Why does phosphine not backbond as well as CO in metal complexes?

From Perturbation MO theory (just accept it!) the strength of an A-B bond (σ or π) is inversely related to the energy difference between the interacting MOs (the H*OMO(A) -LUMO(B) gap or vice versa). CO has low lying π* MOs that have the right symmetry and energy to overlap with the filled M d AOs (high lying since M(0) not M(II)) hence the bond between M(0)-C≡O has double bond character and the synergic nature of the bond: O≡C:→M(0) M(d)→π*C≡O makes the bond strong.

The M(0)-PH3 bond is much stronger than the corresponding M(0)-NH3 bond despite ammonia being a stronger Lewis acid. This is attributed to the PH3 ligand having π acceptor properties. The discussion has been the nature of the π acceptor MOs. In PF3 the the P carries a partial +ve charge (electronegative F) and it is thought the 3d AOs on P are lowered sufficiently in energy to take part in significant M(d)→P(3d) bonding. Certainly the bonding properties of PF3 are almost identical to those of CO.

However in PMe3 the 3d AOs are believed too high in energy for there to be any significant π bonding via the 3d AOs on PMe3. It is believed that the π acceptor properties occurs via the P-C σ* MOs. And PH3 is probably much the same.

The bottom line is that the π acceptor MOs are much higher in energy that those in CO.