First steps towards reproducing chemical shifts spectra of c-LiSi alloys using a semi-empirical force field

F. FERNANDEZ; D. E. BARRACO DÍAZ; M. OTERO; E. P. M. LEIVA; S. A. PAZ

Antofagasta

Workshop; 7th INTERNATIONAL WORKSHOP ON LITHIUM, INDUSTRIAL MINERALS AND ENERGY; 2020

Centro de Investigacion Avanzada del Litio y Minerales Industriales

Nuclear magnetic resonance (NMR) has proven to be a useful tool for investigating the structural changes that occur in electrode materials in electrochemistry. ​ In order to correlate Li NMR shifts with specific local environments, Key et al. acquired the 7​ ​ Li MAS NMR spectra of different crystalline phases that they synthesized by mixing stoichiometric amounts of lithium metal and silicon. Key et al. correlated the chemical shift spectrum of different c-LiSi alloys (Li​ 12​ Si​ 7​ , Li​ 7​ Si​ 3​ , Li​ 13​ Si​ 4​ , Li​ 15​ Si​ 4 and Li​ 21​ Si​ 5​ ) with the local chemical environment. The authors observed in the spectrums a shift of the center of the peak to lower frequencies when the ratio Li/Si increases. They concluded that between the Si neighbors of Li ions, those that are also bonded to other Si shift the signal to a higher frequency in comparison to those that are isolated, i.e. only bonded to other Li ions. However, by measuring the 7​ ​ Li NMR Li​ 12​ Si​ 7 spectrum at lower temperatures, Köster et al. have shown that the understanding of the chemical shift is far more complex and requires much more than two different sets of Si neighbors. In the present work, we will not consider Li​ 21​ Si​ 5 alloy in our analysis because it is metallic while the others alloys are semiconductors, which affects the NMR spectra. With the others c-LiSi alloys we hypothesize that the chemical shift contribution of each lithium ion depends only on the charge associated to this particular lithium. The respective charges are calculated using a reactive forcefield (ReaxFF) obtained by Fan et al. with the molecular dynamics package LAMMPS, using conjugate gradient of the crystalline structures of Materials Project.We find that the average charge of the lithium ions decreases when the ratio Li/Si increases, as perceived experimentally by Key et al. We associate these average values of the charge with the respective center of the peak for each alloy spectrum of Key et al. This allows us to relate the chemical shift with the charge of the lithium ions. From this relation we calculate a contribution in the chemical shift from the individual charge of each Li ion in each c-LiSi alloy, assuming a Lorentzian shape with a similar width to the experimental one. The obtained spectra for c-LiSi alloys are shown in Figure 1. We observed that the shape of the NMR spectra obtained by our simulations is in agreement with the experimental results.