Bonding in Metal Carbonyls

A neutral ligand can bind a low-oxidation-state metal

Carbon monoxide is neutral, yet it forms stable complexes with many transition metals in low formal oxidation states. Ni(CO)₄, Fe(CO)₅ and Cr(CO)₆ are familiar examples. If every CO is assigned charge zero and the entity is neutral, the formal metal oxidation state is zero. That formal assignment does not mean there is no electron-density redistribution. Oxidation states are bookkeeping integers, whereas real bonding involves continuous sharing and polarisation of electron density.

Terminal CO attaches to one metal through carbon. A metal–carbonyl unit is usually approximately linear at M–C–O because the donor and acceptor orbitals overlap effectively along that direction. Drawing the metal attached to oxygen instead would represent a different and unusual binding situation, not the ordinary carbonyl ligand being taught here. A bridging carbonyl attaches to more than one metal and needs a different structural picture.

Two complementary electron-density transfers

In σ donation, an occupied CO orbital with substantial carbon lone-pair character overlaps with an accepting metal orbital. Electron density flows from CO toward the metal along the M–C axis. In π back-donation, an occupied metal d orbital of suitable symmetry overlaps with an empty antibonding π* orbital on CO. Electron density flows from metal toward ligand. The arrows in the diagram label orbital interactions, not a stepwise chemical reaction mechanism.

σ donation supplies electron density to the metal and can facilitate back-donation. Back-donation in turn helps accommodate that density and strengthens the metal–ligand interaction. Populating CO π* weakens the internal C–O bond, often lengthening it and lowering its stretching frequency relative to free CO. Thus metal–carbon bonding can become stronger while carbon–oxygen bonding becomes weaker. Saying that back-donation strengthens both bonds confuses the interfragment bond with the internal ligand bond.

CO is consequently a strong-field ligand in many complexes, a result that a point-charge model alone cannot explain. The metal and CO do not transfer an entire pair irreversibly in each direction as two independent ionic events. The donation/back-donation language decomposes a shared molecular-orbital interaction into useful components.

Structures with one metal

In trigonal-bipyramidal Fe(CO)₅, three equatorial CO ligands lie in a plane around the metal with 120° relationships, and two axial CO ligands lie above and below at 180° to one another and 90° to the equatorial directions. Five-coordinate does not always mean this geometry, but this is the established geometry of Fe(CO)₅. In Cr(CO)₆, six CO donors occupy octahedral positions. In Ni(CO)₄, the four CO donors are tetrahedral and all electrons are paired in the formal d¹⁰ description.

Two metals: separate bridging from terminal groups

Mn₂(CO)₁₀ has a Mn–Mn bond and ten terminal CO ligands, five on each manganese. It can be viewed as two Mn(CO)₅ fragments joined through the metal–metal bond; the five carbonyl positions around each metal form a square-pyramidal fragment. Counting five terminal CO groups at one Mn is not the same as claiming that an isolated Mn(CO)₅ molecule is the complete structure. The metal–metal bond is an additional essential connection.

A commonly depicted form of Co₂(CO)₈ has six terminal CO groups, three on each cobalt, and two bridging CO groups, plus a Co–Co bond. Each bridge is one CO ligand bonded to both metals, so two bridges do not become four separate molecules. Other non-bridged forms are also known, especially depending on conditions. The bridged drawing is a representative structure, not a claim that the formula has only one possible arrangement in every phase.

Formal d count and electron sharing

In neutral homoleptic carbonyls with only neutral CO and homonuclear metal–metal bonds, the metals have formal oxidation state zero. The formal d counts are Ni d¹⁰, Fe d⁸, Cr d⁶, Mn d⁷ and Co d⁹ by group-number bookkeeping. These counts do not directly equal the numbers of unpaired electrons in the complete molecules: bonding, metal–metal interactions and orbital filling must be considered. In particular, do not infer paramagnetism of a dimer by treating its metals as isolated free atoms.

The carbonyl model makes the limitations of purely ionic CFT visible. The ligand is neutral but its π-acceptor character has a major influence on bonding and splitting. Keep the two electron-flow directions, actual carbon connectivity and internal C–O antibonding consequence together; learning only the word synergic leaves the chemical explanation incomplete.

Quick Quiz