Incongruent Melting and Phase Diagram of SiC from Machine Learning Molecular Dynamics (Part II)
Authors/Creators
-
1.
Harvard University
-
2.
Robert Bosch (United States)
Description
Silicon Carbide Machine Learning Molecular Dynamics Dataset (Part II)
This dataset resolves the controversy regarding silicon carbide (SiC)'s high-temperature and high-pressure phase behavior through machine learning molecular dynamics simulations (MLMD). We provide MLMD trajectories demonstrating SiC's incongruent melting behavior through two-phase simulations under varied conditions:
- Pristine trajectories at 10 and 90 GPa (No_Defect files)
- Low defect concentration (1/1000) trajectories at varied pressures (Defect_16 files)
- High defect concentration (1/100) trajectories at varied pressures (Defect_160 files)
- FLARE on-the-fly (OTF) active learning MD trajectories and DFT data.
Our approach addresses the challenges of investigating SiC under extreme conditions critical for applications in power electronics, nuclear technology, and quantum computing. Additional MLMD trajectories with different sizes and conditions are available in https://doi.org/10.5281/zenodo.14648292. The following table refers to figures in our research paper arXiv:2502.17804v1.
| Dataset | Description | Related Figure |
Defect_16_script.zip |
LAMMPS scripts for two-phase simulation (2 ns) with defect concentration 1/1000 | |
|
|
Two-phase 16K-atom simulations with defect concentration 1/1000 starting with B3 zinc blende phase at 10, 30, 60, 90 GPa | Supplementary Figure 7 |
|
|
LAMMPS scripts for two-phase simulation (2 ns) with defect concentration 1/100 | |
|
|
Two-phase 16K-atom simulations with defect concentration 1/100 starting with B3 zinc blende phase at 10, 30, 60, 90 GPa | Supplementary Figure 7 |
|
|
Two-phase 16K-atom simulations starting with B3 zinc blende phase at 10 GPa | Supplementary Figure 5, 6, 7 |
|
|
Two-phase 16K-atom simulations starting with B3 zinc blende phase at 90 GPa | Supplementary Figure 7 |
|
|
FLARE active learning MD trajectories that collect high-uncertainty MD snapshots with density functional theory (DFT) energies and forces | Figure 1 |
Files
Defect_16_P30.zip
Files
(200.2 GB)
| Name | Size | |
|---|---|---|
|
md5:46cbcbe29bcc53c363298225daf0e077
|
16.4 GB | Preview Download |
|
md5:1598ee7a1fc5c0a0684515456979207c
|
16.3 GB | Preview Download |
|
md5:03b7b107c63e218a40e02d1e26073368
|
16.3 GB | Preview Download |
|
md5:171c5d7533d82027ce3c8b3a771234ff
|
16.3 GB | Preview Download |
|
md5:c30cb66fb5bead15238286c90162d95c
|
3.1 MB | Preview Download |
|
md5:5fdae408467edf9066b62e4e771373ec
|
16.5 GB | Preview Download |
|
md5:ca69c12087a556c1e0cca6fad30faa01
|
16.5 GB | Preview Download |
|
md5:a767c0af1a37cc8d1a0bdb40a052a614
|
16.5 GB | Preview Download |
|
md5:2719cbd7a0f012ba9f98bf999f35885b
|
16.5 GB | Preview Download |
|
md5:90bbcf3819c5b8a1a331aec6a711b605
|
3.1 MB | Preview Download |
|
md5:cb7bb7e2e8e3da6c50896b403ea55a7f
|
16.4 GB | Preview Download |
|
md5:3caad44bcf6659201e8ce53519aef55d
|
16.3 GB | Preview Download |
|
md5:8062c3c90faa4e5f61de64485d3a2d54
|
6.5 GB | Preview Download |
|
md5:f25fe044563ecb5eeea621914ea8dd1c
|
4.6 GB | Preview Download |
|
md5:170effed512b2feefc5280cffa28213a
|
13.8 GB | Preview Download |
|
md5:328028648bd0482da4d259f98fef66c9
|
4.7 GB | Preview Download |
|
md5:feadf94d9db58b975da1bf7a1f1059d9
|
2.5 GB | Preview Download |
|
md5:1c11576326b7e88ebeb9562c734d24a5
|
2.5 GB | Preview Download |
|
md5:321ba7b249591ccf4a4b46da83bb9ea1
|
1.6 GB | Preview Download |
Additional details
Funding
- National Energy Research Scientific Computing Center
- National Energy Research Scientific Computing Center BES-ERCAP0024206
- United States Department of Energy
- Design & Assembly of Atomically-Precise Quantum Materials & Devices DE-SC0020128
- U.S. National Science Foundation
- Harvard University Materials Research Science and Engineering Center DMR-2011754