| The positions of -[Al-O(H)-Si]-Bronsted acid sites in zeolites have important implications on their performance as catalysts in terms of their accessibility and their effective acidity. We have modelled a range of zeolites with different frameworks and compositions in order to identify the most probable acid site distributions and their properties. In order to obtain realistic distributions, acid site formation must be regarded as a two-stage process - aluminium incorporation into the framework followed by bridging site protonation. Al distribution stability depends on individual Al substitution energies (influenced by disruption to the lattice), interactions between Al (minimised for larger Al...Al distances and ordered arrangements), and complex interactions between Al and cations (dependent on cation species and charge). Thus, we predict that ordered Al3, Al8 and Al12 distributions are most probable in TPA-ZSM-5 with Si/Al = 23. Likewise, the most stable Al distribution modelled in Na-Mordenite with Si/Al = 5 is an ordered distribution with occupation of the T1 and T4 sites double that of the T2 and T3 sites. Verification of these Al distributions is limited by the lack of experimental data apart from those estimated from X-ray diffraction T-O bond lengths, which we found to include a large error. Protonation of bridging sites with smaller T-O-T bond angles is favoured, and is also influenced by complex interactions between acid sites dependent on their separation and relative orientations. In general, the more stable acid sites are less acidic. Acidity was measured predominantly by the acid site OH vibrational frequencies, the major influences on which were the magnitude of the electric field, and related O-H bond length. The lack of correlation between the OH vibrational frequency and the size of ring or pore into which it points suggests that deconvolution of the experimental OH peak into low-frequency main channel and high-frequency side-pocket contributions may be over-simplified. |
