previously he was at Sandia National Laboratories. His research concerns the energetics and kinetics of the reactions between minerals and aqueous solutions, and in recent work, he has determined rate coefficients for dissociation of Al−O bonds in various dissolved aluminum complexes.

No abstract

High-resolution solid-state NMR techniques were used to investigate the surface structure of Cab-O-Sil fumed silica. 1H NMR results obtained from CRAMPS, MAS-only and relaxation studies reveal the existence of both hydrogen-bonded silanols and isolated silanols on the Cab-O-Sil surface. A systematic dehydration study of fumed silica was carried out, with results on the quantity of each type of silanol on the surface at various dehydration stages. 29Si CP-MAS experiments, including CP spin dynamics studies and various other relaxation studies, were employed to probe hydrogen bonding and the local structural environments of various hydroxyl groups of silica surfaces. 29Si CP-MAS experiments on water-treated and deuterium-exchanged Cab-O-Sil indicate the existence of interparticle silanols and internal silanols in fumed silica. 1H and 29Si NMR show that for fumed silica both isolated and hydrogen-bonded silanols are present on the surface of an untreated sample, in contrast to the case of silica gel, where all silanols of an untreated sample are hydrogen bonded.

No abstract

No abstract

A refined and generalized version of a previously suggested model of the silica surface, in which geminal silanols are situated on surface segments similar to (100)-type faces of the β-cristobalite structure and single silanols are situated on surface segments similar to corresponding (111)-type faces, is supported by extensive spectroscopic data. In this model single silanols on the same (111)-type surface segment cannot form hydrogen bonds with each other. Whether or not adjacent geminal silanols on the same (100)-type surface segment can form hydrogen bonds with each other depends on the relative orientation of their hydroxyl groups. When two surface segments of either (100)- or (111)-type intersect convexly, hydroxyl groups cannot participate in hydrogen bonding across the intersection; but when two surface segments intersect concavely, those silanols situated at the intersection can form hydrogen bonds with their counterparts across the intersection. All the hydrogen-bonding silanols in this generalized β-cristobalite model have a common feature:  when any two silanols are hydrogen bonded to each other, the two silicon atoms containing them are also situated on the same (100)-type surface segment. This idealized structure of the surface of silica gel, which is clearly known from X-ray diffraction to be an amorphous material, may be distorted for various thermodynamic or kinetic reasons during its formation; therefore, a wide range of hydrogen-bonding strengths between two hydroxyls is likely on a real silica gel surface. The generalized β-cristobalite surface model can also explain the reversible dehydroxylation and rehydroxylation processes on silica surfaces. Both single and geminal silanols participating in hydrogen bonding are most easily dehydroxylated under evacuation at temperatures between 170 and 450 °C and form low-strain bicyclo[3.3.0]octasiloxane rings. The mode of dehydroxylation on a silica surface undergoes a transformation between 450 and 650 °C, yielding highly strained trisiloxane rings for dehydroxylation at T ≥ 650 °C.

A 29Si NMR study was carried out on silica gel, using cross polarization and magic-angle spinning (CP/MAS). Spectra were examined on silica samples prepared at various stages of dehydration. It was found that the spectral changes observed in these experiments could not be accounted for by a single structural model of the types that have been advanced previously for silica surfaces. However, these 29Si spectral features are consistent with a heterogeneous silica surface consisting of separate regions resembling the 100 face and 111 face of 0-cristobalite.

5208electrons is not taken into consideration, and individual species such as S, C, and N are assumed to be rigid. Accordingly, the real effective ionic radius is expected to be larger than 1.93 A. Jindal and HarringtonZ7 have already pointed out that the effective ionic radii of NO< and SCNare calculated to be 1.60 and 1.86 A, respectively, and J. Phys. Chem. 1982, 86, 5208-5219 concluded that these values are small compared with the published data (see the literature4 where the ionic radii so far reported are summarized). Therefore, our value 2.15-2.20 A evaluated as the effective ionic radius of SCNis thought to be reasonable. The above discussion implies that the electronic polarizability of an ion can be used to estimate the effective ionic radius. Further investigation, the true nature of electronic polarization.29Si NMR experiments, using cross polarization (CP) and magic-angle spinning (MAS), were carried out on a variety of silica gels and the products of their trimethylsilylation reactions with the silylating agent, hexamethyldisilazane (HMDS). A methodology has been developed to provide quantitative relationships on structure and reactivity from the 29Si CP/MAS spectra. Geminal-hydroxyl silanol sites were found to be more reactive to HMDS than lone-hydroxyl silanol sites. Measured surface hydroxyl densities and trimethylsilane coverages are discussed in terms of structural models. Experimental SectionNMR Measurements. Solid-state 29Si NMR spectra were obtained at 11.88 MHz on a prototype JEOL FX-6OQS NMR spectrometer and at 39.75 MHz on a modified Nicolet NT-200 spectrometer. Details of these spectrometer systems are described elsewhere.A piston-cylinder apparatus was used to measure the transition curves and corresponding volume changes in condensed NH3 from 200 to 305 K (0.9-10.5 kbar). The melting curve has two distinct branches: a lower one bounding solid I and an upper one bounding solid 11. The two melting curves intersect at 217.34 K (3.070 kbar), forming a solid I-solid 11-liquid triple point. The solid 1-olid I1 transition line was determined from the triple point up to 224 K (8.0 kbar). From sound-velocity measurements, it appears that ultrasonic signals are completely attenuated in both solids over wide P-T regions adjoining the melting curves.