Evidence of the Presence of Two Different Framework Ti(IV) Species in Ti−Silicalite-1 in Vacuo Conditions: an EXAFS and a Photoluminescence Study

Abstract

In a previous contribution (J. Phys. Chem. 1994, 98, 4125) we reported EXAFS data collected at ADONE on a high quality Ti−silicalite (TS-1) sample (Ti = 1.47 wt %) dehydrated in a carefully controlled atmosphere at 400 K. The importance of that paper was to definitively clarify that Titanium atoms occupy silicon substitutional framework positions. However, due to the large incertitude associated with the estimated first shell Ti-coordination number (N = 4.4 ± 0.6), it was not possible to clarify if only perfect “closed” Ti(OSi)₄ sites are present or if a considerable fraction of defective “open” Ti(OSi)₃ (OH) sites must also be considered (in an ideal EXAFS measurement N is expected to be five if only open sites are present, due to the insertion of a fifth oxygen of the silanol formed in the adjacent Si center when the Ti−O−Si bridge is broken). We report more recent EXAFS data collected at LURE DCI on a high quality TS-1 (Ti = 2.03 wt %) sample: the higher photon flux of that facility and the higher Ti content of the sample has allowed us to remarkably increase the signal/noise ratio with a consequent reduction of the errors associated with the fitted parameters. The quality of the fit, performed on a wider Δk range, has also been strongly improved. Our new measurements indicate that a first shell coordination number higher than four (4.44 ± 0.25). This suggests that, even for samples dehydrated in a carefully controlled atmosphere at 400 K, a considerable fraction of sites exhibits the substitution of a bridged oxygen with two OH groups. Being the value of the error bars associated to the EXAFS analysis of fundamental importance for the solidity of our conclusions, both statistical and systematic errors associated with our measurement and data analysis have been deeply discussed. EXAFS results are also qualitatively confirmed by a parallel photoluminescence study, where the presence, in both emission and excitation spectra, of two distinct bands clearly indicates that we are dealing with two family of Ti sites characterized by slightly different environments. For both techniques, the comparison with measurements performed on the Ti(OSi(CH₃)₃)₄ model compound, where Ti has both the first and the second coordination shells identical to those of Ti (perfect sites) in TS-1, strongly validates our conclusions.