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 ∆k 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(CH 3 ) 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.

No abstract

The room-temperature adsorption of NH3 on Ti-silicalite (TS-1) was studied by means of IR and XANES-EXAFS spectroscopies and by microcalorimetry. The propensity of framework tetrahedral Ti(IV) heteroatoms to expand their coordination shell upon adsorption of NH3 was monitored by the perturbation induced on the Ti-sensitive framework stretching band (960 cm -1 ) and on the preedge peak at 4967 eV in XANES spectra. In both cases a modification of the local geometry of titanium sites upon adsorption of NH3 as an additional ligand was deduced. The amount of NH3 specifically adsorbed on the Ti(IV) sites was estimated by comparing volumetric data obtained for TS-1 with those obtained for a Ti-free silicalite taken as reference material. At pNH 3 ) 50 Torr, the number of NH3 molecules adsorbed per Ti atom was found to be close to 2, indicating that virtually all Ti atoms are involved in the interaction and have completed their 6-fold coordination shell. The molar heats of adsorption on the Ti(IV) sites were estimated following the same procedure. It was found that the molar heat of NH3 adsorption on titanium is higher (qm ) 95-83 kJ/mol) than that measured for the silicalite matrix (qm ) 66-58 kJ/mol). The interaction with ammonia was found to be essentially reversible at room temperature but for a significant amount of adsorbed species irreversibly held on titanium sites. The zero-coverage heat of adsorption was quite high (q ≈ 200 kJ/mol) not only for the first run of adsorption involving both irreversible and reversible interaction but also for the second run involving only the reversible component (q ≈ 130 kJ/mol). The evolution of the heat of adsorption with coverage was typical of heterogeneous surfaces; this is due not only to the presence of sites active toward ammonia on the silica matrix but also to the heterogeneous distribution of Ti(IV) sites, suggesting that a considerable number of framework sites (among the 12 available in the orthorhombic MFI framework) are occupied in a nearly equidistributed way. EXAFS measurements allowed the elongation of the Ti-O bond upon interaction with NH3 (0.05 ( 0.03 Å), the Ti-N distance (1.93 ( 0.03 Å), and the average number of adsorbed NH3 molecules per Ti site (1.9 ( 0.3) to be estimated. The results were in good agreement with the volumetric-calorimetric data. The effect of the pretreatment of TS-1 samples with ammonium acetate aqueous solution was also investigated by comparing the spectroscopic and the quantitative calorimetric results for both untreated and treated TS-1 samples.

No abstract

No abstract