A color too loud for a d-d band
In the previous guide you learned that the soft, washed-out colors of most transition-metal complexes come from d-d transitions — an electron hopping from the lower t2g set up into the eg set across the gap delta-o. Those bands are gentle for a deep reason: they are forbidden. The Laporte rule says a d-to-d jump within the same shell shouldn't happen at all, and only a little wobbling of the molecule (vibronic coupling) sneaks it through. The result is a faint absorption and a pale tint — the pale pink of [Co(H2O)6]2+, the soft blue-green of [Ni(H2O)6]2+.
Now hold a crystal of potassium permanganate, KMnO4, up to the light. The purple is so saturated it stains your skin and a single grain tints a whole beaker. That color cannot be a d-d band — and the giveaway is that the manganese in permanganate is Mn(VII), a d0 ion with no d electrons at all to jump between d orbitals. Something else entirely is absorbing the light. That something is a charge-transfer transition, and instead of shuffling an electron between two d orbitals on the same atom, it hurls an electron all the way from the ligand to the metal, or from the metal to the ligand.
Why charge-transfer bands are so loud
The reason a charge-transfer band dwarfs a d-d band is the selection rules. A d-d transition is Laporte-forbidden because it starts and ends in orbitals of the same kind (both d, both gerade — symmetric about the metal's center). A charge-transfer transition does not have that problem: the electron starts on an orbital that is mostly on the ligand and ends on one that is mostly on the metal, so it changes its character and its place in space dramatically. That motion is fully Laporte-allowed, so the molecule absorbs light eagerly rather than reluctantly.
Translate "allowed versus forbidden" into numbers and the contrast is stark. Chemists measure absorption strength with the molar extinction coefficient, epsilon. A Laporte-forbidden d-d band typically has an epsilon of roughly 1 to 100, a faint smudge. A charge-transfer band routinely runs from a few thousand up to tens of thousands — hundreds to thousands of times stronger. That is the whole story of permanganate's outrageous color: each MnO4- ion is an extravagantly efficient little antenna for green-yellow light, so even a trace solution looks intensely purple.
Two directions: LMCT and MLCT
An electron can be flung in either direction, and the two cases have names. In ligand-to-metal charge transfer (LMCT) the electron jumps from a filled ligand orbital onto an empty or half-empty metal d orbital — the ligand effectively reduces the metal for an instant. This is favored when the metal is in a high oxidation state, hungry for electrons, sitting next to a ligand that gives electrons up easily (a good reducing ligand like oxide, O^2-, or sulfide, or bromide). Permanganate and chromate are textbook LMCT: an electron leaps from an oxide-based orbital onto the electron-starved d0 metal center.
The opposite case is metal-to-ligand charge transfer (MLCT): the electron jumps the other way, from a filled metal d orbital onto an empty orbital on the ligand. This needs a metal in a low oxidation state, electron-rich and willing to give, paired with a ligand that has low-lying empty orbitals ready to catch — exactly the pi-acceptor ligands you met in the spectrochemical series, like CO, CN-, or the aromatic chelates bipyridine and phenanthroline. The famous orange of [Fe(bipy)3]2+ and the deep red of the ruthenium complex [Ru(bipy)3]2+ are MLCT bands; that ruthenium dye is the workhorse of dye-sensitized solar cells precisely because its MLCT absorption is so strong and well-placed in the visible.
LMCT (ligand -> metal): ligand orbital --e-> metal d orbital needs: high oxidation-state, oxidizing metal + easily-oxidized ligand e.g. MnO4^- (Mn d0), CrO4^2-, [FeBr4]^- MLCT (metal -> ligand): metal d orbital --e-> ligand pi* orbital needs: low oxidation-state, reducing metal + pi-acceptor ligand e.g. [Fe(bipy)3]^2+, [Ru(bipy)3]^2+, many M-CO complexes rule of thumb: electron flows from the easy giver to the eager taker
Telling a CT band from a ligand-field band
When you look at a real spectrum you often see both kinds of band at once, and you need to tell which is which. The fastest tell is sheer intensity: read the epsilon. A weak band (epsilon up to ~100) is almost certainly a ligand-field d-d transition; a band a thousand times stronger is charge transfer. A second clue is position — charge-transfer bands usually sit at higher energy, often in the ultraviolet, with only their low-energy tail spilling into the visible to give the color, whereas d-d bands sit squarely in the visible. A third clue is d-electron count: if the metal is d0 (like Mn(VII)) or d10 (like Zn(II)), there are no d-d transitions possible at all, so any color you see must be charge transfer.
- Read the intensity: if epsilon is tens of thousands it is charge transfer; if it is only single or double digits it is a Laporte-forbidden d-d band.
- Check the d-electron count: a d0 or d10 center has no d-d transitions, so its color can only be charge transfer.
- Note the energy: a band lurking in the UV with a tail reaching into the visible is the classic charge-transfer signature; a band centered in the visible is more likely d-d.
- Decide the direction: if the band shifts toward lower energy as the ligand gets easier to oxidize it is LMCT; if it shifts as the metal gets easier to oxidize and the ligand is a pi-acceptor it is MLCT.
There is an elegant confirmation of the LMCT direction. Across a series of halide complexes [MX4]^n-, swapping the halide from fluoride to chloride to bromide to iodide shifts the charge-transfer band steadily to lower energy (redder). That is exactly what LMCT predicts: iodide gives up its electrons most easily, so it takes less energy to pull one onto the metal, and the band moves to lower energy. The trend tracks how easily each ligand is oxidized — a beautiful, testable fingerprint that the electron really is moving from ligand to metal.
Honest limits and where the colors show up
A few honest cautions. First, calling permanganate's transition "ligand to metal" does not mean the manganese is literally Mn(VII) sitting beside neat O^2- point charges that hand over a whole electron — those oxidation states are a bookkeeping device, not real charges, and the Mn-O bonds are heavily covalent. The orbitals that the electron jumps between are delocalized molecular orbitals spread over the whole MnO4- unit; LMCT just names which end each orbital leans toward. Second, the line between a "ligand-field" band and a "charge-transfer" band is genuinely blurry in strongly covalent complexes, where every orbital is part metal and part ligand; the labels are useful, not absolute.
Once you have the idea, you start seeing charge transfer everywhere the world is loudly colored. The deep blue of lapis lazuli and ultramarine pigment is a charge-transfer band in trapped polysulfide ions. The blood-red of the iron-thiocyanate test, [Fe(SCN)]^2+, used to detect iron, is LMCT. Prussian blue, the first modern synthetic pigment, owes its intensity to an electron transferring between Fe(II) and Fe(III) sites bridged by cyanide. Whenever a transition-metal compound is not merely tinted but genuinely vivid — a stain, a pigment, a dye — your first guess should be charge transfer, not a polite little d-d band.