A recently developed compact effective core potential (ECP) scheme for lanthanides [T. R. Cundari and W. J. Stevens, J. Chem. Phys. 98, 5555 (1993)] is applied to molecular species—the lanthanide trihalides (LnX3, Ln=Ce to Lu; X=F, Cl, Br, I). This research is the first ECP study of lanthanide trihalides in which 4f electrons are explicitly included in the lanthanide valence space. The high-spin 4fn configuration of Ln(III) is described using multiconfiguration (MC) and unrestricted Hartree-Fock (UHF) wave functions. Both methods give excellent agreement with experiment for Ln–X bond lengths, although UHF is faster with no evidence of extensive spin contamination. There is no loss of accuracy from early to late members of the lanthanide series and from the lightest to heaviest halogen for the MC calculations. The root-mean-square difference between MC/ECP Ln–X bond lengths and those estimated from gas-phase experiments is only 0.05 Å (≊2%). The UHF calculations on LnF3 yield only three nonplanar minima (CeF3, PrF3, and PmF3). For the three complexes with pyramidal minima, the planar transition states are very close in energy (ΔEUHF<100 cm−1), suggesting these complexes will probably be fluxional at ambient conditions.

A computational study, using relativistic effective core potentials, is presented of transition metalmain group multiply bonded complexes, of interest in the context of catalysis and chemical vapor deposition of TM/MG materials. Model do transition metal complexes chosen are of the general form CI,ME where M = Zr (n = 2). Ta (n = 3), and W (n = 4). Main group elements of interest are the tetrels (E = C, Si, Ge, Sn), pnictogens (E = N, P, As, Sb), and chalcogens (E = 0, S, Se, Te). A comparison between calculated metric data and available experimental data for a wide range Of TM =MG complexes will help in further assessing efficient computational approaches to TM complexes, particularly of the heavier MG elements, as a function of metal, ligand and level of theory. In the present work restricted Hartree Fock (RHF) and Mdler-Plesset second order perturbation theory (MP2) wavefunctions were employed. In most cases there are small differences between RHF and MP2 calculated geometries, with both methods showing good agreement with experimental data, suggesting these approaches will be suitable for the study of larger, more experimentally relevant models. Changes in ZrE bond lengths for E = chalcogen (upon going from RHF to MP2) suggest a fundamentally different description between the Zr-0x0 bond and heavier chalcogens, a result supported by recent experimental data for a series of Zr-chalcogenidos. To date no examples have been reported of arsinidene and stibinidene complexes. Computational results show similar behavior among the heavier pnictogen complexes, i.e., /M=EH (E = P, As, Sb), suggesting that strategies used to synthesize phosphinidenes may be suitable in the search for the first L,M=AsR and L,M =SbR complexes. Additionally, calculations suggest that design of ligand sets which yield linearly coordinated phosphinidenes (and presumably As and Sb analogues) will lead to phosphinidenes with stronger metal-pnictogen bonds and increased thermodynamic stability versus nonlinearly coordinated examples. 0 I994 John Wiley & Sons, Inc.

No abstract