Unveiling fullerene formation and interconversion through molecular dynamics simulations with deep neural network potentials

The development and application of a deep learning potential trained on carbon clusters provide new insights into the nucleation and growth of fullerenes, expanding upon previous assumptions.