First-Principles Determination of Ultrahigh Thermal Conductivity of Boron Arsenide: A Competitor for Diamond?

L. Lindsay, D. A. Broido, T. L. Reinecke
2013 Physical Review Letters  
We have calculated the thermal conductivities (κ) of cubic III-V boron compounds using a predictive first principles approach. Boron Arsenide (BAs) is found to have a remarkable room temperature κ over 2000Wm -1 K -1 ; this is comparable to those in diamond and graphite, which are the highest bulk values known. We trace this behavior in BAs to an interplay of certain basic vibrational properties that lie outside of the conventional guidelines in searching for high κ materials. We also find that
more » ... . We also find that cubic BN and BSb will have high κ with isotopic purification. This work provides new insight into the nature of thermal transport at a quantitative level and predicts a new ultra-high κ material of potential interest for passive cooling applications., As microelectronic devices become smaller, faster and more powerful, thermal management is becoming a critical challenge in, e.g., microprocessors, LEDs and high power RF devices. As a result, the ability to identify and understand materials with high thermal conductivities (κ) is becoming increasingly important [1]. Carbon based materials, including diamond and graphite, have long been recognized as having the highest κ of any bulk material with room temperature (RT) values around 2000Wm -1 K -1 [2, 3] . Other high κ materials, such as copper, are not even close, having four to five times smaller values [4] . However, diamond is scarce and its synthetic fabrication suffers from slow growth rates, high cost and low quality [5] . Thus it is of current interest to identify new materials with ultra-high thermal conductivities. Commonly accepted criteria [4] to guide the search for high κ non-metallic crystals include: 1) simple crystal structure, 2) low average atomic mass, M avg , 3) strong interatomic bonding, and 4) low anharmonicity. Items 2) and 3) imply a large Debye temperature, θ D , and give the frequently used rule of thumb that κ decreases with increasing M avg and with decreasing θ D . However, little progress has been made over the years in identifying new high κ materials. Until recently fully microscopic, parameter-free computational materials techniques for electronic properties have been more advanced than are those for thermal transport. In the last few years quantitative ab initio techniques have been developed for thermal transport [6-12], and we have made contributions to this development [6, 7, 12] . These techniques open the way to a fuller understanding of the key physical features in thermal transport and to the ability to predict accurately the κ of new materials. Here, we present quantitative, first principles calculations of κ for the class of boron based cubic III-V 3 compounds, Boron Nitride (BN), Boron Phosphide (BP), Boron Arsenide (BAs) and Boron Antimonide (BSb), referred to collectively as BX compounds. We find that BAs has an exceptionally high RT κ above 2000 Wm -1 K -1 , which is comparable to that of the bulk carbon crystals. This result is surprising given that a prediction based on the commonly accepted criteria above [4] gives RT κ for BAs of only around 200Wm -1 K -1 , comparable to that of silicon.
doi:10.1103/physrevlett.111.025901 pmid:23889420 fatcat:oddmzbdh5zepblkpinsfpsbzbe