抽象的

Toluene oxidation on metal-oxide catalysts: Theoretical modeling

Vjeran Gomzi

  

 Removal of reactive organic species from the environment becomes a standard in modern society. In many reactions catalytic centers are used to enable reactions which would otherwise be energetically demanding. Apart from the ac­tivity of the catalyst, the material availability and the cost of production of the catalyst should also be taken into account. Here, several relatively inexpensive materials are investigated as potential catalysts for toluene degradation.

Interaction of toluene in oxygen atmosphere has been in­vestigated by modeling the most abundant fraction in sev­eral powder catalysts as investigated in toluene degrada­tion experiments. The repeated unit cell has been formed from MnO2, Mn2O3, Fe2O3, NiO and CuO crystal slab consisting of approximately 1000 atoms from crystal sur­rounded by vacuum layer to which three or four toluene and oxygen molecules were added. For calculation, the re­active force-field method was used as implemented in the ReaxFF code developed and by Duin et al [1]. Although a semiempirical method, this code has shown considerable success in modeling binding, due to the implemented bond order changes depending on the interatomic dis­tances [2]. Temperature controlled Berendsen thermostat (NVT) with damping constant of 0.1 fs is used for 2.5fs molecular dynamics (MD) calculation on the initially en­ergy minimized crystal structure at 500K. Toluene is then added and additional 25ns

MD calculation is performed (100000 steps of 0.25fs) at 500K. Additionally, calculations involving temperature rise from 500K to 700K, followed by 6.25 fs calculation at 700K and then temperature drop back to 500K were per­formed keeping all the other parameters as in constant temperature calculations. These were done in an attempt to model the much longer time available for reactions under experimental conditions more realistically. Unfor­tunately the experimental times of the order of 1s are not accessible to the MD model due to time and size limita­tions.