Proton transfer from a zeolitic cluster to NH3 and subsequent coordination of the ammonium cation onto the zeolitic cluster are studied by using a b initio quantum chemical cluster calculations. Proton transfer from the zeolite cluster to NH3 is favorable if, after proton transfer, the resulting NH4+ cation is coordinated to the zeolitic cluster with two or three hydrogen bonds. These structures are referred to as 2H and 3H, respectively. Their adsorption energies the energy needed for the process of proton transfer followed by the binding of the NH4+ cation, are calculated to be -1 14 and -1 13 kJ/mol, respectively. The geometries were optimized at the S C F level and the adsorption energies were calculated at the second-order Mprller-Plesset perturbation theory level (MP2), using the counterpoise correction (CPC) to avoid the basis set superposition error (BSSE). The basis set is the 6-3 1 l+G(d,p)/STO-3G one, which has previously been shown to give proper binding and proton transfer energies. The calculated heats of adsorption compare well with experimental heats of desorption. Proton transfer also occurs when another NH3 molecule is coadsorbed. However, the process of coadsorption is energetically less favorable than the 2H and 3 H structures: the adsorption energy per NHs molecule is only -30 kJ/mol. For the clusters the N-H stretching frequencies have been calculated at the SCF level in the harmonic approach. They have been compared with experimental spectra of the NH4+ forms of some zeolites.The N-H stretching region of these spectra can be explained as a superposition of the spectra of the 2H and 3H structures. By comparison of the adsorption energy on a geometry optimized cluster and on a fixed geometry cluster, it was found that the choice of the geometry is important. On enlarging the fixed geometry cluster the adsorption energy remained constant.

The calculation of adsorption energies in zeolites in the cluster approximation has the advantage of a reduced computational effort compared to that of the periodic approach. However, in the cluster approximation, the long-range electrostatic effects of the infinite crystal are ignored and there are boundary effects. In order to remove a part of the disadvantages of the cluster approach, we developed a method to embed a cluster in a zeolite crystal by imposing an electrostatic potential on it. This potential adds the long-range electrostatic effects of the crystal and subtracts the electrostatic potential of the boundary of the cluster. It is calculated from the charge distribution of the crystal obtained with an ab initio calculation using the CRYSTAL program. We calculated the adsorption energies of NH, and NH: on three different clusters embedded in the potential of a chabazite crystal, and we have compared them with the adsorption energies of NH3 and NH: in the crystal. If a cluster is used that has dangling bonds four, or more, bonds away from the adsorbate, i.e., the adsorption site is described well covalently, embedding the cluster reduces the error in adsorption energies (relative to the crystal calculations) from -20 to -2-3 kJ/mol.

chemisorption of CO or HZ. As is usually the case with other systems, there is some uncertainty in the stoichiometry of titration. However, the merit of this type of titration procedure is to provide a standard that permits an easy comparison between the work of different investigators. Hence, consistency, reproducibility, and convenience should be attributes of the procedure. For molybdenum and molybdenum carbides, both CO and H2 chemisorption methods appear to satisfy the requirements. The recommended procedures are summarized as follows: (i) For molybdenum carbides, CO chemisorption at 196 K with the stoichiometry of one CO per surface Mo atom is recommended. For molybdenum, CO chemisorption at RT with the stoichiometry of two CO per surface Mo atom is better. (ii) For both Mo and carbides, H2 chemisorption while cooling the sample from 620 K to RT and the stoichiometry of one H per surface Mo atom are r m m e n d e d .The estimation of particle size by chemisorption is difficult due to the presence of polymeric carbon. The success by CO chemisorption in the present study should be considered rather extraordinary. The XRD could be used for unsupported molybdenum carbides. Yet it is difficult to apply to small particles on a support, alumina in particular. The TEM is the best method for the purpose. The interaction of NH3 and a zeolitic cluster as well as the protonation of NH3 by zeolitic protons are studied by quantum-chemical calculations on small clusters at different levels of approximation. The focus of the paper is on a comparison of results obtained by the different methods. The clusters are studied at the SCF level as well as at the correlated level. Electron correlation is included through second-order Maller-Plesset perturbation theory. The basis-set superposition error (BSSE) was avoided by using the counterpoise scheme. Monodentate singly bonded NH3, that is NH, being attached to one oxygen atom, forms a strong hydrogen bond with the zeolitic OH group. This bond has a strength of 60 or 67 kJ/mol, depending on the geometry of the zeolitic cluster. This value is approximately half the experimentally found heat of desorption. Acknowledgment.For this case, the 0-N distance is found to be very short (2.74 or 2.73 A) and the intermolecular 0-H-N stretching frequency is calculated to be 185 or 193 cm-l. The latter values agree reasonably with experimental data. Upon complexation with NH,, the OH stretching frequency shows a red shift of 551 cm-I. Proton transfer from the zeolitic cluster to NH3 is calculated to be unfavorable by 52 kJ/mol, as long as NH4+ is considered to be monodentate coordinated. The description of the hydrogen-bonded form is only slightly dependent on the basis set used. However, the proton-transfer energy does strongly depend on the basis set used. Electron correlation makes the proton transfer more favorable. The BSSE has a large influence on the description of the structures, especially if electron correlation is included. Although electron correlation has a nonncgbgible effect on...

We have calculated the adsorption energy of NH3 and NHZ in a zeolite crystal and on four different dusters cut out from the crystal. The adsorption energies were calculated for two different geometries of the chabazite: a geometry determined with x-ray diffraction and a shell model optimized geometry. The effect of the geometry optimization of the crystal is large. In the crystal for which the geometry has been optimized the adsorbates are less stable.For NH, the difference between the two geometries is 29 kJ mol-' (adsorption energies -101 and -72 W mol-'), and for NHZ the difference between the two geomevies is 137 ! d mol-' (-15 and 122 kJ mol-'). The adsorption energy of NHI is a local process. The effect of the Madelung potential is relatively small: for the two geometries we found -6 and +5 kJ mol-'. The adsorption of NHf is more affected by the Madelung potential -53 and t 2 3 kJ mol-' for the two geometries. Clusters thal have bonds saturated with hydrogen atoms close to the adsorbate do not give a proper descnption of the adsorption process: this type of boundary error can be as large as 60 W mol-',

The adsorption of N H 3 in acidic zeolites has been studied extensively experimentally. Therefore, it can be used very well to verify a model used in a quantum chemical calculation. Here, we present a calculation that, from a quantum chemical point of view, should give a reliable description of the adsorption process. We studied the adsorption of N H 3 and NJ%+ in chabazite with the embedded cluster method using a reasonable basis set, applying the counterpoise correction and including electron correlation. The geometry was partially optimized. With this calculation we verified the reliability of our model and obtained information that cannot be obtained experimentally. The adsorption energies of hydrogen-bonding NH3 and of NH4+ were -70 f 10 kJ/mol and -120 f 15 kJ/mol, respectively. The latter value compares very well with the experimental heat of adsorption. N&+ has a high coordination with the zeolite wall; this is confirmed experimentally. A good geometry is obtained if the boundary of the embedded cluster is kept fixed to that of the zeolite crystal.