Research

Multicomponent Diffusion in Zeolites

One major goal in this area is to predict multicomponent diffusivities given single-component values. Despite the high level of activity on single-component diffusion in zeolites, multicomponent diffusion is largely unexplored territory, even though knowledge of multicomponent zeolite diffusivities is important in both separations and catalysis. This knowledge also seems crucial for understanding behavior in novel zeolitic membranes. A big problem is the dearth of experimental data on multicomponent systems.

Our strategy has been to use a combination of molecular modeling and NMR experiments, with the ultimate goal of developing engineering correlations. We performed the first molecular dynamics (MD) simulations of binary diffusion in a zeolite and found good agreement with experimental data from NMR. We have carried out extensive MD studies of n-alkane diffusion in faujasite and MD simulations providing the first dynamical evidence for `molecular traffic control' in zeolite channels, a hypothesis going back 20 years. Recently, we used MD to study mixtures in faujasite. By predicting the so-called `transport' or Fickian diffusivities from MD and then using these in traditional engineering models, we have been able to model co-diffusion, counter-diffusion, and other scenarios in faujasite membranes on macroscopic time and length scales in a completely predictive way. Our simulations allow us to directly predict the cross-term coefficients, test theories for approximating them, and assess their importance in practical applications. We find that at higher loadings the cross terms cannot be neglected.

MD simulations are a powerful tool, but they cannot access longer time scales relevant to many systems of interest. To overcome this problem, we are developing methods for simulation of infrequent-event dynamics. These methods are based on the picture from transition-state theory of molecules diffusing by overcoming energetic barriers.

In addition to modeling, we are also performing pulsed field gradient (PFG) NMR diffusion measurements. This powerful method allows the measurement of self-diffusivities of individual components in mixtures, so it is well suited for comparison with MD.

This work has been funded by the ACS Petroleum Research Fund and the National Science Foundation. Students and post-docs involved in this work are Marty Sanborn, Louis Clark, Amit Gupta, and Yoo Joong Kim.