Force field validation for molecular dynamics simulations of IRMOF-1 and other isoreticular zinc carboxylate coordination polymers
JA Greathouse and MD Allendorf, JOURNAL OF PHYSICAL CHEMISTRY C, 112, 5795-5802 (2008).
Molecular dynamics simulations were conducted to validate a hybrid force field for metal-organic framework-5 (IRMOF-1). In this force field, only nonbonded parameters are used to describe Zn-O interactions. The CVFF force field was used with slight modifications to describe the benzene dicarboxylate linker. The force field correctly predicts a wide range of structural properties of this MOF, including a negative thermal expansion of approximately 1% at 30 and 293 K, in agreement with both theory and experiment. Compressibility results and the associated elastic moduli are in also good agreement with published density functional theory calculations and nanoindentation experiments. The force field predicts a decrease in elastic moduli as temperature increases, which would greatly affect the mechanical properties of MOFs. Calculated vibrational frequencies for Zn-O modes agree with experiment, and a low-frequency mode representing a 180 degrees rotation of the phenyl groups is seen. This rotation becomes more prevalent as the temperature is increased from 300 to 400 K, in agreement with NMR data. Simulations were also carried out with adsorbed guests, including ethanol, cyclohexane, and several chloromethanes. It is shown that the IRMOF-1 lattice parameter depends on the nature of the guest-framework interaction; strongly hydrophilic guests, such as ethanol, cause a decrease (-0.9%) in unit cell volume, while hydrophobic guests cause an increase (0.7-1.5%) in unit cell volume. The calculated free volumes in IRMOF-1 range from 53.5% to 56.0%, in good agreement with experiment. Finally, the activation energy for benzene self-diffusion calculated at low loadings is in good agreement with previous simulations and NMR results, but the magnitude of the diffusion constant is underestimated, most likely because of deficiencies in the CVFF portion of the force field. The results demonstrate, however, that employing a rigid force field results in much poorer agreement with experimental data. Additionally, a flexible force field approach is required when simulating framework stability because of physical changes or the presence of adsorbates. The use of a general-purpose force field for the organic components allows our approach to be extended to other Zn-based frameworks.
Return to Publications page