Modellizzazione di sistemi avanzati per lo stoccaggio di litio

Nowadays, rechargeable batteries are the best developed renewable energy technology, because they are safe, cost-effective, eco-friendly, portable, with high energy and power density and good life cycle. At present the most successful rechargeable battery is the Li-ion battery (LIB), thus much of the research in the field of batteries in recent years has focused on the improvement of this technology, which is dependent on the properties of its components. This thesis work was developed within the Theoretical Chemistry Group of the University of Turin, with the aim of performing a quantum-mechanical study at the atomic scale of the processes underlying the functioning of the relevant materials for cathode. The system of interest for practical application is LiNi0.8Mn0.1Co0.1O2 (NMC-811), but before characterizing it, we focused on a structurally simpler one, namely LiNiO2 (LNO). We focused on geometry and electronic structure characterization, since they significantly influence the performance of the battery and constitute the essential starting point to build a meaningful theoretical model, and then on the determination of the energy barrier (Ea) and the diffusion coefficient (D), since it is one of the most important intrinsic physical properties of Li insertion materials for LIBs. A main challenge in the development of this thesis work has been the determination of D of Li in the investigated systems. In fact, looking at the available literature about this theme, it becomes clear that D has not yet been reliably determined, neither experimentally nor theoretically, since the published results differ of several orders of magnitude. Therefore, the main goal of this project is that of setting up a computational procedure able to provide reliable results for D of these materials, appliable also to more complex systems. In the first place, we tried to set up a procedure with the CRYSTAL program for finding Ea, fundamental to calculate D, by defining the Li path by hand. The resulting Ea for the ODH path was 0.51 eV, which is coherent with other results found in literature, but the procedure was time-consuming and the TSH path, more thermodynamically stable, resulted to be less favoured. Thus, we tried to use other methods, namely ab initio molecular dynamics (AIMD) and metadynamics (MeD), exploiting CPK program. With AIMD, we applied different values of external potential along x to make Li move. Studying the variation of D as a function of the potential, the results suggested a barrier of 0.7 eV to be overcome for having diffusion in LNO, which matches with literature results, even if the calculated values of D were at least two orders of magnitude larger than expected and the real value and the direction of the potential experienced by Li is unknown. With MeD, the main effort was the selection of the best parameters for the calculation, namely the height of spawned gaussians, the simulation time and the choice of collective variables. In the end, thanks to the reconstruction of the free energy surface, we determined the diffusion paths of Li and the presence of three minima and two energy barriers of 1.16 and 1.67 eV, which are higher than expected due to limitations of the method related to the amount of sampling. Anyway, we are rather confident that, exploiting MeD for the definition of the path and the minima and calculating the energy of the system at each point of the path, Ea will be determined more accurately, together with D.