Colour in coordination compounds

A coloured complex absorbs part of the visible light passing through it; its colour is the complementary colour of what is left. If green is absorbed, the complex looks red.

Absorbed wavelength and colour seen: [CoCl(NH₃)₅]²⁺ 535 nm (yellow) looks violet; [Co(NH₃)₅(H₂O)]³⁺ 500 nm (blue-green) looks red; [Co(NH₃)₆]³⁺ 475 nm (blue) looks yellow-orange.

More pairs: [Co(CN)₆]³⁻ absorbs at 310 nm in the ultraviolet and is pale yellow; [Cu(H₂O)₄]²⁺ absorbs 600 nm (red) and looks blue; [Ti(H₂O)₆]³⁺ absorbs 498 nm (blue-green) and looks violet.

[Ti(H₂O)₆]³⁺ is d¹: the lone electron sits in t₂g. Blue-green light promotes it to eg (t₂g¹eg⁰ → t₂g⁰eg¹), so the complex looks violet. CFT puts colour down to such d-d transitions.

No ligand, no splitting, no colour: heating [Ti(H₂O)₆]Cl₃ to drive off water leaves a colourless solid, and anhydrous CuSO₄ is white while CuSO₄·5H₂O is blue.

Changing the ligand changes the colour. Adding en to green [Ni(H₂O)₆]²⁺ in en:Ni ratios 1:1, 2:1 and 3:1 gives pale blue [Ni(H₂O)₄(en)]²⁺, blue/purple [Ni(H₂O)₂(en)₂]²⁺ and violet [Ni(en)₃]²⁺.

Ruby is Al₂O₃ with about 0.5-1% Cr³⁺ (d³) in some Al³⁺ sites; these act like octahedral Cr(III) complexes whose d-d transitions give the colour. In emerald, Cr³⁺ sits in octahedral sites of beryl (Be₃Al₂Si₆O₁₈); its bands move to longer wavelengths (yellow-red and blue), so green is transmitted.

Limits of CFT: treating ligands as point charges predicts that anions should split most, yet anionic ligands sit at the weak end of the spectrochemical series; and CFT ignores the covalent side of metal-ligand bonding. Ligand field and molecular orbital theories handle these.