Density functional theory (DFT, B3PW91) calculations have been carried out on the reactivity of ethene with model systems M(NR)(=CHCH3)(X)(Y) for M = Mo or W, R = methyl or phenyl, X = CH2CH3, OCH3, or OSiH3, and Y = CH2CH3, OCH3, or OSiH3, which are representative of experimental olefin metathesis catalysts, and the results are compared to those previously obtained for Re(CCH3)(=CHCH3)(X)(Y). The general pathway comprises four steps: olefin coordination, [2+2] cycloaddition, cycloreversion, and olefin de-coordination. Two key factors have been found to control the detailed shape of the energy profiles: the energy of distortion of the tetrahedral catalyst and the stability of the metallacycle intermediate, which is controlled by the M-C bond strength. The efficiency has been evaluated by calculating the turnover frequency (TOF) based on the steady-state approximation, and the most striking feature is that the unsymmetrical catalysts (X not equal to Y) are systematically more efficient for all systems (Mo, W, and Re). Overall, the Re complexes have been found to be less efficient than the Mo and W catalysts, except when Re is unsymmetrically substituted: it is then calculated to be as efficient as the best Mo and W catalysts.

Both industrial and biochemical ammonia syntheses are thought to rely on the cooperation of multiple metals in breaking the strong triple bond of dinitrogen. Such multimetallic cooperation for dinitrogen cleavage is also the general rule for dinitrogen reductive cleavage with molecular systems and surfaces. We have observed cleavage of dinitrogen at 250 degrees C and atmospheric pressure by dihydrogen on isolated silica surface-supported tantalum(III) and tantalum(V) hydride centers [(identical with Si-O)2Ta(III)-H] and [(identical with Si-O)2Ta(V)H3], leading to the Ta(V) amido imido product [(identical with SiO)2Ta(=NH)(NH2)]: We assigned the product structure based on extensive characterization by infrared and solid-state nuclear magnetic resonance spectroscopy, isotopic labeling studies, and supporting data from x-ray absorption and theoretical simulations. Reaction intermediates revealed by in situ monitoring of the reaction with infrared spectroscopy support a mechanism highly distinct from those previously observed in enzymatic, organometallic, and heterogeneous N2 activating systems.

The efficiency of silica supported d(0) ML(4) alkene metathesis catalysts [([triple bond]SiO)M(NR(1))(=CHR(2))(X)] (M = Mo, W; R(1) = aryl and alkyl) is influenced by the nature of the X ancillary ligand. Replacing the alkyl ligand by a pyrrolyl ligand dramatically increases the performance of the catalyst. DFT calculations on the metathesis, the deactivation, and the byproduct formation pathways for the imido Mo and W and the alkylidyne Re complexes give a rational for the role of pyrrolyl ligand. Dissymmetry at the metal center leads to more efficient catalyst even when the difference in sigma-donating ability between X and OSi is not large. beta-H transfer at the square based pyramid metallacyclobutane is the key step for catalyst deactivation and byproduct formation. Overall, the greatest benefit of substituting the ancillary alkyl by a pyrrolyl ligand, [([triple bond]SiO)M(ER(1))(=CHR(2))(pyrrolyl)], is in fact not to improve the efficiency of the catalytic cycle of alkene metathesis, but to shut down deactivation and byproduct formation pathways. Pyrrolyl ligand, and more generally ligands having metal-bound-atoms more electronegative than carbon, disfavor mostly the two first steps (beta-H transfer at the metallacyclobutane and subsequent insertion of an ethene in the M-H bond) of the deactivation channel. The [([triple bond]SiO)M(ER(1))(=CHR(2))(pyrrolyl)] catalyst is thus highly efficient because pyrrolyl ligand is optimal: (i) it is still a better electron donor than the siloxy group, thus, favoring the metathesis pathway (dissymmetry at the metal center); and (ii) the nitrogen of the pyrrolyl ligand is more electronegative than the carbon of the alkyl group, thus, specifically disfavoring the decomposition of the metallacyclobutane intermediate via beta-H transfer.

DFT(B3PW91) calculations show that the reaction pathways for ethylene metathesis with Re([triple bond]CMe)(=CHMe)(X)(Y) (X/Y = CH2CH3/CH2CH3; CH2CH3/OSiH3; OSiH3/CH2CH3; OCH3/OCH3, CH2CH3/OCH3, and OCF3/OCF3) occur in two steps: first, the pseudo-tetrahedral d0 Re complexes distort to a trigonal pyramid to open a coordination site for ethylene, which remains far from Re (early transition state for C-C bond formation). The energy barrier, determined by the energy required to distort the catalyst, is the lowest for unsymmetrical ligands (X not equal Y) when the apical site of the TBP is occupied by a good sigma-donor ligand (X) and the basal site by a poor sigma-donor (Y). Second, the formation of metallacyclobutanes (late transition state for C-C bond formation) has a low energy barrier for any type of ligands, decreasing for poor sigma-donor X and Y ligands, because they polarize the Re-C alkylidene bond as Re(+delta)=C(-delta), which favors the reaction with ethylene, itself polarized by the metal center in the reverse way. The metallacyclobutane is also a TBP, with apical alkylidyne and Y ligands, and it is stabilized by poor sigma-donor X and Y. The best catalyst will have the more shallow potential energy surface, and will thus be obtained for the unsymmetrical set of ligands with X = a good sigma-donor (alkyl) and Y = a poor sigma-donor (O-based ligand). This rationalizes the high efficiency of well-defined Re alkylidene supported on silica, compared to its homogeneous equivalent, Re([triple bond]CMe)(=CHMe)(OR)2.

The reaction of [W(dNAr)(dCHtBu)(CH 2 tBu) 2 ] (1; Ar ) 2,6-iPrC 6 H 3 ) with a silica partially dehydroxylated at 700 °C, SiO 2-(700) , giVes syn-[(tSiO)W(dNAr)(dCHtBu)(CH 2 tBu)] (2) as a major surface species, which was fully characterized by mass balance analysis, IR, NMR, EXAFS, and DFT periodic calculations. Similarly, complex 1 reacts with [(c-C 5 H 9 ) 7 Si 7 O 12 -SiOH] to giVe [(SiO)W(dNAr)(dCHtBu)(CH 2 tBu)] (2m), which shows similar spectroscopic properties. Surface complex 2 is a highly actiVe propene metathesis catalyst, which can achieVe a TON of 16 000 within 100 h, with only a slow deactiVation.

The surface complex [([triple bond]SiO)Re([triple bond]CtBu)(=CHtBu)(CH2tBu)] (1) is a highly efficient propene metathesis catalyst with high initial activities and a good productivity. However, it undergoes a fast deactivation process with time on stream, which is first order in active sites and ethene. Noteworthy, 1-butene and pentenes, unexpected products in the metathesis of propene, are formed as primary products, in large amount relative to Re (>>1 equiv/Re), showing that their formation is not associated with the formation of inactive species. DFT calculations on molecular model systems show that byproduct formation and deactivation start by a beta-H transfer trans to the weak sigma-donor ligand (siloxy) at the metallacyclobutane intermediate having a square-based pyramid geometry. This key step has an energy barrier slightly higher than that calculated for olefin metathesis. After beta-H transfer, the most accessible pathway is the insertion of ethene in the Re-H bond. The resulting pentacoordinated trisperhydrocarbyl complex rearranges via either (1) alpha-H abstraction yielding the unexpected 1-butene byproduct and the regeneration of the catalyst or (2) beta-H abstraction leading to degrafting. These deactivation and byproduct formation pathways are in full agreement with the experimental data.

Density functional calculations have been carried out to analyze the origin of the differences in reactivity, selectivity, and stability toward deactivation in metathesis of d0 oxo alkylidene complexes vs their isoelectronic imido counterparts. DFT calculations show that the elementary steps and geometries of the extrema are similar for the oxo and imido complexes, but that the energy profiles are different, the greatest difference occurring for the deactivation pathway. For the alkene metathesis pathway, replacing the imido by an oxo ligand slightly lowers the energy barrier for alkene coordination but raises that for the [2+2]-cycloaddition and cycloreversion; it also destabilizes the trigonal bipyramidal (TBP) metallacyclobutane intermediate with respect to the separated reactants. The isomeric square-based pyramid (SP) metallacyclobutane is in general more stable, and its stability relative to the separated reactants is similar for oxo and imido systems. Consequently, the oxo complex is associated with a slightly larger energy difference between the lowest energy intermediate (SP or separated reactants) and the highest energy transition state (cycloreversion) than the imido complex, which accounts for a slightly lower activity. Changing the imido into an oxo ligand disfavors strongly the deactivation pathway by raising considerably the energy barrier of the β-H transfer at the SP metallacycle that begins the entry into the channel for deactivation and byproduct formation as well as that of the subsequent ethene insertion. This makes the oxo catalysts more selective and stable toward deactivation than the corresponding imido catalysts, when dimerization can be avoided.

The molecular dynamics of a series of organometallic complexes covalently bound to amorphous silica surfaces is determined experimentally using solid-state nuclear magnetic resonance (NMR) spectroscopy and density functional theory calculations (DFT). The determination is carried out for a series of alkylidene-based catalysts having the general formula [([triple bond]SiO)M(ER)(=CH(t)Bu)(R')] (M = Re, Ta, Mo or W; ER = C(t)Bu, NAr or CH2(t)Bu; R' = CH2(t)Bu, NPh2, NC4H4). Proton-carbon dipolar coupling constants and carbon chemical shift anisotropies (CSA) are determined experimentally by solid-state NMR. Room-temperature molecular dynamics is quantified through order parameters determined from the experimental data. For the chemical shift anisotropy data, we validate and use a method that integrates static values for the CSA obtained computationally by DFT, obviating the need for low-temperature measurements. Comparison of the room-temperature data with the calculations shows that the widths of the calculated static limit dipolar couplings and CSAs are always greater than the experimentally determined values, providing a clear indication of motional averaging on the NMR time scale. Moreover, the dynamics are found to be significantly different within the series of molecular complexes, with order parameters ranging from = 0.5 for [([triple bond]SiO)Ta(=CH(t)Bu)(CH2(t)Bu)2] and [([triple bond]SiO)Re([triple bond]C(t)Bu)(=CH(t)Bu)(CH2(t)Bu)] to = 0.9 for [([triple bond]SiO)Mo([triple bond]NAr)(=CH(t)Bu)(R') with R' = CH2(t)Bu, NPh2, NC4H4. The data also show that the motion is not isotropic and could be either a jump between two sites or more likely restricted librational motion. The dynamics are discussed in terms of the molecular structure of the surface organometallic complexes, and the orientation of the CSAs tensor at the alkylidene carbon is shown to be directly related to the magnitude of the alpha-alkylidene CH agostic interation.