Theoretical Studies On Sodium Cation Location In ZSM-5 By Cluster Model And ONIOM Methods

Extra-framework metal cations in zeolites play important roles in catalytic processes. However, accurate coordination structures of ion adsorption can hardly be obtained by means of experimental techniques. In complements, reliable theoretical calculations can provide structural information of different adsorption modes, as well as the distribution of different adsorption sites. In this work, we perform a theoretical investigation on sodium cation location in ZSM-5 by using traditional cluster models as well as various ONIOM methods. The target of present study is to obtain cost-effective model which can achieve both reliable structures and energetics. The main results are summarized as follows:1) A number of cluster models in a range of 3T to 192T were investigated at the level of B3LYP/6-31G(d). Our calculation shows that when cluster models are no larger than 33T, either interaction energies or substitution energies of all the four sites are far from convergence. When the cluster size increases to 75T, the substitution energies converged to those of even bigger clusters, such as 128T and 192T. This indicated that 75T clusters are good enough to describe the interaction of metal ion with the Al substituted zeolite.2) To explore the effect of dividing schemes of ONIOM, different high-level layers, varied from 3T to 33T, have been tested. We show that the success of ONIOM calculations are strongly dependent on the choice of high-level layer. Our calculations demonstrated that if the high-level layer only included the first coordination circle (Type I), ONIOM method fails to give reasonable adsorption structures and energetics. This finding downplays the significance of previous ONIOM studies where the sizes of high-level layers were usually less than 6T.3) We find that ONIOM model with a moderate high-level layer can achieve reliable structures. Specially, the MAD of Na-0 distances in 33T@75T is less than 0.04A as compared with those in 75T. Our calculation also showed that the convergence of energetics for ONIOM method is slower than that of geometry optimization such that even 33T@75T gave an erroneous stability sequence of different ion sites. Fortunately, a single-point correction by using B3LYP/6-31G(d) could attain reasonable results.4) When properly used, ONIOM is able to provide accurate results with good account of the environment effect and at the same time affordable computational costs. In our calculations, ONIOM method needs only 1/10 to 1/5 computational time for each SCF (self-consistent field) iteration as compared to the full 75T model calculation, highlighting its pivot role in zeolite simulation.