Exploring Graphene Oxide with Machine Learning
Author Information
Author(s): Z. El-Machachi, D. Frantzov, A. Nijamudheen, T. Zarrouk, M. A. Caro, V. L. Deringer
Primary Institution: University of Oxford
Hypothesis
Can machine learning methods accelerate the exploration of the chemical and configurational space of graphene oxide?
Conclusion
The study demonstrates that machine learning can effectively model and predict the structural transformations of graphene oxide during thermal reduction.
Supporting Evidence
- The study provides a platform for routine, accurate simulations of carbonaceous materials.
- Machine learning potentials were fitted to accelerate the exploration of functional groups in graphene oxide.
- Simulations revealed how functional groups evolve during thermal reduction, correlating with experimental findings.
Takeaway
This study shows how computers can help scientists understand the tiny details of materials like graphene oxide, which is important for making new technologies.
Methodology
The study combines on-the-fly accelerated first-principles molecular dynamics with message-passing neural-network potentials to explore the structures of graphene oxide.
Limitations
The study is limited by the timescales accessible to molecular dynamics simulations, which are on the order of nanoseconds.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website