Small titanium dioxide nanoparticles feature complex potential energy surfaces with multiple local minima, and, consequently, exploration of these potential energy surfaces becomes a nontrivial global optimization task requiring significant computational resources. To reliably identify the lowest energy conformers accessible at room temperature requires combining an accurate electronic structure method with an efficient potential energy surface traversing algorithm. So far, due to computational cost, a rigorous search for low-lying conformers has been carried out only for (TiO2)n, n=1-15, nanoparticles using various density functional theory methods. In this talk, we present a detailed exploration of potential energy surfaces of larger (TiO2)n, n=16-60, nanoparticles using GFN2-xTB method, which is computationally approximately ten-fold less expensive than conventional DFT methods. The efficient global optimization algorithm is a key to identify the lowest conformers of nanoparticles. To enable our study of potential energy surfaces of (TiO2)n, n=16-60, nanoparticle, we developed a stochastic algorithm identifying the lowest energy conformers, which combines the tree growth scheme for building nanoparticle structures from a given set of seeds with the robust filter for selecting viable nanoparticle structures based on common neighbors analysis. In most cases, the lowest energy conformers can be found only from several thousands of preoptimized conformer candidates. The developed algorithm enabled us to reduce the number of conformer candidates required to identify the lowest energy conformers of nanoparticles. Applying our global optimization algorithm to (TiO2)n, n=16-60, nanoparticles, we identified these nanoparticles' putative lowest energy conformers and carried out a detailed analysis of their geometrical structure evolution with the increasing size of the nanoparticle. Furthermore, application of common neighbors analysis to individual conformers of nanoparticles allowed us to trace the formation of rutile-like cores in (TiO2)n, n=43-60, nanoparticles and to identify other repeating bonding patterns in smaller titanium dioxide nanoparticles. Based on these results, in this talk, we present a geometrical structure evolution map for (TiO2)n, n=16-60, with clearly identifiable structural motives common for most nanoparticles.

Structural evolution of atomically precise titanium dioxide nanoparticles: GFN2-xTB exploration of (TiO2)n, n=16-60, potential energy surfaces

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.