Ca. Hijar et al., The siting of Ti in TS-1 is non-random. Powder neutron diffraction studiesand theoretical calculations of TS-1 and FeS-1, J PHYS CH B, 104(51), 2000, pp. 12157-12164
We present the first direct evidence of non random siting of Ti and Fe in T
S-1 and FeS-1, nanoporous metallosilicate selective oxidation catalysts of
MFI topology. This was accomplished by using Rietveld analysis of powder ne
utron diffraction data and exploiting the differences in neutron scattering
lengths between Ti or Fe and Si. Previous spectroscopic, X-ray diffraction
, and computational approaches have suggested a random substitution of Ti a
nd Fe ions among the 12 crystallographically distinct Si sites in the frame
work of TS-1 and FeS-1. In contrast, our results indicate that titanium is
distributed among only 4 or 5 of the 12 silicon sites with Ti occupying T3,
-7, -8, -10, and -12. Of the 2.47 total Ti atoms per unit cell the Ti site
occupancies and estimated standard deviation for sample B are as follows:
T3 0.30(0.11), T7 0.34(0.14), TX 0.92(0.10), T10 0.41(0.14), and T12 0.50(0
.14). In FeS-1 synthesized with 1.5 Fe atoms per unit cell, iron is found o
nly at T8. Several starting models were chosen for initial refinement, and
each returned the same specific, nonrandom distribution of Ti in the framew
ork of MFI. We have examined several computational approaches that involve
thermodynamic arguments to rationalize the experimental observations, and a
ll have failed to predict the experimentally observed substitution pattern,
This suggests that the kinetics of framework formation may play a role in
directing the observed metal substitution.