Charge-Transfer Spectra: LMCT and MLCT

Drop a colorless solution of Fe(III) salt into a colorless solution of thiocyanate. The mixture turns blood red instantly. No d-d transition can do that. Permanganate is an intense purple with no d electrons at all. There are none to promote between crystal-field levels. Ruthenium tris(bipyridine) is the workhorse chromophore of solar energy conversion and photoredox catalysis. It absorbs visible light with a molar absorptivity thousands of times larger than any spin-allowed d-d band. In every one of these cases the color is not d-d in origin. It comes from charge-transfer (CT) transitions. In these, an electron moves between orbitals that are predominantly ligand in character and orbitals that are predominantly metal in character. Charge-transfer bands matter because they are the loudest voice in the electronic spectrum of a coordination compound. Where d-d bands are Laporte-forbidden, they gain intensity only through vibronic borrowing or loss of centrosymmetry. They survive with molar absorptivities of 1–100 M⁻¹cm⁻¹. Charge-transfer transitions, by contrast, are fully allowed, both by parity and by spin. They routinely reach ε ~ 10³–10⁵ M⁻¹cm⁻¹. This single fact of intensity is diagnostic. An intense, broad, solvent-sensitive band in the UV-visible region is a near-certain sign of charge transfer. This recognition underlies practical applications from colorimetric metal analysis (the Fe(III)-thiocyanate test, the molybdenum blue phosphate assay). It also guides the design of chromophores for dye-sensitized solar cells and photocatalysts. The physical picture builds directly on the same molecular-orbital framework used for Selection Rules for Electronic Transitions and for understanding Intensity of d-d Bands and Relaxation Mechanisms. In an octahedral or other ligand-field complex, filled and empty orbitals are not purely "metal d" or "ligand p." Instead they are mixtures. A charge-transfer transition is simply an electric-dipole-allowed excitation. It goes between an orbital of predominantly one parentage and an orbital of predominantly the other. Ligand-to-metal charge transfer (LMCT) promotes an electron from a filled, mostly-ligand orbital. It moves into an empty or partially filled, mostly-metal d orbital. Metal-to-ligand charge transfer (MLCT) does the reverse. It promotes an electron from a filled, mostly-metal d orbital. That electron moves into an empty, mostly-ligand π* orbital. Both processes are governed by orbital overlap, energy matching, and oxidation-state/electronegativity trends. These trends can be predicted, and often manipulated by design, before ever recording a spectrum. This article develops the molecular-orbital origin of LMCT and MLCT bands. It derives the energy trends that let you predict which type of CT band a complex will show. It also shows roughly where that band will fall. The article works through the classic diagnostic cases — permanganate, ferric thiocyanate, and [Ru(bpy)₃]²⁺ — in full quantitative detail. It closes with the chemistry where charge-transfer transitions are put to work, from photoredox catalysis to blue pigments.

Key takeaways

LMCT: electron moves from filled, ligand-based MOs to empty/partly-filled metal d-based MOs; favored when the metal is highly oxidized/electronegative-poor and the ligand is easily oxidized (e.g., MnO₄⁻, FeCl₄⁻, [Fe(SCN)]²⁺).
MLCT: electron moves from filled, metal d-based (mostly nonbonding, "t₂g-like") MOs to empty ligand π* orbitals; favored for low-oxidation-state metals with π-acceptor ligands (bpy, phen, CO, CN⁻) — e.g., [Ru(bpy)₃]²⁺, [Fe(CN)₆]⁴⁻, M(CO)ₙ complexes.
Both transitions are Laporte- and (usually) spin-allowed, giving ε ≈ 10³–10⁵ M⁻¹cm⁻¹, one to four orders of magnitude more intense than d-d bands.
CT energy tracks periodic trends: LMCT drops (redshifts) as metal oxidation state rises or ligand basicity/reducing power rises; MLCT drops as metal is easier to oxidize (lower oxidation state, more electron-rich) and ligand π* is lower-lying (better acceptor).
Charge-transfer bands often show pronounced solvatochromism and can be resolved from d-d bands by their much larger ε, broader vibronic envelope, and characteristic blue/red shift with solvent polarity or ligand substitution.
Key idea: an intense visible/UV band with ε > 1000 M⁻¹cm⁻¹ signals a fully-allowed electron jump between orbitals of different atomic parentage (ligand vs. metal), and whether it is LMCT or MLCT is decided by which of "oxidize the ligand" or "reduce the metal" costs less energy.

Where this lives today

Dye-sensitized solar cells (DSSCs). The Grätzel-cell architecture uses ruthenium polypyridyl dyes (the classic "N3"/"N719" dyes, close relatives of $[Ru(bpy)_3]^{2+}$ with carboxylate-substituted bipyridines for anchoring to TiO₂) whose broad, intense MLCT absorption across the visible captures sunlight; the MLCT excited state — an electron promoted onto a ligand π* orbital that is electronically coupled to the TiO₂ conduction band — injects the excited electron directly into the semiconductor on a femtosecond-to-picosecond timescale, the key photophysical step of the whole device.

Photoredox catalysis. $[Ru(bpy)_3]^{2+}$ and its relatives ( $[Ir(ppy)_3]$ and cyclometalated Ir(III)/Ru(II) analogs) are the standard photocatalysts of modern visible-light-driven organic synthesis. The MLCT excited state is simultaneously a stronger reductant and a stronger oxidant than the ground state (because the promoted electron sits in a ligand-based orbital, effectively separating charge and making both the oxidized "hole" on the metal and the reduced electron on the ligand kinetically accessible for electron transfer to substrates) — a direct chemical consequence of the CT (not d-d) nature of the excited state.

Colorimetric analytical chemistry. The intense, highly diagnostic LMCT color of $[Fe(SCN)]^{2+}$ (blood-red, ε in the thousands) remains a standard qualitative/quantitative test for either Fe³⁺ or SCN⁻ in analytical laboratories; the molybdenum blue method for phosphate determination and various vanadium- and titanium-peroxide colorimetric assays rely on the same principle — an intensely colored LMCT or mixed-valence CT band whose absorbance is directly proportional to analyte concentration via Beer's law, exploiting the very large ε that CT transitions provide.

Pigments and materials. Prussian blue, $\text{Fe}_4[\text{Fe}(\text{CN})_6]_3$ , owes its deep blue color to an intervalence charge-transfer (IVCT) band — a close cousin of LMCT/MLCT in which an electron transfers between Fe(II) and Fe(III) centers bridged by cyanide, a mixed-valence phenomenon whose theoretical treatment (Robin-Day classification, Hush theory) builds directly on the same MO/energetics framework developed here. Chromate/dichromate yellows and oranges, and the permanganate-based purple used historically as a dye and oxidant indicator, are textbook LMCT pigments.

Bioinorganic chemistry and spectroscopic diagnosis. LMCT bands are used routinely to characterize metalloprotein active sites: the intense S(cysteinate) → Fe(III) LMCT bands of rubredoxins and Fe-S clusters (giving their characteristic brown/red color), and the O(tyrosinate) → Fe(III) LMCT bands of transferrin, are spectroscopic fingerprints for the identity and oxidation state of the metal-binding ligand set, often used alongside EPR and Mössbauer data to assign active-site structure without crystallography.

Phosphorescent OLED emitters. Heavy third-row metal (Ir(III), Pt(II), Os(II)) MLCT/mixed MLCT-ligand-centered excited states, strongly spin-orbit coupled (as previewed in Example 3 above), give efficient room-temperature phosphorescence exploited in the emissive layers of organic light-emitting diodes, harvesting both singlet and triplet excitons for near-unity internal quantum efficiency.

:::quiz[High-spin [Fe(H₂O)₆]³⁺ (d⁵) is pale yellow even though it has zero spin-allowed d-d transitions. What is the best explanation?]

d⁵ ions cannot absorb light at all
The color arises from an O(2p) → Fe(3d) LMCT transition, which is spin- and (effectively) orbital-allowed ✓
The color is due to a forbidden MLCT transition
Water ligands absorb light independently of the metal

Ask the tutor: In [Ru(bpy)₃]²⁺, why does the MLCT excited state act as both a better oxidant and a better reductant than the ground state, and how is this exploited in photoredox catalysis?

The physical picture

Begin from the molecular-orbital (MO) diagram of an octahedral complex, the same diagram used to explain why d-d bands are weak. Ligand σ-donor orbitals combine with metal s, p, and eg (dz²,dx²-y²) orbitals to give bonding σ and antibonding σ sets; if the ligands also have π symmetry orbitals (filled lone pairs on halides, O, S donors, or empty π on CO, CN⁻, bipyridine, phenanthroline), these combine with metal t₂g orbitals to give bonding/antibonding π and π sets. The "d orbitals" chemists draw in a crystal-field diagram are, more accurately, the largely metal-centered members of this MO manifold: an antibonding eg (σ) level and a t₂g level that is either metal-nonbonding, weakly π-bonding, or weakly π-antibonding depending on whether the ligand π orbitals are filled (donor) or empty (acceptor). Four classes of orbitals matter for CT spectra: 1. Filled, mostly-ligand orbitals — ligand lone pairs (σ or π) that combine only weakly with the metal, staying close to their free-ligand energy, slightly stabilized (bonding) by donation into the metal. 2. Filled, mostly-metal orbitals — the t₂g-like set (nonbonding or π-bonding combinations), holding the d electrons. 3. Empty, mostly-metal orbitals — the eg-like set (σ-antibonding, strongly destabilized) or empty t₂g if the metal is in a high oxidation state. 4. Empty, mostly-ligand orbitals — ligand π orbitals (as in CO, CN⁻, bpy, phen) that combine with filled metal t₂g to form a π/π pair, with the low-lying π remaining largely ligand in character but pulled down further by conjugation. LMCT is the transition (1) → (3): filled ligand orbital to empty (or partly filled) metal d orbital. MLCT is the transition (2) → (4): filled metal d orbital to empty ligand π. Both are transitions between orbitals of different parentage and, critically, of different symmetry character with respect to the metal-ligand bonding framework, so they are not subject to the same g → g Laporte restriction that suppresses d-d transitions in a centrosymmetric complex — the initial and final MOs typically have different parities or belong to representations connected by the dipole operator, making the transition orbitally allowed. Spin selection rules still apply: LMCT/MLCT bands are intense only when they do not require a spin flip.

Figure

Simplified MO energy ladder: LMCT vs MLCT origins

Which of these two transitions dominates a given complex's low-energy spectrum is decided by relative orbital energies, which in turn are set by oxidation state, electronegativity, and π-donor/acceptor character of the ligand — the periodic-trend content of the next section.

Figure

Periodic-trend map: which complexes show LMCT vs MLCT

The core ideas

The energy of a charge-transfer band can be organized, to first approximation, as an ionization/electron-affinity balance dressed by ligand-field and electron-repulsion corrections — the "optical electronegativity" approach pioneered by Jørgensen.

\tilde\nu_{\text{CT}} \;\approx\; \big[\chi_{\text{opt}}(\text{donor orbital}) - \chi_{\text{opt}}(\text{acceptor orbital})\big]\times 30{,}000\ \text{cm}^{-1}

where: $\tilde\nu_{\text{CT}}$ is the CT band energy in wavenumbers; $\chi_{\text{opt}}$ are Jørgensen's empirical optical electronegativities assigned to the ligand (donor, for LMCT) and to the metal in its actual oxidation state and coordination environment (acceptor, for LMCT); the constant 30,000 cm⁻¹ converts an electronegativity difference into an approximate transition energy, calibrated against measured LMCT maxima across many complexes. Reading: a bigger electronegativity mismatch between the electron-rich ligand and the electron-poor metal center means a lower-energy (redder) CT band, because it is energetically "cheaper" to move charge from a very electropositive donor site to a very electronegative acceptor site — exactly analogous to how a small HOMO–LUMO gap gives a low-energy transition in organic chromophores. A second, complementary relationship makes the oxidation-state and π-bonding dependence explicit for a family of isostructural complexes:

E_{\text{LMCT}} \;=\; E_{\text{ligand HOMO}\to\text{vacuum}} \;+\; \big(-\,\text{EA}_{\text{eff}}(M^{n+})\big) \;-\; \Delta_{\text{stab}}

where: $E_{\text{ligand HOMO}\to\text{vacuum}}$ is (loosely) the ionization energy of the coordinated ligand's donor orbital; $-\text{EA}_{\text{eff}}(M^{n+})$ is the effective electron affinity of the metal ion in that oxidation state and ligand field (larger for higher, more electron-poor oxidation states — Mn(VII) ≫ Mn(II)); $\Delta_{\text{stab}}$ collects the Coulombic/ligand-field stabilization gained once the electron actually sits on the metal, which lowers the observed transition energy below the naive atomic-orbital gap. Reading: raising the metal oxidation state (Fe²⁺ → Fe³⁺ → Fe⁴⁺, or Mn²⁺ → Mn⁷⁺) makes $-\text{EA}_{\text{eff}}$ more negative — the metal is a hungrier electron acceptor — which lowers $E_{\text{LMCT}}$ and redshifts (and usually intensifies visibility of) the LMCT band; this is why Fe(III) complexes are colored while analogous Fe(II) complexes with the same ligand often are not. For MLCT the mirror-image relationship holds:

E_{\text{MLCT}} \;\approx\; \text{IE}_{\text{eff}}(d^n,\ t_{2g}) \;-\; \text{EA}(\pi^*_{\text{ligand}}) \;-\; \Delta_{\text{stab}}'

where: $\text{IE}_{\text{eff}}(d^n, t_{2g})$ is the effective ionization energy of the filled metal t₂g-like orbital (smaller — easier to ionize — for lower oxidation states and more electron-rich metals, e.g., Ru(II) vs. Ru(III)); $\text{EA}(\pi^*_{\text{ligand}})$ is the electron affinity of the ligand's lowest π orbital (larger, i.e., π lower-lying and easier to fill, for more extended/conjugated π-acceptor ligands such as phenanthroline > pyridine, or for ligands substituted with electron-withdrawing groups); $\Delta_{\text{stab}}'$ is the analogous stabilization once the electron localizes in the ligand π system (partly relieved by solvent reorganization in polar media). Reading: lowering the metal's oxidation state (easier to oxidize) and choosing better π-acceptor ligands (lower π) both shrink $E_{\text{MLCT}}$ , redshifting the MLCT band — precisely the strategy used to tune [Ru(bpy)₃]²⁺-type photosensitizers by ligand substitution.

Derivation AWhy CT transitions are fully allowed while d-d transitions are not

Goal. Show, using group theory and the parity of the MOs involved, why an LMCT or MLCT transition typically has a transition dipole moment integral that survives (nonzero), whereas a pure d-d transition within the t₂g/eg manifold of a centrosymmetric complex vanishes to first order.

Step 1 — set up the transition moment integral. The intensity of any electric-dipole transition between initial state $\psi_i$ and final state $\psi_f$ is proportional to

|\boldsymbol\mu_{fi}|^2 = \left|\int \psi_f^{*}\,\hat{\boldsymbol\mu}\,\psi_i\,d\tau\right|^2 ,

which is nonzero only if the direct product of the representations of $\psi_i$ , $\hat{\boldsymbol\mu}$ (which transforms as $x,y,z$ ), and $\psi_f$ contains the totally symmetric representation of the point group.

Step 2 — apply this to d-d transitions in Oₕ. In an octahedral (centrosymmetric) complex, the metal d orbitals split into $t_{2g}$ and $e_g$ , both of which are gerade (g) under inversion, because d orbitals themselves are even functions ( $l=2$ ). The dipole operator $\hat{\boldsymbol\mu}$ transforms as $x,y,z$ , which is ungerade (u). The product $g \times u \times g = u$ , which is never totally symmetric ( $A_{1g}$ ) in a centrosymmetric group — the integral is forced to zero by the parity (Laporte) selection rule. d-d bands survive only because vibrations that destroy the inversion center (odd-parity normal modes) mix a little $u$ character into the wavefunctions, borrowing intensity vibronically; this is why d-d ε values stay in the 1–100 M⁻¹cm⁻¹ range, as covered under Intensity of d-d Bands and Relaxation Mechanisms.

Step 3 — apply the same logic to LMCT/MLCT. The relevant ligand orbitals — σ or π lone pairs on the donor atoms, or ligand π systems delocalized over the ligand framework — do not, in general, have the same g/u symmetry classification as the pure metal d set, because they are built from p orbitals on atoms sitting off the inversion center (or, for a set of six equivalent ligand donor orbitals arranged octahedrally, they decompose into both g and u combinations, e.g., $a_{1g}+e_g+t_{1u}$ for six σ donors). Crucially, ligand-based orbitals of $t_{1u}$ (u) symmetry exist alongside the metal-based $t_{2g}$ / $e_g$ (g) orbitals. A transition from a $t_{1u}$ ligand orbital to a $t_{2g}$ or $e_g$ metal orbital has product $u \times u \times g = g$ ... more carefully: the dipole operator itself is $t_{1u}$ , so we need (ligand $t_{1u}$ ) × ( $t_{1u}$ , dipole) × (metal $t_{2g}$ or $e_g$ ) to contain $A_{1g}$ . Because $t_{1u}\times t_{1u} = a_{1g}+e_g+t_{1g}+t_{2g}$ , and this contains $e_g$ and $t_{2g}$ pieces that directly match the metal acceptor orbital's symmetry, the triple product does contain $A_{1g}$ : the transition is orbitally allowed. The essential point is that ligand donor/acceptor orbitals commonly carry odd*-parity representations that metal d orbitals structurally lack, so a ligand → metal or metal → ligand jump can connect $g$ and $u$ partners exactly as required for an allowed $u$ -type dipole transition, unlike a pure $g\to g$ d-d jump.

Step 4 — spin selection rule and consequence for intensity. Provided the CT transition does not require a change in total spin quantum number $S$ (true for the great majority of LMCT/MLCT transitions in diamagnetic or high-symmetry paramagnetic complexes), both the parity and spin rules are satisfied, and the transition dipole integral is not forced to zero by symmetry. The observed oscillator strength is then limited only by the actual orbital overlap and mixing coefficients, typically giving

\varepsilon_{\text{CT}} \sim 10^3\text{–}10^5\ \text{M}^{-1}\text{cm}^{-1},

one to four orders of magnitude above vibronically-borrowed d-d intensities — the empirical signature used throughout the rest of this article to identify CT bands by eye.

Go deeper · Derivation B · 2 minPredicting LMCT energy trends across an oxidation-state series

Derivation BPredicting LMCT energy trends across an oxidation-state series

Goal. Derive, from the effective nuclear charge and crystal-field stabilization felt by the transferred electron, why LMCT energies decrease (redshift) as metal oxidation state increases, using the isoelectronic/isostructural halide series $[MX_6]^{n-}$ as the worked framework, then check the limiting case of a very electropositive, low-oxidation-state metal.

Step 1 — identify the two energy costs. An LMCT transition removes an electron from a ligand-centered orbital (cost = ionization energy of that orbital while ligated, $I_L$ ) and places it into a metal-centered acceptor orbital (this releases energy equal to the effective electron affinity of the metal center in oxidation state $n{+}$ , $A_M^{(n)}$ , since the metal "wants" the extra electron density). To first order,

E_{\text{LMCT}} \approx I_L - A_M^{(n)} .

Step 2 — express $A_M^{(n)}$ in terms of oxidation state. For a given metal, $A_M^{(n)}$ increases (more energy released) as $n$ increases, because a higher formal charge means a larger effective nuclear charge $Z_{\text{eff}}$ felt at the metal center and hence a stronger pull on any additional electron density. Empirically, across a series such as Fe²⁺/Fe³⁺/Fe⁴⁺ or Mn²⁺/…/Mn⁷⁺ with the same ligand, each unit increase in oxidation state increases $A_M^{(n)}$ by a roughly constant increment $\delta$ (tens of thousands of cm⁻¹, reflecting the steep rise in successive ionization energies of the free ion, moderated by ligand-field covalency):

A_M^{(n+1)} \approx A_M^{(n)} + \delta .

Step 3 — combine to get the oxidation-state trend. Substituting,

E_{\text{LMCT}}^{(n+1)} - E_{\text{LMCT}}^{(n)} \approx -\,\delta < 0 ,

so each unit increase in metal oxidation state redshifts (lowers the energy of) the LMCT band by roughly $\delta$ . This is exactly the trend seen going from colorless (or pale) Mn(II) aqua/halide complexes, whose lowest LMCT bands lie deep in the UV, to intensely colored Mn(VII) as permanganate, whose O(2p) → Mn(3d) LMCT band sits at about 18,000–20,000 cm⁻¹ (≈ 525 nm, giving the characteristic purple color) — several $\delta$ -increments lower in energy than a hypothetical Mn(II)–oxo LMCT band would be.

Step 4 — limiting case check. Consider the limit of a very electropositive, low-oxidation-state metal, e.g. an alkaline-earth-like $M^{2+}$ with a very weakly acidic, essentially closed-shell electronic acceptor. Here $A_M^{(n)} \to$ small (little driving force to accept ligand electron density), so

E_{\text{LMCT}} \to I_L \ \ (\text{large, deep-UV}),

consistent with the fact that complexes of hard, low-charge-density, high-oxidation-resistance metals (e.g., alkali and alkaline-earth aqua/halo complexes) show no visible LMCT bands at all — their would-be CT transitions are pushed into the vacuum-UV, well outside the range that produces color. This correctly reproduces the everyday observation that NaCl and MgCl₂ solutions are colorless while FeCl₃ solutions are pale yellow (LMCT tailing into the visible) and permanganate is intensely purple (LMCT squarely in the visible) — the same $I_L - A_M^{(n)}$ balance, just evaluated at very different $A_M^{(n)}$ .

Worked examples

Example 1 (foundational): distinguishing LMCT, MLCT, and d-d by counting and by intensity in $[Fe(H_2O)_6]^{3+}$ vs. $[Fe(CN)_6]^{4-}$ . Take high-spin $[Fe(H_2O)_6]^{3+}$ (Fe³⁺, d⁵, weak-field). Water is a poor π-donor/π-acceptor and only a modest σ-donor; the metal is in a fairly high (+3) oxidation state, and there are no low-lying ligand π orbitals available to accept an electron from the metal. That rules out MLCT (no good acceptor) and predicts, via the LMCT energy argument above, a moderately low-lying LMCT band because Fe³⁺ is a fairly good electron acceptor even without exotic ligands. Experimentally, $[Fe(H_2O)_6]^{3+}$ shows a very pale, yellowish color from an O(2p) → Fe(3d) LMCT tail beginning around 40,000 cm⁻¹ (250 nm) trailing into the near-UV/visible edge, superimposed on weak, spin-forbidden d-d bands (ε ~ 0.1–1 M⁻¹cm⁻¹, high-spin d⁵ has no spin-allowed d-d transitions at all, by the same reasoning as <a class="wiki-link" href="/wiki/chem-inorg-selection-rules-transitions" data-track="wiki_article.body.wiki_link">Selection Rules for Electronic Transitions</a>). The pale color of "ferric" solutions is overwhelmingly LMCT, not d-d, precisely because d⁵ high-spin has zero spin-allowed d-d intensity to compete with. Now take $[Fe(CN)_6]^{4-}$ (Fe²⁺, d⁶, low-spin, strong-field). Here the metal oxidation state is lower (+2, easier to oxidize) and the ligand, cyanide, is an excellent π-acceptor with a low-lying empty π orbital. Both conditions favor MLCT over LMCT: the metal is now a good electron donor rather than a hungry acceptor, and there is a low-energy ligand orbital ready to receive the electron. Indeed, $[Fe(CN)_6]^{4-}$ is pale yellow from a moderately weak, higher-energy Fe(t₂g) → CN⁻(π*) MLCT band; strengthening the π-acceptor (replacing some CN⁻ by better acceptors, or oxidizing to Fe³⁺ in $[Fe(CN)_6]^{3-}$ , which shifts the balance back toward LMCT contributions) is exactly how the classic Prussian blue chemistry — mixed-valence Fe(II)/Fe(III) cyanide bridges — generates its intense intervalence/CT color, discussed further in the applications below. The qualitative lesson: swapping a σ-only, poor-π ligand (H₂O) for a strong π-acceptor (CN⁻) while simultaneously lowering the metal oxidation state (Fe³⁺ → Fe²⁺) flips the dominant CT type from LMCT to MLCT — both ligand and oxidation-state changes push in the same direction.

Example 2 (applied): estimating the LMCT energy of permanganate, $MnO_4^-$ , from optical electronegativities. Permanganate is Mn(VII), formally d⁰ — there are no d electrons at all, so its intense purple color (ε ≈ 2,300 M⁻¹cm⁻¹ at λmax ≈ 525 nm) must be charge transfer, specifically an O(2p, filled π/σ) → Mn(3d, empty, largely $e$ symmetry in the tetrahedral field) LMCT transition. Using Jørgensen's optical electronegativity scale, oxide/hydroxide-type oxygen donors have $\chi_{\text{opt}}(\text{O}) \approx 3.2$ , and Mn(VII) — an extremely electron-poor, high-oxidation-state center — has $\chi_{\text{opt}}(\text{Mn}^{7+}) \approx 1.9$ (values consistent with Jørgensen's original tabulation for early/high-oxidation-state transition metals). Applying the optical electronegativity relation:

\tilde\nu_{\text{CT}} \approx \big[\chi_{\text{opt}}(\text{O}) - \chi_{\text{opt}}(\text{Mn}^{7+})\big]\times 30{,}000\ \text{cm}^{-1}

\tilde\nu_{\text{CT}} \approx \big[3.2 - 1.9\big]\times 30{,}000\ \text{cm}^{-1} = 1.3 \times 30{,}000\ \text{cm}^{-1}

\tilde\nu_{\text{CT}} \approx 39{,}000\ \text{cm}^{-1}.

This overestimates the lowest, most intense LMCT band (observed near 18,500 cm⁻¹, 540 nm) because $MnO_4^-$ actually has several O → Mn LMCT transitions of differing symmetry (the $t_1 \to e$ and $t_1 \to t_2$ components split by tetrahedral ligand-field and exchange effects), and the simple electronegativity-difference formula is calibrated as an average/upper picture across a whole family rather than a precise predictor for any one transition; the lowest-energy, symmetry-allowed component falls well below the naive electronegativity gap once the actual $t_1(\pi) \to e(d)$ orbital energies (rather than atomic electronegativities) are used. The value of the calculation is qualitative and comparative: it correctly places $MnO_4^-$ 's LMCT deep in the visible (compare $CrO_4^{2-}$ , Cr(VI), $\chi_{\text{opt}} \approx 1.7$ , predicted and observed to absorb at somewhat higher energy, in the near-UV/violet, giving chromate's yellow rather than purple color) and correctly predicts that replacing Mn(VII) with a less oxidizing, lower-charge center (e.g., hypothetical " $MnO_4^{3-}$ ", Mn(V)) would blue-shift the LMCT band, consistent with the oxidation-state trend derived above. As a limiting-case check: setting the electronegativity difference to zero (a hypothetical donor and acceptor of identical optical electronegativity) predicts $\tilde\nu_{\text{CT}} \to 0$ — the electron would move at essentially no energy cost, i.e., the "ligand" and "metal" orbitals become degenerate and charge transfer becomes barrierless resonance rather than a discrete absorption band, which is the correct qualitative limit (no CT band forms because there is no gap to absorb into).

Example 3 (exam/competition level): rationalizing the MLCT band shift on going from $[Ru(bpy)_3]^{2+}$ to $[Ru(bpy)_2(CN)_2]$ and predicting the direction of shift for $[Os(bpy)_3]^{2+}$ . $[Ru(bpy)_3]^{2+}$ (Ru²⁺, low-spin d⁶, $O_h$ -like $D_3$ symmetry) shows its famous, intense MLCT absorption at λmax ≈ 452 nm (ε ≈ 14,600 M⁻¹cm⁻¹), assigned to $\text{Ru}(4d, t_{2g}) \to \text{bpy}(\pi^*)$ excitation. Replace two bpy ligands with two cyanide ligands to give $[Ru(bpy)_2(CN)_2]$ : cyanide is a strong σ-donor/π-acceptor but a much weaker π-acceptor than the low-lying, extended π system of bipyridine, and it also raises electron density at ruthenium (net σ-donation without matching π-back-bonding demand). Two effects both push the same way: the Ru center becomes more electron-rich (lower $\text{IE}_{\text{eff}}$ of the $t_{2g}$ set — easier to oxidize/ionize the metal electron) — which by itself would redshift MLCT — while the remaining acceptor character of the complex is now split between two different, weaker π manifolds. Experimentally the dominant bpy-based MLCT band in the cyanide-substituted complex is observed to redshift substantially relative to $[Ru(bpy)_3]^{2+}$ (into the 480–500 nm region for related $[Ru(bpy)_2(CN)_2]$ -type complexes), consistent with the metal-donor-energy term dominating the shift once a strong σ-donor/weak π-competitor like CN⁻ replaces a bpy ligand. Now predict $[Os(bpy)_3]^{2+}$ relative to $[Ru(bpy)_3]^{2+}$ , same ligand set, one row down the periodic table (3rd-row vs. 2nd-row d⁶ metal). Two competing periodic effects operate: (i) 5d orbitals are more diffuse and higher in energy than 4d orbitals of the same formal charge, so the filled $t_{2g}$ -like HOMO of Os(II) sits higher (smaller ionization energy) than that of Ru(II) — this lowers $\text{IE}_{\text{eff}}$ and should redshift the MLCT band; (ii) increased spin-orbit coupling in the third row (much larger $\zeta$ for 5d than 4d) mixes singlet and triplet MLCT states more effectively, broadening bands and enabling additional formally spin-forbidden MLCT absorption at lower energy that has appreciable intensity (unlike the negligible spin-forbidden bands in first-row d-d spectra). Both effects point the same way: $[Os(bpy)_3]^{2+}$ is observed to absorb at lower energy (redshifted, extending further into the visible/near-IR, λmax around 480 nm plus additional lower-energy shoulders) than $[Ru(bpy)_3]^{2+}$ , and shows a more prominent, spin-orbit-enabled lower-energy MLCT/phosphorescence manifold — the correct qualitative prediction from the periodic trend in $\text{IE}_{\text{eff}}$ combined with heavy-atom spin-orbit coupling, the same effect responsible for Os(II) polypyridyl complexes' strong phosphorescent (formally spin-forbidden) emission used in OLED phosphors.

Limiting cases, scope & common traps

Situation	What happens to CT bands	Why
Metal oxidation state ↑ (isostructural series)	LMCT redshifts (lower energy); may become visible	Higher $Z_{\text{eff}}$ /effective electron affinity of metal, per the derivation above
Metal oxidation state ↓, ligand is π-acceptor	MLCT redshifts	Metal $t_{2g}$ easier to ionize (lower IE); more electron density available to donate
Ligand is a strong π-donor (O²⁻, S²⁻, halide) with high-valent metal	LMCT dominates, often intense and low-energy	High-lying filled ligand π orbital, hungry metal acceptor
Ligand is a strong π-acceptor (CO, CN⁻, bpy, phen) with low-valent, electron-rich metal	MLCT dominates	Low-lying empty ligand π*, easily ionized metal d electrons
$d^0$ or $d^{10}$ metal center	Only LMCT (or ligand-ligand/intraligand) possible; MLCT impossible for $d^0$ (no d electrons to donate); LMCT impossible in the usual sense for $d^{10}$ into filled d shell but MLCT can occur from filled d into ligand π*	No filled d electrons (d⁰) or no empty d acceptor states (d¹⁰, but MLCT still needs the d-shell to donate, which is fine for d¹⁰)
Solvent polarity increases	MLCT bands typically blue-shift in more polar solvents (ground state less polar, more stabilized in polar solvent, than the more charge-separated excited state is destabilized — solvatochromism direction depends on relative dipole change)	Differential solvation of ground vs. CT excited state; sign of the shift diagnoses the direction of charge flow
CT and d-d bands overlap in energy	The much larger ε of CT dominates the observed spectral shape; underlying d-d structure may be masked or only visible as shoulders	Order-of-magnitude intensity difference (10³–10⁵ vs. 1–100 M⁻¹cm⁻¹)

The chemist behind it

The systematic, quantitative language for charge-transfer spectra — optical electronegativities, the recognition that intense visible bands in transition-metal compounds are frequently not d-d at all — was built up chiefly through the work of the Danish inorganic spectroscopist Christian Klixbüll Jørgensen in the 1950s–60s, working alongside and in dialogue with the crystal-field/ligand-field theorists of the era (notably Carl Ballhausen, with whom Jørgensen co-authored foundational ligand-field texts). Jørgensen synthesized decades of scattered spectroscopic observations on halide, oxide, and cyanide complexes into a single predictive scale — assigning each ligand and each metal ion (in each oxidation state) an "optical electronegativity" extracted directly from measured CT band positions — turning what had been a collection of curiosities (why is permanganate purple? why does ferric thiocyanate turn red?) into a coherent, periodic-trend-based theory that could predict new CT band positions before they were measured.

Origins · 1 minChristian Klixbüll Jørgensen and how this was discovered

Christian Klixbüll Jørgensen

1931–2001 · Danish inorganic and theoretical chemist working at Cyanamid European Research Institute (Geneva) and later the University of Geneva, in the era when ligand-field theory was being extended beyond simple crystal-field electrostatics to a full molecular-orbital treatment of transition-metal spectra

History

- 1953 — Ilse and Hartmann, and independently others, apply developing ligand-field ideas to interpret transition-metal absorption spectra, setting the stage for distinguishing CT from d-d bands.

- 1957–1962 — Jørgensen systematically compiles LMCT band energies across metal/ligand combinations and introduces the optical electronegativity concept to place them on a common predictive scale.

- 1962 — Jørgensen and Ballhausen's Molecular Orbital Theory consolidates the MO picture that explains why CT bands are orbitally allowed while d-d bands are not.

- 1970s–1980s — the MLCT chemistry of

[Ru(bpy)_3]^{2+}

and related polypyridyl complexes is developed intensively (Meyer, Balzani and coworkers) as excited-state electron-transfer photochemistry emerges as a field.

- 1990s–present — CT chromophore design becomes central to dye-sensitized solar cells, photoredox catalysis, and phosphorescent OLED emitters built on Ru(II)/Os(II)/Ir(III) MLCT and related excited states.

What is the practical difference between "charge transfer" and "d-d" bands when looking at a UV-vis spectrum?

The clearest practical signal is intensity: charge-transfer bands typically show molar absorptivities ε in the thousands to tens of thousands (M⁻¹cm⁻¹), while spin-allowed d-d bands rarely exceed a few hundred and spin-forbidden d-d bands are far weaker still (often <1 M⁻¹cm⁻¹). CT bands are also usually broader, often show pronounced solvatochromism (shifting with solvent polarity because the ground and CT excited states have very different dipole moments/charge distributions), and their positions correlate strongly with ligand electronegativity and metal oxidation state rather than simply with crystal-field splitting parameters like Δ_o.

Can a single complex show both LMCT and MLCT bands?

Yes, and this is common. A complex with both a good π-donor and a good π-acceptor coordinated to a metal of intermediate oxidation state can show both types, often at different energies. Mixed-ligand complexes such as $[Ru(bpy)_2Cl_2]$ can display an MLCT band (Ru → bpy π) alongside a distinct, usually higher-energy LMCT band (Cl⁻ → Ru), because bipyridine and chloride play opposite roles (acceptor vs. donor) toward the same metal center. Careful comparison of how each band shifts on ligand substitution — LMCT tracking ligand donor strength, MLCT tracking ligand π-acceptor strength — is how the two are experimentally distinguished.

Why does permanganate's color come from LMCT if manganese has no d electrons in MnO₄⁻?

Precisely because Mn(VII) is d⁰, there are no d electrons available to be promoted (ruling out d-d and MLCT transitions, both of which require filled or partially filled d orbitals as the electron source). The only electrons available to be excited into the empty, low-lying Mn(3d)-based acceptor orbitals are the filled oxygen 2p (π and σ) lone pairs, so the intense purple color must be an O(2p) → Mn(3d) ligand-to-metal charge-transfer transition. This is one of the cleanest textbook demonstrations that color does not require partially-filled d orbitals at all.

Why do MLCT bands in ruthenium(II) polypyridyl complexes shift with substituents on the bipyridine ligand?

Electron-withdrawing substituents (e.g., ester or carboxylic acid groups, as in DSSC dyes) lower the energy of the ligand π orbital that accepts the excited electron, directly shrinking the metal-to-ligand energy gap and redshifting the MLCT band; electron-donating substituents (e.g., methyl or amino groups) raise the π energy and blue-shift it. This tunability is exactly why bipyridine/phenanthroline ligand design is such a powerful lever for engineering the absorption window of ruthenium- and osmium-based photosensitizers and photocatalysts.

Is charge-transfer intensity always higher than d-d intensity, with no exceptions?

As a strong rule of thumb, yes for spin-allowed comparisons, but there are subtleties. A CT transition that happens to require a spin change (rare, but possible in some high-spin systems) can be substantially weakened, sometimes down toward d-d-like intensities. Conversely, some non-centrosymmetric or strongly distorted complexes (e.g., tetrahedral d-d transitions, which lack a strict inversion center to begin with) can show unusually intense, Laporte-allowed-in-part d-d bands (ε up to a few hundred), narrowing—but not closing—the gap with CT intensities. The general ordering (CT ≫ d-d) remains reliable for diagnostic purposes across the vast majority of real complexes.

How does solvent polarity help identify whether a band is CT in character?

Charge-transfer transitions physically move electron density from one region of the molecule to another, creating (or reversing) a substantial dipole moment change between ground and excited state; this makes CT bands strongly solvatochromic, shifting position measurably as solvent polarity changes. D-d transitions, by contrast, redistribute electron density only subtly within the same metal center and show comparatively little solvatochromism. Recording a spectrum in two or three solvents of very different polarity (e.g., water vs. acetonitrile vs. dichloromethane) and looking for a substantial shift is a fast, low-cost experimental check that complements the intensity criterion.

Does a complex need to be octahedral for LMCT/MLCT to occur?

No. LMCT and MLCT are general MO phenomena that occur in tetrahedral (MnO₄⁻, CrO₄²⁻), square-planar (many Pt(II)/Pd(II) complexes), and organometallic carbonyl/cyclopentadienyl complexes of essentially any geometry, because the underlying requirement is simply the existence of filled and empty orbitals of appropriately different parentage and matching symmetry for an allowed dipole transition — a requirement met across the full range of coordination geometries, not a special feature of octahedral symmetry.