First Principle Kinetic Studies of Zeolite-Catalyzed Methylation Reactions

Abstract

Methylations of ethene, propene, and butene by methanol over the acidic microporous H-ZSM-5 catalyst are studied by means of state of the art computational techniques, to derive Arrhenius plots and rate constants from first principles that can directly be compared with the experimental data. For these key elementary reactions in the methanol to hydrocarbons (MTH) process, direct kinetic data became available only recently [J. Catal.2005, 224, 115-123; J. Catal.2005, 234, 385-400]. At 350 °C, apparent activation energies of 103, 69, and 45 kJ/mol and rate constants of 2.6 × 10(-4), 4.5 × 10(-3), and 1.3 × 10(-2) mol/(g h mbar) for ethene, propene, and butene were derived, giving following relative ratios for methylation k(ethene)/k(propene)/k(butene) = 1:17:50. In this work, rate constants including pre-exponential factors are calculated which give very good agreement with the experimental data: apparent activation energies of 94, 62, and 37 kJ/mol for ethene, propene, and butene are found, and relative ratios of methylation k(ethene)/k(propene)/k(butene) = 1:23:763. The entropies of gas phase alkenes are underestimated in the harmonic oscillator approximation due to the occurrence of internal rotations. These low vibrational modes were substituted by manually constructed partition functions. Overall, the absolute reaction rates can be calculated with near chemical accuracy, and qualitative trends are very well reproduced. In addition, the proposed scheme is computationally very efficient and constitutes significant progress in kinetic modeling of reactions in heterogeneous catalysis.

Complex Reaction Environments and Competing Reaction Mechanisms in Zeolite Catalysis: Insights from Advanced Molecular Dynamics

“…[6] Consequently, numerous studies focused on the methylation by methanol, dimethyl ether or methoxides of aromatics and alkenes in the framework of the MTO process. [16,17,26,42,44,46,82,83] It should be mentioned that the static approach with finite cluster models as applied in most of these studies has been very successful in reproducing and predicting experimentally measured kinetic data. [16,17] However, a similar approach is not applicable for the study of methylation reactions in the large pore AFI structured H-SSZ-24 as will be demonstrated in this case study.…”

“…[8,[15][16][17] However, theoretical chemists are still confronted with paramount challenges to thoroughly explain experimental observations. The true challenge lies in linking the model system with experimental or industrial conditions.…”

The coordinatively saturated vanadium MIL-47 as a low leaching heterogeneous catalyst in the oxidation of cyclohexene

“…In this case bimolecular transition state theory should be applied as outlined in the supporting information. More details are also given by Van Speybroeck et al on the computational procedures for discriminating between intrinsic and apparent reaction kinetics 44 .…”

Ab Initio Calculation of Rate Constants for Molecule–Surface Reactions with Chemical Accuracy

“…However, for large chemical systems, such as those encountered in heterogeneous catalysis, where adequate models typically include several hundred atoms, the “chemical accuracy” (4 kJ mol −1 for the energy barriers, one order of magnitude for the pre‐exponential factors) required for useful predictions is not attained. The exponential scaling with the system size of both the electronic structure methods (potential energy surface) and the nuclear motion problem (vibrational partition function) limits the applicable method to density functional theory (DFT) for the potential energy surface and the harmonic approximation for the vibrations 4, 5, 6, 7, 8. Depending on the functional, energy barriers may be in error by 10–20 kJ mol −1 , and the harmonic approximation is most problematic for low‐frequency modes, which are known to make the largest contribution to the partition functions 9…”

“…A deviation of a factor of ten reflects the experimental uncertainty5, 25 and is considered “accurate”. This is indicated by the gray region in the logarithmic plot of Figure 3.…”