First high-nuclearity thallium–palladium carbonyl phosphine cluster, [Tl2Pd12(CO)9(PEt3)9]2+, and its initial mistaken identity as the unknown Au2Pd12 analogue: structure-to-synthesis approach concerning its formation

Our exploratory research objective to obtain new high-nuclearity Au–Pd carbonyl phosphine clusters from reactions in DMF of preformed Pd10(CO)12(PEt3)6 with Au(PPh3)Cl in the presence of TlPF6 (a frequently utilized chloride-scavenger) has given rise unexpectedly in 40% yield to the first example of a heterometallic Tl–Pd carbonyl phosphine cluster, [Tl2Pd12(CO)9(PEt3)9]2+ (1-Et), as the [PF6]− salt. Its initial incorrect formulation as the unknown Au2Pd12 cluster, obtained from a well-refined low-temperature CCD X-ray diffraction analysis of its crystal structure, was primarily based upon its related molecular geometry to that of the previously reported [Au2Pd14(CO)9(PMe3)11]2+ (as the [PF6]− salt) prepared from an analogous reaction of Pd8(CO)8(PMe3)7 and Au(PCy3)Cl in the presence of TlPF6. (Because X-ray scattering occurs via the electrons of atoms, an assignment in the crystal-structure determination of 1-Et of the two independent “heavy” atoms as either Tl (at. no. 81) or Au (at. no. 79) would result in non-distinguishable refinements). 1-Et was originally characterized by IR and 31P{1H} NMR; attempted MALDI-ToF mass-spectrometric measurements were unsuccessful. The geometrically unprecedented pseudo-C3h core of 1-Et may now be described as edge-fusions of three trigonal bipyramidal Pd5 fragments to a central trigonal bipyramidal Tl2Pd3 kernel. Its formation was originally viewed as the condensation product of three partially ligated butterfly Pd4(CO)3(PEt3)3 fragments that are also linked to and stabilized by two capping naked Au+ cations. This proposed “structure-to-synthesis” approach led to the isolation of 1-Et in ca. 90% yield from the reaction in DMF of the butterfly Pd4(CO)5(PEt3)4 with the phosphine-scavenger Au(SMe2)Cl together with TlPF6. Our later realization and resulting conclusive evidence that its metal-core stoichiometry is Tl2Pd12 instead of Au2Pd12 was a consequence of: (1) our bothersome inability based upon a presumed Au2Pd12 core-geometry to interpret its complex 31P{1H} NMR spectrum despite 31P{1H} COSY experiments clearly showing couplings between the seven major resonances that are consistent with intramolecular processes involving only one species; (2) our subsequent direct preparation of the same Tl2Pd12 cluster (90% yield) from the reaction in THF of Pd4(CO)5(PEt3)4 with TlPF6 (mol. ratio, 3/2), and the ensuing low-temperature CCD X-ray determination revealing a virtually identical solid-state structure (as expected) but with 31P{1H} NMR measurements displaying an analogous complex spectrum that now can be interpreted; and (3) an elemental analysis (Tl, Au, Pd, P), which had been delayed because of the misleading confidence concerning our initially assigned stoichiometry, that ascertained its present formulation; noteworthy is that an elemental analysis of a sample of this compound would not disclose its true identity unless directly tested for Tl (and the absence of Au). Gradient-corrected DFT calculations performed on the PH3-model of the crystallographically known butterfly Pd4(CO)5(PPh3)4 and on its hypothetical Tl+, Au+, and [Au(PH3)]+ adducts (where the optimized geometries consisted of a trigonal bipyramidal MPd4 core with an equatorial M = Tl+, Au+, or [Au(PH3)]+) revealed: (a) that the monocationic Tl+ charge is primarily localized on thallium in contrast to the monocationic Au+ charge being much more delocalized over the entire molecule with charge density having been withdrawn mainly from CO ligands (relative to that of the neutral Pd4(CO)5(PH3)4); (b) that the interactions of Tl+, Au+, or [Au(PH3)]+ adducts with a stable butterfly Pd4(CO)5(PH3)4 model are energetically favorable processes, with Au+ bonding being stronger than Tl+ bonding to Pd4(CO)5(PH3)4; and (c) that the presence of an additional PH3 ligand on the Au+ significantly weakens the Au–Pd bonding interactions such that its bonding energy is comparable with that of the Tl–Pd interactions.