Photochemical synthesis and stereophysical characterization of tungsten complex W(CO)4[cyclo-(PW(CO)5)4]: experimental-theoretical bonding analysis of its unprecedented pentametal-coordinated cyclo-P4 ligand

A photochemical reaction of elemental white phosphorus (P4) with W(CO)6 has resulted in the isolation (in 67% yield) of W(CO)4[cyclo-(PW(CO)5)4] (1), which is only the second example of a metal complex containing a cyclo-P4 ligand. 1 was characterized by single-crystal X-ray diffraction, laser-desorption FT mass spectrometric, solution P-31, C-13, and H-1 NMR, solid-state P-31 NMR, infrared, and electrochemical measurements. This complex possesses a novel pentametal-coordinated cyclo-P4 ligand, which is linked not only to an apical W(CO)4 fragment by its four pi-electrons but also to four Lewis-acid W(CO)5 fragments by its four normally unshared electron pairs. Evidence that the cyclo-P4 ligand possesses an instantaneous C(s)-m kite-shaped geometry (which was not detected in the X-ray diffraction study due to the molecule possessing crystallographic C4-4 site symmetry) is given by solution P-31 NMR spectra, which exhibit three resonances with an AM2X pattern at room temperature in high-polarity solvents and at lower temperatures in low-polarity solvents. Of particular interest is that a recent X-ray crystallographic determination of Nb(eta-5-C5Me5)(CO)2P4 (2) revealed that its cyclo-P4 ring is slightly distorted from a square configuration toward a planar C(s)-m kite-shaped configuration in accordance with molecular pseudo-C(s) symmetry. A low-temperature solution P-31 NMR spectrum of 2 also exhibited three well-resolved signals, which were attributed to hindered rotation of the cyclo-P4 ligand. That a presumably analogous C(s)-m kite-shaped configuration is exhibited by the cyclo-P4 ligand in both 1 and 2, which have cylindrical C4 and noncylindrical C(s) symmetry, respectively, provides convincing evidence that this distortion of the cyclo-P4 ligand in solution is due solely to electronic interactions involving the pi-coordinated metal-ligand fragment and appears to be independent of the overall molecular symmetry. Unfortunately, room- and low-temperature solid-state P-31 NMR spectra of 1 did not yield additional information about the cyclo-P4 ring. Molecular orbital calculations were carried out with the nonparametrized Fenske-Hall model on the hypothetical W(CO)4P4 molecule and on 1 in order to describe the bonding of a cyclo-P4 unit pi-coordinated to a metal tetracarbonyl fragment of overall C4 symmetry. The MO results revealed that the doubly degenerate HOMOs in W(CO)4(eta-4-P4) are primarily composed of slightly antibonding in-plane cyclo-P4 orbitals, while the doubly degenerate LUMOs mainly possess bonding W(5d-pi)-CO(pi*) orbital character with smaller contributions of antibonding W(5d-pi)-P(3p-pi) orbital character. The main contributors to the doubly degenerate HOMOs of 1 are nonbonding W(CO)5 orbitals; the much higher energy LUMOs remain essentially unchanged in orbital character. Hence, the MO calculations are consistent with the observed irreversible electrochemical behavior of 1 but provide no clear-cut explanation for a distortion of the cyclo-P4 ligand. 1 crystallizes as solvated 1.CH2Cl2 in two nonseparated crystalline forms. X-ray crystallographic studies showed that both forms have tetragonal unit cells of virtually identical volumes (with Z = 2) but of different symmetries (P4nc versus I4); structural analyses showed that 1 has an analogous molecular configuration in both crystal forms. The formulation of 1 as W5(CO)24P4 was confirmed from a positive-ion LD/FT mass spectrum, which showed the existence of the parent-ion peak envelope (m/z approximately 1714).