Institute of Nuclear Physics,
Polish Academy of Sciences, Cracow, Poland · 2025
Read Institute of Nuclear Physics, by Polish Academy of Sciences, Cracow, Poland (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-11008.txt.
Introduction
Paweł Jochym and Jan Łażewski compute heat flow in cubic silicon carbide in this 2025 paper. Silicon carbide is a hard ceramic used in power electronics, where heat must escape fast. Its thermal conductivity depends on how lattice vibrations carry energy.
They built an atomic interaction model from displacement and force data. A sampling method called HECSS gathered those data with density functional theory forces. A new weighting scheme in the sampler let them use every computed sample with proper weight. That raised efficiency a lot. Their anharmonic phonon results matched published experiments closely. The match supports the sampling change as a sound upgrade.
Catalog record
| CREATOR | Polish Academy of Sciences, Cracow, Poland |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-11008
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Institute of Nuclear Physics, Polish Academy of Sciences, Cracow, Poland Abstract Thermal conductivity of 3C-SiC from configuration space sampling Paweł T. Jochym \star and Jan Łażewski Abstract Cubic silicon carbide phonon thermal conductivity has been calculated using anharmonic phonon analysis. The atomic interaction model was built using displacement-force data obtained with the High Efficiency Configuration Space Sampling (HECSS) technique and density functional theory calculated forces. In the new version of HECSS we replaced the Markov chain scheme of Metropolis-Hastings Monte-Carlo with weighting of the final sampling according to the target distribution. This increased the efficiency of the method and allowed to use – with appropriate weight – all generated and ab-initio evaluated samples. The quality of the proposed method is confirmed by the accuracy with which the experimental results taken from the literature were reproduced. Copyright attribution to authors. This work is a submission to SciPost Physics. License information to appear upon publication. Publication information to appear upon publication. Received Date Accepted Date Published Date 1 Introduction Silicon Carbide (SiC) can crystallize in the rich family of polytypes. It is quite exceptional, having more than 250 possible structures identified [ 1 , 2 , 3 ] . The stoichiometry of all polytypes is the same (50:50 proportion of Si and C), the only difference is the stacking order and arrangement of consecutive atomic layers. This results in cubic, hexagonal, or rhombohedral structures as well as lower symmetry polytypes with very large lattice vectors (even above 300 nm [ 2 ] ). Despite this large spectrum of structures, only hexagonal (4H-SiC, 6H-SiC) and cubic (3C-SiC) are commercially available. The structure determines the properties of the material, but since the composition of all polytypes is the same, they all fall in the same ballpark. All structures are classified as wide band gap semiconductors, with a gap energy of 2.3-3.2 eV, which is much higher than silicon. At the same time, they have a large breakdown field (200-400 MV/m). These properties result in low on-state resistance and, simultaneously, low leakage current. Combined with high saturated electron velocity, which is important for high-frequency behavior, and thermal conductivity approximately three times higher than silicon all these properties make silicon carbide a promising material for all kinds of industrial applications, particularly for devices working in high-temperature environments, and potential alternative to currently used semiconductors (e.g. gallium nitride). What is more, SiC is also known to be bio-compatible [ 4 , 5 ] and radiation resistant [ 6 ] . The material is not without some drawbacks, the major one being the difficulty of growing large-volumes good-quality single crystals, similarly to silicon monocrystals. Although one can grow 3C-SiC epilayers on silicon wafers investigated and developed starting from the early 1980s [ 7 , 8 , 9 , 10 , 11 , 12 , 13 ] , they are still of relatively low crystallographic quality, which limits their applications and favors more expensive hexagonal bulk SiC wafers in the fabrication of devices requiring larger volume and/or higher crystal quality. The electrical properties mentioned above make silicon carbide a promising material for the production of high-power, high-temperature devices. The application range is, however, not limited to these fields. The electro-optical properties of this material, including large second-order ( \chi^{2} ) susceptibility [ 14 ] and high third-order non-linearity ( \chi^{3} ) [ 15 , 16 ] , make it also valuable for optical frequency conversion [ 17 ] devices. Particularly, due to its aforementioned favorable properties as well as high chemical resistance [ 18 ] , hardness ( E\approx 450 GPa), and high thermal conductivity [ 19 ] . This last property plays a significant role in high-power applications since it helps to quickly remove the dissipated heat from the device using the substrate as a heat conductor instead of elaborated heat transport solutions (e.g. bonded copper heat spreaders which need to be electrically insulated from the device, while being in good thermal contact with the heat-producing active parts). The ability to accurately predict thermal conductivity for a novel material or structure using ab-initio computation has been one of the goals for computational material science. While the theoretical framework of heat transport in crystalline solids is well understood [ 20 , 21 ] , its practical application encounters numerous difficulties. Firstly, heat transport is connected with phonon interactions and energy exchange between vibration modes. Thus, it is an intrinsically anharmonic process – since harmonic phonons do not interact. Secondly, the interaction model used to construct lattice dynamics should correspond to a physically realistic state of the system, which is a thermodynamic equilibrium at a given temperature. Thus, we need to build an anharmonic interaction model for the system. Such a model requires a large set of parameters – interatomic force constants (IFCs), which are often obtained from the molecular dynamics [ 22 , 23 , 24 ] . This approach, while quite expensive computationally, simultaneously solves the second of the above difficulties by deriving data for the interaction model from the representation of the thermal equilibrium provided by the molecular dynamics. In our work, we have used a similar approach of constructing the interaction model from the representation of the thermal equilibrium state, but we have replaced the molecular dynamics with the recently introduced HECSS method [ 25 ] which samples the thermodynamic ensemble with high efficiency. The anharmonic lattice dynamics model is then used to calculate the temperature-dependent heat conduction coefficient for the finite 3C-SiC crystal. 2 Calculation methods The determination of the thermal conductivity of the crystal necessitates the development of an anharmonic lattice dynamics model that corresponds to the thermal equilibrium. Traditionally, this is done with the displacement-force data set generated with ab-initio molecular dynamics [ 23 ] . This method is computationally expensive and quite wasteful – since one throws away most of the data computed along the trajectory. Instead, we have used the thermodynamic ensemble sampling technique (HECSS) based on the generalized equipartition theorem. The details of the approach are described in our earlier work [ 25 ] . This approach was already successfully used to investigate negative thermal expansion [ 26 ] , structural phase transitions [ 27 , 28 ] , orbital order [ 29 ] and chiral phonons [ 30 ] . These results demonstrate that it is possible to effectively generate samples providing an appropriate representation of the system in thermal equilibrium at a given temperature and build an adequate lattice dynamics model for the material extending to the analysis of negative thermal expansion effects [ 26 ] . Here, we present a series of calculations for the 3C-SiC using VASP as a source of energies and forces, using a range of the system sizes (from 2^{3} to 5^{3} supercell) and standard PAW-PBE calculation setup [ 31 , 32 , 33 , 34 ] . The aim of our work is to calculate phononic thermal conductivity based on anharmonic lattice dynamics derived from the DFT data. This goal requires substantially more data points, comparing to standard lattice dynamics calculation, due to the large number of independent interatomic force constants (IFCs) of the third and fourth order required to derive an anharmonic lattice dynamics model. The DFT calculations used standard PAW-PBE