Comprehensive DFT Study of Nitrous Oxide Decomposition over Fe-ZSM-5

Abstract

The reaction mechanism for nitrous oxide decomposition has been studied on hydrated and dehydrated mononuclear iron sites in Fe-ZSM-5 using density functional theory. In total, 46 different surface species with different spin states (spin multiplicity M(S) = 4 or 6) and 63 elementary reactions were considered. Heats of adsorption, activation barriers, reaction rates, and minimum energy pathways were determined. The approximate minimum energy pathways and transition states were calculated using the "growing string method" and a modified "dimer method". Spin surface crossing (e.g., O(2) desorption) was considered. The minimum potential energy structure on the seam of two potential energy surfaces was determined with a multiplier penalty function algorithm by Powell and approximate rates of spin surface crossings were calculated. It was found that nitrous oxide decomposition is first order with respect to nitrous oxide concentration and zero order with respect to oxygen concentration. Water impurities in the gas stream have a strong inhibiting effect. In the concentration range of 1-100 ppb, the presence of water vapor influences the surface composition and the apparent rate coefficient. This is especially relevant in the temperature range of 600-700 K where most experimental kinetic studies are performed. Apparent activation barriers determined over this temperature range vary from 28.4 (1 ppb H(2)O) to 54.8 kcal/mol (100 ppb H(2)O). These results give an explanation why different research groups and different catalyst pretreatments often result in very different activation barriers and preexponential factors. Altogether perfect agreement with experimental results could be achieved.

Experimental techniques for investigating the surface oxygen formation in the N2O decomposition on Fe-MFI zeolites

“…(2) has been favoured for a long time [8,14,15,19] in agreement with findings in a fixed bed reactor under steady state conditions [15]. However, recent studies have shown that also the O 2 desorption can determine the rate of total N 2 O decomposition [26,27], in concordance with DFT-model calculations [28]. Finally, the reaction of N 2 O with surface oxygen (Eq.…”

“…Thus, the surface oxygen amount, calculated by the N 2 formed, increases with higher reaction temperatures (Table 3). Wood et al [19] and Perez-Ramirez et al [8,26] also found this effect on ZSM5 zeolites containing 0.38% Fe (w/w) and 0.68% Fe (w/w), respectively, in the range of 500-573 K. In contrast, Kiwi-Minsker et al [16] observed no influence of the reaction temperature for zeolites containing ≤1000 ppm Fe (w/w) in the range of 523-603 K. A possible explanation could be that the temperature increases the number of active iron species in the Fe-MFI, particularly in zeolites [27,28] and by Kondratenko and Perez-Ramirez [26], or the formation of oxygen species bound to different sites. DFT calculations in Refs.…”

“…DFT calculations in Refs. [27,28] revealed that the activation energy increases if additional oxygen species are deposited through N 2 O decomposition at sites which are already occupied by surface oxygen. In this case, higher temperatures would favour this accumulation.…”

“…Notice that the surface oxygen formation was not measured independently in this study. DFT calculations by Heyden et al [28] show that the surface oxygen formation at empty sites is not activated at all (E A = 0 kJ mol −1 ), but at sites already occupied by an atomic oxygen species an activation energy of 115-130 kJ mol −1 was determined. Thus, it is possible to get higher and lower values for the activation energy than the ones presented in the present study depending on the oxidation state of the zeolite and the iron species therein.…”

“…Although investigations suggest that the recombination of atomic surface bound oxygen (Eq. (2)) [15,[19][20][21][22][23]28] or the O 2 desorption [26][27][28] are rate determining for the total N 2 O decomposition at high iron loaded zeolites, the reaction of surface oxygen with N 2 O (Eq. (3)) should not be excluded in these considerations, as shown by results in Refs.…”