Speaker: Prof. Cristiano Francisco Woellner, professer from the Federal University of Paraná

Title: Designing Nanomaterials with Predictive AI Molecular Dynamics

Time: 9:30 September 4th, 2025 (Thursday)

Venue: Conference room on the 6th floor, Scientific Research Building

Host: Prof. Huiqiong Zhou

Info. of Speaker：

Cristiano Francisco Woellner received his B.Sc. (2004), M.Sc. (2006), and Ph.D. (2010) degrees in Physics from the Federal University of Paraná (UFPR), Brazil. He carried out postdoctoral research in the United States at the University of California, Santa Barbara (2012–2013) and Rice University (2016–2017), and in Brazil at the University of Campinas – UNICAMP (2014–2016 and 2017–2018). He is currently an Associate Professor in the Department of Physics at UFPR. His research focuses on Condensed Matter Physics, with emphasis on Nanotechnology and Computational Simulation. He works with classical and reactive molecular dynamics, first-principles calculations (DFT), Monte Carlo methods, and machine learning approaches for modeling a variety of properties of materials at the nanoscale. He has particular interest in adsorption, transport, and mechanical properties of two-dimensional materials and complex structures such as schwarzites, graphene oxides, and high-entropy alloys, as well as in the modeling of organic solar cells. He is also involved in the development of artificial intelligence–based interatomic potentials and active learning strategies to accelerate computational materials discovery.

Abstract:

Graphynes (GYs) and graphdiynes (GDYs) are two-dimensional carbon allotropes that integrate sp and sp² hybridized atoms, offering structural motifs distinct from graphene. Using equilibrium molecular dynamics (MD) simulations combined with density functional theory (DFT) calculations, we benchmark the electronic and thermal properties of α-GY and α-GDY against graphene. Traditional MD results reveal that the incorporation of acetylenic groups lowers the thermal conductivity of α-GY and α-GDY by nearly an order of magnitude compared to graphene, due to reduced atomic density, weaker sp bonds, and flattened phonon dispersions. When configured as nanoscrolls, these materials exhibit an even further reduction in thermal transport, while maintaining a zero-band-gap semiconducting nature. To extend beyond conventional MD, we introduce a predictive AI-driven molecular dynamics framework and apply it to graphene as a test case. The AI approach captures thermal and vibrational properties with high accuracy, demonstrating its potential to accelerate nanomaterials design by combining predictive capability with computational efficiency. Together, these results establish a comparative baseline for graphyne-based systems while showcasing the advantages of AI-enhanced molecular simulations for next-generation materials discovery.