Adsorption and diffusion of fructose in zeolite HZSM-5: Selection of models and methods for computational studies

Lei Cheng, Larry A. Curtiss, Rajeev Surendran Assary, Jeffrey Greeley, Torsten Kerber, Joachim Sauer

Research output: Contribution to journalArticlepeer-review

29 Scopus citations

Abstract

The adsorption and protonation of fructose in HZSM-5 have been studied for the assessment of models for accurate reaction energy calculations and the evaluation of molecular diffusivity. The adsorption and protonation were calculated using 2T, 5T, and 46T clusters as well as a periodic model. The results indicate that the reaction thermodynamics cannot be predicted correctly using small cluster models, such as 2T or 5T, because these small cluster models fail to represent the electrostatic effect of a zeolite cage, which provides additional stabilization to the ion pair formed upon the protonation of fructose. Structural parameters optimized using the 46T cluster model agree well with those of the full periodic model; however, the calculated reaction energies are in significant error due to the poor account of dispersion effects by density functional theory. The dispersion effects contribute -30.5 kcal/mol to the binding energy of fructose in the zeolite pore based on periodic model calculations that include dispersion interactions. The protonation of the fructose ternary carbon hydroxyl group was calculated to be exothermic by 5.5 kcal/mol with a reaction barrier of 2.9 kcal/mol using the periodic model with dispersion effects. Our results suggest that the internal diffusion of fructose in HZSM-5 is very likely to be energetically limited and only occurs at high temperature due to the large size of the molecule.

Original languageEnglish
Pages (from-to)21785-21790
Number of pages6
JournalJournal of Physical Chemistry C
Volume115
Issue number44
DOIs
StatePublished - Nov 10 2011
Externally publishedYes

Fingerprint

Dive into the research topics of 'Adsorption and diffusion of fructose in zeolite HZSM-5: Selection of models and methods for computational studies'. Together they form a unique fingerprint.

Cite this