintaining the load/pressure even during the cooling step. Due to the high pressure and sintering temperature, as well as the very short exposure time to the sintering temperature, several smaller particles might start diffusing into each other and aggregate to form larger grains, while some other particles remain small. During the cooling stage that maintained the load/pressure on the compacted powder, the diffusion of small particles might be enhanced; thus, the sample labeled as nano Cu₂Se_l has much larger grain size in comparison to the sample designated as Cu₂Se_nl. Nano and micro grain sizes and their distribution in the structure have a strong influence on the TE properties of the compound. In order to assure the observed difference on the transport is mainly due to the microstructure differences, elemental composition analysis of powder and SPS compacted samples have been performed using ICP-OES, which revealed no significant difference revealing stoichiometric phase where Cu : Se atomic ratio is obtained as 2 : 1.

Fig. 5 SEM micrographs of (a) Cu₂Se_Nano_nl and (b) Cu₂Se_Nano_l at similar magnifications.

TE transport properties of nanostructured Cu₂Se samples were determined based on measurements of the Seebeck coefficient, electrical conductivity, and thermal conductivity between 300 and 900 K. The results are illustrated in Fig. 6. The electrical conductivity (Fig. 6a) decreases with the increasing temperature for all samples, likely due to electrons being scattered by acoustic phonons, which is a typical behavior of heavily doped semiconductor materials. The sample with no load applied during cooling shows a higher electrical conductivity of 1126 S cm⁻¹ compared to the conductivity of 884 S cm⁻¹ measured on the sample on which the load was maintained during the cool down. Both of the samples show a sudden drop in the electrical conductivity at the phase transition temperature, followed by an increase and the final values of 170 S cm⁻¹ for Cu₂Se_Nano_nl and about 100 S cm⁻¹ for Cu₂Se_Nano_l reached at 900 K. Comparing the electrical conductivity with the values reported by Gahtori et al.²⁶ and Yu et al.,¹⁵ some 15% (≈1000 S cm⁻¹) and almost 75% (≈725 S cm⁻¹) enhancements are observed at RT, respectively. This might be due to the percolated network of large (micron-sized) grains in the Cu₂Se_nl compound. The conductivity values follow the same trend in all 3 samples and fall to a value of 150 S cm⁻¹ at 900 K in Cu₂Se_Nano_nl, similar to the work of Gahtori et al.²⁶ and about 50% higher than the value (≈95 S cm⁻¹) reported by Yu et al.¹⁵

Fig. 6 Thermoelectric transport data for the two samples measured between 300 K and 900 K: (a) electrical conductivity, (b) Seebeck coefficient, (c) thermal conductivity, and (d) estimated ZT value.

The positive sign of the Seebeck coefficient confirms that holes are the main carriers in both samples (Fig. 6b). The RT Seebeck coefficient value is 70 μV K⁻¹ and increases to about 250 μV K⁻¹, which is roughly 25% more than values reported in the work of Gahtori et al.²⁶ (≈200 μV K⁻¹), but roughly similar to the values reported by Yu et al.¹⁵ (≈255 μV K⁻¹). The temperature dependence of the Seebeck coefficient indicates that there is a tendency for an extrinsic to intrinsic transition at high temperatures. Using the empirical relation for the band gap, E_g = 2eS_maxT_max, and our Seebeck coefficient data,⁴⁶ the band gap energy is estimated to be 0.72 eV for the sample Cu₂Se_Nano_nl, and 0.83 eV for the sample Cu₂Se_Nano_l. Both values are relatively lower than the band gap energy of nano-Cu₂Se reported by Riha et al.⁴⁷ but close to the band gap energy reported by Al-Mamun et al.⁴⁸ Several groups have reported different values of energy gap for Cu₂Se.^47–50 The wide range in observed band gap energies is attributed to local differences in the Cu to Se stoichiometry, large grain size distributions, and grain size effects in this compound.

The thermal conductivity values follow the same trend as reported in the literature, namely, the thermal conductivity of Cu₂Se_Nano_l starts at about 0.75 W m⁻¹ K⁻¹ at RT (Fig. 6c), abruptly drops at the phase transition region, and continues to decrease until it reaches a value of about 0.38 W m⁻¹ K⁻¹ at 900 K. The reduced thermal conductivity value in our Cu₂Se samples originates from the highly disordered structure due to randomly distributed Cu atoms in the structure,^17,26,51,52 as well as the phonon filtering effect due to different grain sizes, and grain boundary scattering present in the compacted pellet. We used the Wiedemann–Franz law to calculate the lattice contribution to the thermal conductivity of all samples by subtracting the electronic term (κ_el = L₀σT where L₀, σ and T are the Lorenz number, electrical conductivity and absolute temperature, respectively) from the total thermal conductivity. L₀ is a physical parameter that strongly depends on the band structure, scattering mechanism, carrier density, temperature and chemical potential. It is difficult to calculate its absolute value due to requirement to know all these parameters. However, there are few approximations for calculating this number for different materials. One approximation is made by Larsen et al., where he proposed an analytical solution of L₀ based on the Single Parabolic Band (SPB) model as a function of carrier concentration (n) and (m*T)^−3/2 (where m* is the effective mass) along with various sets of parameters for distinct carrier scattering mechanisms.⁵³ However, when the Hall carrier concentration, n_H of a material is not available or materials have complex band structure, the use of the Larsen approximation is not possible. Another approximation regarding the calculation of L₀ is proposed by Kim et al.⁵⁴ According to the model proposed by Kim et al., L can be calculated via where L is in the order of 10⁻⁸ W Ω K⁻² and S in μV K⁻¹, for each temperature. Their model is based on assuming SPB model with acoustic phonon scattering (APS) in which both L₀ and S are taken as parametric functions of only the reduced chemical potential (η = ξ/k_BT, where k_B is Boltzmann constant); thus, there is no need of explicit knowledge of temperature (T), carrier concentration (n), or effective mass (m*) is required to relate them. We used the model proposed by Kim et al. for calculating L₀ for each temperature and obtained L₀ as 2.02 × 10⁻⁸ V² K⁻² for Cu₂Se_Nano_nl and 1.96 × 10⁻⁸ V² K⁻² for Cu₂Se_Nano_l at room temperature.

Looking at the thermal conductivity plot presented in Fig. 7 as well as the initial thermal conductivity values illustrated in Table 2, a clear 30% reduction in κ_tot is noticeable between the two samples at 850 K. The estimated κ_el of the total thermal conductivity at room temperature for the sample Cu₂Se_Nano_nl and for Cu₂Se_Nano_l reaches levels of 0.67 W m⁻¹ K⁻¹ and 0.47 W m⁻¹ K⁻¹, compared to the total thermal conductivity of 0.82 W m⁻¹ K⁻¹ and 1.02 W m⁻¹ K⁻¹, respectively. Moreover, the lattice thermal conductivity at the highest temperature of measurement (900 K) shows a 51% reduction at 900 K, which is the result of the nano-sized grains, causing enhanced phonon scattering, and filtering of low and high frequency phonons, both processes resulting in a reduced lattice contribution. The lattice contribution at phase transition temperature does not reflect its actual value. Because the crystal structure, hence the carrier density and scattering mechanism change dramatically around this temperature.⁹ Therefore, L₀ predicted by SPB model approximation for this temperature does not reflect the actual value.

Fig. 7 Thermal conductivity contributions in the two Cu₂Se_Nano samples. Cu₂Se_Nano_nl: square symbols with solid, dashed and dotted lines; Cu₂Se_Nano_l: circle symbols with solid, dashed and dotted lines.

Table 2 Room temperature electrical conductivity, total thermal conductivity and electronic thermal conductivity of the two samples

Sample tag	σ (S cm^−1)	κtot (W m^−1 K^−1)	κel (W m^−1 K^−1)
Cu2Se_Nano_nl	1126	0.82	0.67
Cu2Se_Nano_l	884	1.02	0.47

---

Any decrease in the size of the grains in a material results in an electrical conductivity reduction until the quantum confinement region is reached. There are more interfaces and grain boundaries in nanostructured materials in comparison to bulk samples. Electrons are scattered passing these interfaces and grain boundaries, resulting in suppressed electrical conductivity compared to their bulk counterparts.⁵⁵ However, recently it was shown that multiple scattering mechanisms in bulk samples enhance the power factor (PF) of the material.⁵⁶ In other words, in nano/micro mix-grained composites, the dispersed nanoparticles are designed to preferably scatter phonons, while micro particles form a percolated network for electron transport.^57,58 The Seebeck coefficient increased in nano/micro mix-grained composites. We believe a similar phenomenon is causing an enhancement in the Seebeck coefficient and electrical conductivity in our samples as compared to reported values in the literature. Fig. 6d represents the overall TE figure of merit ZT of both samples prepared. High power factor and low thermal conductivity result in an exceptionally high TE figure of merit ZT for Cu₂Se_Nano_nl as compared to the highest reported ZT values in Fig. 8. Due to the methodology that preserved the nanostructure, ZT values as high as 2 at 900 K were achieved for the nanostructured Cu₂Se, an almost 25% higher ZT than the maximum value reported by Gahtori et al.²⁶ at the same temperature.

Fig. 8 Comparison of ZT values of Cu₂Se samples prepared in this work with earlier reports.

The increase in ZT in our samples originates from the reduced thermal conductivity compared to other works (at least 10% reduction compared to the work of Gahtori et al.²⁶ and more in comparison to others) as well as the higher Seebeck coefficient achieved with the optimized compaction process. The high ZT at a lower temperature enables the usage of the material at temperatures lower than the material's instability region. According to the phase diagram of copper and selenium, slight variations in the value of x in the Cu_2−xSe create big issues regarding the material's stability above 900 K.⁴¹ This might result in the migration of Cu ions and creation of local stoichiometric variations, resulting in local melting of those domains.

4. Conclusions

Using a highly reproducible and fully automated MW-assisted thermolysis route, copper selenide (Cu₂Se) nanopowders were fabricated. Structural properties of the powder and the pellets compacted from the powder were investigated, and the transport properties of the resulting materials were determined. Optimized SPS compaction parameters were identified and used. The samples have a relatively high power factor and the thermal conductivity of the material is about 10–20% lower than the previously reported values in the literature. An exceptionally high ZT value of 2 was obtained at about 900 K, one of the highest reported values in the literature. The MW-assisted synthesis method provides a promising route for a scalable fabrication of the nanostructured bulk thermoelectric Cu₂Se material, which can be easily adopted by the industry for energy harvesting and other TEG applications. We conclude that using this preparation method with optimized compaction parameters, it is possible to enhance the ZT value of nanostructured Cu₂Se based compounds by ∼60% as compared to its bulk counterpart.

Acknowledgements

This work was supported by the Swedish Foundation for Strategic Research (SSF, Grant no. EM11-0002) and Swedish Research Council (VR-SRL 2013-6780). Sedat Ballikaya acknowledges support by Scientific and Technological Research Council of Turkey (TUBITAK) with project number 115F510 and Scientific Research Projects Coordination Unit of Istanbul University (BAP) with project number of 21272 and 20611. The work at the University of Michigan was supported by the Center for Solar and Thermal Energy Research, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Basic Energy Sciences under the award DE-SC0000957.

References

1. D. M. Rowe, Thermoelectrics Handbook: Macro to Nano, CRC Press, 2010.
2. Z. He, C. Stiewe, D. Platzek, G. Karpinski, E. Müller, S. Li, M. Toprak and M. Muhammed, J. Appl. Phys., 2007, 101, 043707.
3. Z. He, C. Stiewe, D. Platzek, G. Karpinski, E. Müller, S. Li, M. Toprak and M. Muhammed, J. Appl. Phys., 2007, 101, 053713.
4. A. Khan, M. Saleemi, M. Johnsson, L. Han, N. Nong, M. Muhammed and M. S. Toprak, J. Alloys Compd., 2014, 612, 293–300.
5. M. S. Toprak, C. Stiewe, D. Platzek, S. Williams, L. Bertini, E. Müller, C. Gatti, Y. Zhang, M. Rowe and M. Muhammed, Adv. Funct. Mater., 2004, 14, 1189–1196.
6. M. K. Jana, K. Pal, U. V. Waghmare and K. Biswas, Angew. Chem., Int. Ed. Engl., 2016, 55, 7792–7796.
7. S. N. Guin, D. Sanyal and K. Biswas, Chem. Sci., 2016, 7, 534–543.
8. A. Banik, B. Vishal, S. Perumal, R. Datta and K. Biswas, Energy Environ. Sci., 2016, 9, 2011–2019.
9. S. N. Guin, J. Pan, A. Bhowmik, D. Sanyal, U. V. Waghmare and K. Biswas, J. Am. Chem. Soc., 2014, 136, 12712–12720.
10. A. A. Sirusi, S. Ballikaya, C. Uher and J. H. Ross, J. Phys. Chem. C, 2015, 119, 20293–20298.
11. H. H. Kou, Y. M. Jiang, J. J. Li, S. J. Yu and C. M. Wang, J. Mater. Chem., 2012, 22, 1950–1956.
12. M. C. Nguyen, J. H. Choi, X. Zhao, C. Z. Wang, Z. Zhang and K. M. Ho, Phys. Rev. Lett., 2013, 111, 165502.
13. C. M. Hessel, V. P. Pattani, M. Rasch, M. G. Panthani, B. Koo, J. W. Tunnell and B. A. Korgel, Nano Lett., 2011, 11, 2560–2566.
14. H. Liu, X. Shi, F. Xu, L. Zhang, W. Zhang, L. Chen, Q. Li, C. Uher, T. Day and G. J. Snyder, Nat. Mater., 2012, 11, 422–425.
15. B. Yu, W. S. Liu, S. Chen, H. Wang, H. Z. Wang, G. Chen and Z. F. Ren, Nano Energy, 2012, 1, 472–478.
16. H. L. Liu, X. Shi, M. Kirkham, H. Wang, Q. Li, C. Uher, W. Q. Zhang and L. D. Chen, Mater. Lett., 2013, 93, 121–124.
17. S. Ballikaya, H. Chi, J. R. Salvador and C. Uher, J. Mater. Chem. A, 2013, 1, 12478–12484.
18. D. Li, X. Y. Qin, Y. F. Liu, C. J. Song, L. Wang, J. Zhang, H. X. Xin, G. L. Guo, T. H. Zou, G. L. Sun, B. J. Ren and X. G. Zhu, RSC Adv., 2014, 4, 8638–8644.
19. X. Su, F. Fu, Y. Yan, G. Zheng, T. Liang, Q. Zhang, X. Cheng, D. Yang, H. Chi, X. Tang, Q. Zhang and C. Uher, Nat. Commun., 2014, 5, 4908.
20. L. L. Zhao, X. L. Wang, J. Y. Wang, Z. X. Cheng, S. X. Dou, J. Wang and L. Q. Liu, Sci. Rep., 2015, 5, 7671.
21. R. D. Heyding, Can. J. Chem., 1966, 44, 1233–1236.
22. H. Kim, S. Ballikaya, H. Chi, J. P. Ahn, K. Ahn, C. Uher and M. Kaviany, Acta Mater., 2015, 86, 247–253.
23. E. Filippo, D. Manno and A. Serra, J. Alloys Compd., 2012, 538, 8–10.
24. P. Lu, H. L. Liu, X. Yuan, F. F. Xu, X. Shi, K. P. Zhao, W. J. Qiu, W. Q. Zhang and L. D. Chen, J. Mater. Chem. A, 2015, 3, 6901–6908.
25. K. Tyagi, B. Gahtori, S. Bathula, M. Jayasimhadri, N. K. Singh, S. Sharma, D. Haranath, A. K. Srivastava and A. Dhar, J. Phys. Chem. Solids, 2015, 81, 100–105.
26. B. Gahtori, S. Bathula, K. Tyagi, M. Jayasimhadri, A. K. Srivastava, S. Singh, R. C. Budhani and A. Dhar, Nano Energy, 2015, 13, 36–46.
27. V. M. Bhuse, P. P. Hankare, K. M. Garadkar and A. S. Khomane, Mater. Chem. Phys., 2003, 80, 82–88.
28. P. Kumar, K. Singh and O. N. Srivastava, J. Cryst. Growth, 2010, 312, 2804–2813.
29. Y. Zhang, C. Hu, C. Zheng, Y. Xi and B. Wan, J. Phys. Chem. C, 2010, 114, 14849–14853.
30. X. Chen, Z. Li, J. Yang, Q. Sun and S. Dou, J. Colloid Interface Sci., 2015, 442, 140–146.
31. X. Q. Chen, Z. Li and S. X. Dou, ACS Appl. Mater. Interfaces, 2015, 7, 13295–13302.
32. L. W. Mi, Z. Li, W. H. Chen, Q. Ding, Y. F. Chen, Y. D. Zhang, S. S. Guo, H. M. Jia, W. W. He and Z. Zheng, Thin Solid Films, 2013, 534, 22–27.
33. Z. H. Han, Y. P. Li, H. Q. Zhao, S. H. Yu, X. L. Yin and Y. T. Qian, Mater. Lett., 2000, 44, 366–369.
34. D. P. Li, Z. Zheng, Y. Lei, S. X. Ge, Y. D. Zhang, Y. G. Zhang, K. W. Wong, F. L. Yang and W. M. Lau, CrystEngComm, 2010, 12, 1856–1861.
35. Y. F. Liu, J. H. Zeng, C. Li, J. B. Cao, Y. Y. Wang and Y. T. Qian, Mater. Res. Bull., 2002, 37, 2509–2516.
36. H. L. Su, Y. Xie, Z. P. Qiao and Y. T. Qian, Mater. Res. Bull., 2000, 35, 1129–1135.
37. L. Yang, Z. G. Chen, G. Han, M. Hong, Y. C. Zou and J. Zou, Nano Energy, 2015, 16, 367–374.
38. J. Choi, N. Kang, H. Y. Yang, H. J. Kim and S. U. Son, Chem. Mater., 2010, 22, 3586–3588.
39. S. Deka, A. Genovese, Y. Zhang, K. Miszta, G. Bertoni, R. Krahne, C. Giannini and L. Manna, J. Am. Chem. Soc., 2010, 132, 8912–8914.
40. H. I. Hsiang, W. H. Hsu, L. H. Lu, Y. L. Chang and F. S. Yen, Mater. Res. Bull., 2013, 48, 715–720.
41. D. J. Chakrabarti and D. E. Laughlin, Bull. Alloy Phase Diagrams, 1981, 2, 305–315.
42. C. J. Vineis, A. Shakouri, A. Majumdar and M. G. Kanatzidis, Adv. Mater., 2010, 22, 3970–3980.
43. F. S. Liu, M. J. Huang, Z. N. Gong, W. Q. Ao, Y. Li and J. Q. Li, J. Alloys Compd., 2015, 651, 648–654.
44. K. Wei, J. Martin and G. S. Nolas, Mater. Lett., 2014, 122, 289–291.
45. C. R. Berry, Phys. Rev., 1952, 88, 596–599.
46. H. J. Goldsmid and J. W. Sharp, J. Electron. Mater., 1999, 28, 869–872.
47. S. C. Riha, D. C. Johnson and A. L. Prieto, J. Am. Chem. Soc., 2011, 133, 1383–1390.
48. M. Al, A. B. M. O. Islam and A. H. Bhuiyan, J. Mater. Sci.: Mater. Electron., 2005, 16, 263–268.
49. A. M. Hermann and L. Fabick, J. Cryst. Growth, 1983, 61, 658–664.
50. W. Wang, L. Zhang, G. Chen, J. Jiang, T. Ding, J. Zuo and Q. Yang, CrystEngComm, 2015, 17, 1975–1981.
51. G. Chen, Int. J. Therm. Sci., 2000, 39, 471–480.
52. W. Kim, J. Zide, A. Gossard, D. Klenov, S. Stemmer, A. Shakouri and A. Majumdar, Phys. Rev. Lett., 2006, 96, 045901.
53. E. Flage-Larsen and Ø. Prytz, Appl. Phys. Lett., 2011, 99, 202108.
54. H.-S. Kim, Z. M. Gibbs, Y. Tang, H. Wang and G. J. Snyder, APL Mater., 2015, 3, 041506.
55. M. S. Dresselhaus, G. Chen, M. Y. Tang, R. G. Yang, H. Lee, D. Z. Wang, Z. F. Ren, J. P. Fleurial and P. Gogna, Adv. Mater., 2007, 19, 1043–1053.
56. K. Biswas, J. He, I. D. Blum, C. I. Wu, T. P. Hogan, D. N. Seidman, V. P. Dravid and M. G. Kanatzidis, Nature, 2012, 489, 414–418.
57. A. Mehdizadeh Dehkordi, M. Zebarjadi, J. He and T. M. Tritt, Mater. Sci. Eng., R, 2015, 97, 1–22.
58. J. P. Heremans, C. M. Thrush and D. T. Morelli, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70, 115334.

---

Footnote

† Present address: KTH Royal Institute of Technology, Department of Applied Physics, School of Engineering Sciences, Roslagstullsbacken 21, SE-106 91 Stockholm, Sweden.

---