Unraveling Fluoride As Structure-Directing Role on Zeolite Self-Assembly Via Integrated DFT and Monte Carlo Simulation
Zeolites – nanoporous aluminosilicate materials – are highly valued for their exceptional properties, including high surface area, uniform pore sizes, acidity, and ion exchange capacities. These properties make zeolite widely used in catalysis, separations, and storage.1 Developing tailor-made zeolites is crucial for sustainable solutions. However, achieving precise control over zeolite synthesis remains challenging due to the complex properties of the sol-gel systems from which zeolite crystals are born. One critical factor influencing zeolite formation is the choice of structure-directing agents (SDAs), which are often organic cations and fluoride used in tandem.2 Understanding the interplay among sol-gel silica, fluoride, and organic SDAs is essential for elucidating the zeolite self-assembly pathway. In this project, I investigated a new Monte Carlo algorithm informed by DFT calculations predicting how fluoride acts as a structure-directing agent during zeolite formation.
A previous member of the Auerbach group developed a Reactive Ensemble Monte Carlo (RxMC, Fig. 1a) simulation approach3 for modeling the formation of zeolites in the presence of organic SDAs, represented as hard spheres (see Figs. 1c and 1d).4,5 The inclusion of organic SDAs successfully produced the zeolite framework, Linde Type A (LTA, Fig. 1b), even within a simplified coarse-grained representation.6 However, the present RxMC approach lacks the inclusion of fluoride anions, which are crucial for the self-assembly of all-silica zeolites. Fluoride is thought to be especially important in the formation of the double-4-ring (D4R) building block, the yellow connecting cubes in Fig. 1b. Thus, although the results in Ref. 4-6 are interesting as base-case models, they must be improved by including fluoride to be considered realistic representations of the sol-gel chemistry of silica during zeolite nucleation.
Flanigen and Patton introduced the hydrothermal synthesis of microporous silica structures using fluoride as a mineralizing agent rather than hydroxide anions.7 Compared to synthesis with hydroxide, synthesis with fluoride reduces the concentration of connectivity defects in pure-Si zeolite formation.8 The charge balance achieved by fluoride eliminates the need for Si-O- charge defects. Fluoride has also been suggested as a structure-directing agent stabilizing small silicate rings and cages such as the D4R.9,10 Several studies have investigated the local environment of fluoride anions residing in cages, finding them to be pentacoordinate with silicon atoms11,12,13 or located at cage centers.14 Despite the number of studies of fluoride stabilization and conformation in the framework, little is known about the detailed mechanism of the fluoride as a structure-directing agent during pre-crystallization.
To accurately simulate the role of fluoride as a structure-directing agent, we will conduct a multi-step approach that integrates DFT energy calculations with RxMC simulations. The diversity of fluoride-silica structures requires the accuracy of DFT to make useful predictions. However, a novel multiscale method is required because RxMC simulations often require on the order of 10 million steps to yield crystals; performing that many DFT calculations is simply not feasible.
Initially, we will perform gas-phase electronic energy calculation to generate an energy dataset detailing the evolution of silica clusters with and without fluoride, leading up to the D4R building block. The energy dataset will be crucial for refining our RxMC simulation by enabling the calculation of acceptance probabilities to accurately captures the energetic landscape of fluoride-mediated zeolite assembly. Finally, we will employ various spectroscopic analyses (NMR, Raman) to compare the simulated structural features with experimental data, providing a comprehensive assessment of the accuracy and predictive power of the simulation.
The stabilization energy by fluoride within the different silica cluster structures will be computed using the Gaussian software package. We will generate an energy landscape detailing the formation of the D4R structure with fluoride. Within the various structures, we will accommodate fluoride anions, both targeting pentacoordinate to different silica atoms within the cage, to assess the preferential location of fluoride for optimal stabilization. By systematically analyzing these factors, including the explicit consideration of fluoride outside the cluster to assess its potential influence on the silica cluster, we aim to identify the most stable configurations of fluoride included in the silica cluster on the potential energy surface and fluoride preferential location and coordinated silicon framework atom within the silica cluster. The additional bias energy allows a more systematically accurate representation of the system's behavior and the likelihood of specific reaction pathways during the simulation.
The effectiveness of our novel additive feature is assessed rigorously by focusing on the reproduction of Qn evolution, where Qn represents the number of bridging oxygen atoms connected to a silicon atom. The Qn evolution is crucial for representing the RxMC simulation, which is used to experimentally study the kinetic of silica polymerization.15,16 Successful reproduction of Qn evolution is a critical validation step, confirming the capability of the RxMC simulation to effectively model the silica polymerization without colliding with fluoride feature in the algorithm. Eventually, we will reach the zeolite framework formation, where the LTA framework is chosen since it was performed recently using the RxMC simulation.6 In the end, we will implement various spectroscopy, including X-ray diffraction, Raman, and NMR, upon forming the zeolite framework. By comparing the simulated local environment with experimental data, we can further test the accuracy of our approach for capturing the behavior of fluoride ions within the zeolite framework. The current project can shed light on the role of F-structure directing during zeolite formation.
Figure 1: (a) Snapshots of simulation cells from RxMC showing both amorphous system (left) and fully crystallized system (right), (b) LTA zeolite framework; (c) LTA alpha-cage with two organic SDAs inside, modeled as a single hard-sphere 10 Å in diameter, and (d) LTA beta-cage with one organic SDA inside, modeled as a hard-sphere 6 Å in diameter (adapted from Bores et al.6). (e) Illustration of two possible pathways from 3 Si clusters (left) to 5 Si clusters (right). The fluoride anion resides in one of the silica framework atoms and is pentacoordinate. For simplicity, the hydrolyzed water molecule and silicic acid, Si(OH)4, are not illustrated in the figure.
Reference:
