High Temperature Boron-based Thermoelectric Materials

By: Takao Mori, Material Matters Volume 4 Article 2

International Center for Materials Nanoachitectonics (MANA)
National Institute for Materials Science (NIMS) Namiki 1-1, Tsukuba Japan 305-0044

Email: takao.mori@nims.go.jp


The modern world is rapidly approaching the limits of classical energy reserves. Alternative sources of energy, such as the conversion of waste heat to electricity, offer potential and are an incentive to developing viable thermoelectric materials. In particular, materials that can function at high temperature and that withstand large temperature differences need to be developed for use in factories, power plants, incinerators, etc., as well as for reliable radiothermal generation of electricity in space applications.1

Boron-rich cluster compounds typically have melting points above 2200 K, and are therefore attractive materials for such purposes. They are also non-toxic, lightweight, and show remarkable stability in corrosive and acidic environments. Their synthesis is relatively straight forward, the addition of small amounts of carbon, nitrogen, or silicon to rare earth-boron systems (RE-B), to serve as bridging sites, results in the formation of novel boron cluster structures.2

Boron-rich cluster materials typically exhibit low intrinsic thermal conductivity (≤0.02 W cm-1 K-1) even for single crystals, which gives them “built in” merit as thermoelectric materials.2-4 The performance of thermoelectric materials is gauged by a dimensionless figure of merit ZT, where ZT = α2σT/κ, and α is the Seebeck coefficient, σ is the electrical conductivity, and κ the thermal conductivity. Therefore, in a system with low κ it is possible to concentrate on maximizing the power factor, P = α2σ.

In this article, I focus on two novel boron-based compounds that have recently been discovered and that show promise in high temperature thermoelectric applications. The compounds are borosilicide, REB44Si2 (RE = rare earth) and homologous RE-B-C(N) borocarbides, including REB17CN, REB22C2N, and REB28.5C4.

back to top 

Emerging Boron-Based Materials and their Preparation

Boron is a versatile element, tending to form atomic networks based on clusters and 2D atomic nets in compounds. In this sense it is similar to carbon, which is able to form atomic network systems such as fullerenes, nanotubes, and graphite-related materials. Yet, the potential of boron in materials science is still largely untapped. Several striking properties of boron-containing compounds have recently been discovered, such as the superconductivity of MgB2 5 (Prod. No. 553913) and boron-doped diamonds,6 the strong magnetic coupling in magnetically dilute insulators,2 and the formation of a novel elemental structure of boron.7

Boron has one fewer electron than does carbon, which enables the formation of electron-deficient multi-atom networks. Such networks have a special affinity for the rare earth elements, which can supply electrons to the framework to stabilize and form novel structures (namely, new RE-B compounds), while the f electron shell is responsible for interesting properties, such as magnetism.2,8

From an application standpoint, the boron cluster framework provides a light robust “armor” that is acid and corrosion resistant, and can withstand very high temperatures. Additional electronic, magnetic, and other useful properties can be supplied by the metal atom from the “inside.” As noted earlier, the addition of small amounts of such elements as carbon, nitrogen, or silicon further increases the number of novel boron-based compounds that can be created.2 The crystal structure of the REB44Si2 compounds is shown in Figure 1a. It consists of five crystallographically independent B12 icosahedra (20-sided polyhedra) and one B12Si3 polyhedron. In the framework, the rare earth atoms form ladders in the direction of the c-axis, along which one of the B12 icosahedra also forms a chain.


Figure 1. View of the (a) structure and (b) grown crystal of REB44Si2 (scale is cm).

Each of the rare earth-boron complexes REB44Si2, REB17CN, REB22C2N, and REB28.5C4, can be synthesized in a similar way.2 First, an appropriate rare earth oxide is reduced with boron (Prod. No. 266620) upon heating under vacuum. Then, required amounts of elemental silicon (Prod. Nos. 633097, 343250, 267414), carbon (Prod. Nos. 699632, 699624, 496596, 496553), or boron nitride (Prod. No. 255475) are added and heated again to produce the desired RE-B-X material. In the case of REB44Si2, it is possible to grow large crystals, as shown in Figure 1b.9

back to top 

Borosilicides—New High Temperature p-Type Thermoelectric Materials

REB44Si2 compounds exhibit attractive high temperature thermoelectric properties.10,11 Their Seebeck coefficients gradually increase with temperature, exceeding 200 μV K-1 above 1000 K (Figure 2a).10 A low intrinsic thermal conductivity of about 0.02 W cm-1 K-1 has also been reported (Figure 2b).11 Unlike most thermoelectric materials, the figure of merit of REB44Si2 shows a steep increase at temperatures beyond 1000 K, where an extrapolated value for ZT of about 0.2 can be estimated.


Figure 2. Temperature dependence of (a) Seebeck coefficient and (b) thermal conductivity of REB44Si2.

Considering that these compounds are not doped and are not composition optimized, their figure of merit and low intrinsic thermal conductivity indicate that they are good starting materials in the development of novel high temperature thermoelectric materials. It is noteworthy that the properties of boron carbide (Prod. No. 378119), a classical thermoelectric material, can be significantly improved by controlling its carbon-boron composition.12 Doping with transition metals is another way to efficiently modify the properties of REB44Si2. The transition metal atoms occupy voids between the boron clusters in the crystal lattice that, as indicated by our preliminary results, significantly increases the figure of merit of the doped material. Such doping efforts should be pursued further.

Control of the material’s morphology is another avenue, which is being pursued since REB44Si2 compounds possess highly anisotropic crystal structures. The preparation of well-aligned materials can offer a powerful way to modify their thermoelectric properties.

Another advantage of REB44Si2 over boron carbide is its relatively low melting point (2200 K versus 2700 K), which allows easier processing. This, in conjunction with the properties previously mentioned, suggests that REB44Si2 can be an alternative to boron carbide as a high temperature p-type thermoelectric material, with the potential for further improvement through composition modification and doping.

RE-B-C(N) Borocarbides—Long Awaited n-Type Counterparts to Boron Carbide

The discovery of n-type RE-B-C(N) compounds is generating substantial interest due to their unique potential to serve as a high temperature thermoelectric counterpart to p-type boron carbide,13,14 since typical thermoelectric applications require both p- and n-type legs.

RE-B-C(N) compounds have a layered structure along the c-axis with B12 icosahedra and C-B-C chain layers residing between the B6 octahedral and the rare earth atomic layers (Figure 3). The number of interlayer B12 icosahedra and C-B-C chain layers increases successively along the series REB17CN, REB22C2N, and REB28.5C4. As the number of C-B-C layers approaches infinity (that is, no rare earth-containing layers), the compound is analogous to boron carbide.

Figure 3. View of the structure of the homologous RE-B-C(N) compounds along the c-axis.

This similarity of RE-B-C(N) structural blocks to B4C is the reason for considering them as compatible n-type counterparts to boron carbide, which has been established as an excellent p-type high temperature thermoelectric material.

One of the greatest challenges and opportunities for the RE-B-C(N) series is the development of an effective compaction procedure. Hot or cold pressing yields materials with densities of only about 50% of the theoretical value.13 Initial attempts to use spark plasma sintering (SPS) yielded an improvement to about 70% of the theoretical value, which is still rather low.14 The increase in density from 50% to 70% resulted in an increase in the figure of merit of close to two orders of magnitude. Therefore, the further development of a densification procedure is highly desirable.

An interesting way to control the morphology of RE-B-C(N) materials is by seeding them with a few percent of metallic borides (such as REB6).14 While seeding does not impact densities, the Seebeck coefficients and the thermal conductivities of doped materials, the resistivities of these materials are reduced by up to two orders of magnitude, offering an effective way to increase the material’s figure of merit. Since the percent of doping is small, it is unlikely that a percolation effect, where the metallic boride particles create channels of high electrical conductivity, is the reason for the observed effect. As seen in Figure 4, the metal-seeded sample has larger grain sizes than the undoped sample, indicating that the addition of metallic borides promotes grain growth in RE-B-C(N). We note that compared with conventional thermoelectric materials, in which it is usually preferred to inhibit grain growth in order to depress thermal conductivity, boron cluster compounds possess low intrinsic thermal conductivity, and therefore grain growth is beneficial. Thus, the seeding provides an efficient tool to control the morphology of the samples and to improve their thermoelectric properties.

Figure 4. SEM microstructures of (a) YB22C2N and (b) ErB22C2N:ErB4.

RE-B-C(N) borocarbides demonstrate high anisotropy of the crystal structure (Figure 3), making further control of the nano- and microstructure of the samples a powerful tool for the enhancement, or tuning, of their thermoelectric properties. Further investigations should continue into the processing and densification of this system. It is hoped that this system will be a viable n-type counterpart to boron carbide.

back to top 


Boron-based compounds are promising systems due to their high temperature thermoelectric properties. They are high temperature materials (typically possessing melting points above 2200 K), non-toxic, lightweight, and show remarkable stability in corrosive or acidic conditions. Two novel boron-based compounds are especially promising. REB44Si2 is a p-type high temperature compound that retains the low thermal conductivity exhibited by well known boron cluster compounds, such as REB66, while improving the electrical properties and thermoelectric figure of merit. REB44Si2 can be more readily melted and therefore might be advantageous in processing when compared to boron carbide. The rare earth borocarbides, REB17CN, REB22C2N, and REB28.5C4, represent the first instance of intrinsic n-type behavior in boron cluster compounds, and are interesting as a highly compatible counterpart to boron carbide. Both compound systems have recently been discovered and further development should prove rewarding, since the benefits of finding materials able to viably convert high temperature waste heat to electricity are immense.

Research into the optimization and processing of these compounds has just begun. Composition control, carbon doping, and doping the voids of the clusters should greatly enhance the materials’ thermoelectric properties. Optimizing densification processes for the borocarbides and ways to control the nano- and microstructure of both anisotropic compounds will be valuable tools in the further development of materials for high temperature thermoelectric applications.

back to top 




  1. Chemistry, Physics and Materials Science of Thermoelectric Materials: Beyond Bismuth Telluride, Kanatzidis, M..G.; Mahanti, S.D.; Hogan, T.P., Ed.; Kluwer: New York, 2003.
  2. Mori, T. Handbook on the Physics and Chemistry of Rare Earths, Gschneidner Jr, K.A.; Bunzli, J.C.; Pecharsky, V., Eds.; Elsevier:Amsterdam, 2008; Vol. 38; p 105.
  3. Mori, T.; Martin, J.; Nolas, G. J. Appl. Phys. 2007, 102, 073510.
  4. Cahill, D.G.; Fischer, H.E.; Watson, S.K.; Pohl, R.O.; Slack, G.A. Phys. Rev. B, 1989, 40, 3254.
  5. Nagamatsu, J.; Nakagawa, N.; Muranaka, T.; Zenitani, Y.; Akimitsu, J.; Nature, 2001, 410, 63.
  6. Ekimov, E. A.; Sidorov V.A.; Bauer E.D.; Mel’nik, N.N.; Curro N.J.; Thompson J.D.; Stishov, S.M.; Nature, 2004, 428, 542
  7. Oganov, A.R.; Chen, J.; Gatti, C.; Ma, Y.; Glass, C.W.; Liu, Z.; Yu, T.; Kurakevych, O.O.; Solozhenko, V.L.; Nature 2009, 457, 863.
  8. Mori, T.; Tanaka, T. J. Phys. Soc. Jpn., 1999, 68, 2033.
  9. Mori, T. Z. Krist., 2006, 221, 464.
  10. Mori, T. J. Appl. Phys., 2005, 97, 093703.
  11. Mori, T. Physica B, 2006, 383, 120.
  12. Wood, C.; Emin, D. Phys. Rev. B, 1984, 29, 4582.
  13. Mori, T.; Nishimura, T. J. Solid State Chem., 2006, 179, 2908.
  14. Mori, T.; Nishimura, T.; Yamaura, K.; Takayama-Muromachi, E. J. Appl. Phys. 2007, 101, 093714.

back to top 

Related Links