covalency in metal-ligand bonding
/ ko-VAY-len-see /
Crystal field theory tells a tidy story: ligands are point charges, the bond is pure electrostatic attraction and repulsion, and the metal d electrons stay entirely on the metal. It is a useful story, but it is not true. Real metal-ligand bonds share electrons — the d electrons spill out onto the ligands and the ligand electrons spill onto the metal. That sharing is covalency, and recognizing it is the whole reason ligand field theory had to be invented.
How do we know the bond is partly covalent rather than purely ionic? Several experiments catch the d electrons living partly on the ligands. Electron paramagnetic resonance shows the metal's unpaired electron interacting with ligand nuclei (for example, hyperfine coupling to fluorine in metal fluorides) — impossible if the electron were locked on the metal. The nephelauxetic effect shows the d-d repulsion shrinking compared with the free ion, as if the d cloud has expanded by spreading onto the ligands. And the spectrochemical series itself, with neutral CO outsplitting charged fluoride, only makes sense once you allow orbital overlap and electron sharing. In molecular-orbital language, covalency simply means the metal and ligand orbitals genuinely mix, so the bonding orbitals are part-metal, part-ligand.
Why does this matter beyond bookkeeping? Covalency controls how strong delta really is, sets the true magnetic and spectroscopic properties, and underlies whole fields: organometallic chemistry, where metal-carbon bonds are deliberately covalent, and bioinorganic chemistry, where covalent metal-sulfur bonds in iron-sulfur proteins tune their reactivity. The honest summary is that the bonding in a complex sits on a sliding scale from mostly ionic (a hard, electropositive metal with hard ligands) to substantially covalent (a soft, low-valent metal with soft, pi-accepting ligands), and the crystal field model is only the ionic extreme of that scale.
In permanganate, MnO4-, a formally Mn(VII) d0 ion would have no d electrons to make a d-d color, yet the ion is intensely purple. The deep color comes from a charge-transfer band — electrons jumping from oxygen to manganese — which is only possible because the Mn-O bonds are heavily covalent, with metal and ligand orbitals strongly mixed.
A d0 ion with no d-d transition is still vividly colored — proof that covalency, not just crystal field splitting, is at work.
Covalency does not abolish the crystal field picture — it refines it; for hard, ionic-leaning complexes the electrostatic model is already a good approximation, and covalency matters most for soft metals, low oxidation states, and pi-bonding ligands.