- You are here:
- Startseite
- About
- Map of International Research Activities (MIRA)
- Detail
Back
Back
Author |
|
| Name | Ramakers, Senja |
| Research field | computational material science |
| Career stage | doctoral researcher |
| Home university/institution | Ruhr-Universität Bochum (RUB) |
| Department/Research unit at home university/institution | Physics and Astronomy |
| Chair/Working group at home institution | Ralf Drautz |
International activity |
|
| Country | United States |
| Location | Cambridge, MA |
| University | Harvard University |
| Fund Research School | PR.INT |
| Type of activity | research stay |
| Period |
starts 14-03-2022 ends 24-06-2022 |
| Keywords | computational material science, silicon carbide, molecular dynamics, machine learning, crystal growth |
| Report | The PR.INT program of the RUB Research School allowed me to visit Harvard University in Cambridge, MA for three months and work in close collaboration with a professor Kozinsky and doctoral researcher Yu Xie. During my research stay, we developed an interatomic potential to model the epitaxial growth of silicon carbide (SiC). SiC is a wide bandgap semiconductor with promising material properties. Its high breakdown voltage and low on-state resistance make it an attractive candidate to outperform Si-based semiconductor technologies. One of the main challenges in the field is the growth of low-defect SiC wafers. To study the defect nucleation process we use atomistic simulations. The standard method, density functional theory (DFT), is accurate but very expensive, which imposes limits to the length and time scale of simulations. We use machine learning methods to train Gaussian processes and neural network on DFT data to achieve highly accurate but orders of magnitude faster models. The developed models allow us to gain more insight in the crystal growth on the atomistic scale. |