Intensity of d-d Bands and Relaxation Mechanisms

Molar absorptivity of d-d bands remains a fast, non-destructive diagnostic in modern inorganic and bioinorganic chemistry: a chemist can often tell, from $\varepsilon$ alone, whether a newly synthesized complex is genuinely octahedral, tetrahedral, or has undergone a symmetry-lowering distortion, without needing a crystal structure. In bioinorganic chemistry, metalloprotein active sites (blue copper proteins, non-heme iron enzymes, cobalamin/vitamin B12 derivatives) are routinely characterized by comparing observed $\varepsilon$ values against these octahedral/tetrahedral/distorted benchmarks to infer coordination number and local symmetry around the metal center in the absence of a structure, or to track a spectroscopically silent-to-visible transition upon substrate binding. Solid-state chemists use the same reasoning in reverse: a strongly intensified, and strongly temperature-dependent, d-d band is taken as direct evidence of significant vibronic (or Jahn–Teller-related) distortion in a nominally centrosymmetric lattice site, informing models of dynamic versus static Jahn–Teller behavior in materials such as manganite and cuprate oxides relevant to colossal magnetoresistance and high-temperature superconductivity research. In gemology and pigment chemistry, the pale colors of high-spin Mn(II) and Fe(III) minerals versus the vivid colors of tetrahedrally coordinated Co(II) or Cu(II) impurities in glasses and ceramics are direct, everyday manifestations of exactly this Laporte/spin-selection-rule bookkeeping. Finally, the entire framework here is the necessary stepping-stone to interpreting fully allowed Charge-Transfer Spectra: LMCT and MLCT bands correctly: recognizing that a very intense visible/UV band ($\varepsilon > 1000$) cannot be a simple d-d transition, precisely because no vibronic-only mechanism can deliver that much intensity, is often the first clue that charge-transfer character — not ligand-field excitation — is responsible for a compound's color.

Ask the tutor: How does spin–orbit coupling quantitatively "borrow" intensity for a spin-forbidden transition, using the same perturbation-theory logic applied here to vibronic coupling?

Begin with why the rule is exact in the static picture. The electric-dipole transition moment between two stationary states $\psi_i$ and $\psi_f$ is

where $\hat{\boldsymbol{\mu}} = -e\sum_k \mathbf{r}_k$ is the electric dipole operator (odd under inversion: $\hat{\mathbf{r}} \to -\hat{\mathbf{r}}$, so it transforms as $u$, ungerade). If the molecule has a center of inversion, every orbital and every electronic state can be classified as gerade (g, even) or ungerade (u, odd) under that inversion. The integral $\int \psi_f^* \hat{\boldsymbol{\mu}} \psi_i\, d\tau$ is non-zero only if the product of parities of $\psi_f$, $\hat{\boldsymbol{\mu}}$, and $\psi_i$ is gerade overall — because only a totally symmetric (g) integrand can integrate to a non-zero number over all space. Since $\hat{\boldsymbol{\mu}}$ is always u, we need $\Gamma(\psi_f)\times u \times \Gamma(\psi_i) = g$, i.e. $\psi_i$ and $\psi_f$ must have opposite parity (one g, one u). This is the Laporte rule: g↔u transitions are electric-dipole allowed; g↔g and u↔u are forbidden. All five d orbitals ($d_{xy}, d_{xz}, d_{yz}, d_{x^2-y^2}, d_{z^2}$) are gerade — they are even functions under inversion, unlike p orbitals, which are odd. Consequently, any transition within a $d^n$ configuration in a centrosymmetric ligand field ($O_h$, $D_{4h}$, etc.) is g→g, and is Laporte-forbidden at the static, equilibrium geometry. This is exact — not approximate — as long as the molecule truly possesses $i$ symmetry and the transition mechanism considered is pure electric dipole. Yet spectra of octahedral $d$-block complexes show real, measurable d-d bands. Two independent relaxation channels are available. First, vibronic coupling: the molecule's normal modes include some of ungerade symmetry (e.g., the $T_{1u}$ asymmetric stretch of an $M\text{L}_6$ octahedron). At any instant other than the equilibrium geometry, the nuclear displacement along a $u$-mode transiently removes the center of inversion. The electronic wavefunctions computed at this distorted geometry are no longer pure g or u — they acquire a small admixture of the opposite parity, because the perturbing (odd) nuclear displacement mixes states of different electronic parity (the Herzberg–Teller mechanism, formally a vibronic coupling term in the Born–Oppenheimer expansion). The transition moment, evaluated including this admixture, is non-zero, though small (it scales with the amplitude of the odd-mode displacement, i.e. with a vibrational coordinate rather than being switched on/off statically). Second, static loss of centrosymmetry: if the complex genuinely lacks an inversion center in its equilibrium structure (tetrahedral $T_d$, or an octahedron distorted by unsymmetrical ligands, e.g. $cis$-$[\text{CoCl}_2(\text{NH}_3)_4]^+$), then d and p orbitals of the metal can mix directly by symmetry at the equilibrium geometry itself — no vibrational assistance needed. This mixing is generally larger than the vibronic admixture because it is not restricted to small instantaneous displacements, and it is temperature-independent. Tetrahedral complexes are consequently 10–100× more intensely colored than octahedral ones of comparable metal and ligands, all else equal.

The observable intensity of an absorption band is quantified by the (dimensionless) oscillator strength $f$, or equivalently the integrated molar absorptivity, related to the transition dipole moment by

where: $\tilde\nu$ is wavenumber (cm⁻¹), $\varepsilon(\tilde\nu)$ is the molar absorptivity at that wavenumber, and the integral runs over the whole band. This connects a directly measurable spectroscopic quantity ($\varepsilon$, from Beer–Lambert law fits) to $f$, which is proportional to $|\boldsymbol{\mu}_{fi}|^2$ and hence to what group theory controls. The transition dipole moment itself is

where: $\psi_i, \psi_f$ are the initial/final electronic (or vibronic) wavefunctions and $\hat{\boldsymbol{\mu}}$ is the electric dipole operator. Reading it: this integral is the "overlap," weighted by the odd dipole operator, between the two states; it is exactly zero whenever symmetry forbids a non-zero totally-symmetric integrand, which is the mathematical content of every selection rule. The group-theoretic non-vanishing condition is

where: $\Gamma(X)$ denotes the irreducible representation spanned by $X$ in the molecular point group, and $\otimes$ is the direct (Kronecker) product of representations. Reading it: a transition is allowed exactly when this triple product contains the totally symmetric representation; because $\hat{\boldsymbol{\mu}}$ transforms as $u$ in any centrosymmetric group, satisfying this condition for a g→g pair is impossible without an additional (vibronic) factor entering the product. For vibronic coupling, the effective (Herzberg–Teller) selection rule becomes a product over three factors — electronic, vibrational, and dipole operator symmetries:

where: $Q_k$ is the normal coordinate of the $k$-th vibrational mode through which the coupling occurs. Reading it: even though $\Gamma(\psi_f^{el})\otimes \Gamma(\hat{\boldsymbol{\mu}})\otimes\Gamma(\psi_i^{el})$ alone is odd (u) and cannot contain $A_{1g}$, multiplying in an ungerade normal mode $Q_k$ (u × u = g) restores the possibility of a totally symmetric product, switching the transition on with an intensity that scales with the vibrational amplitude, hence with temperature.

Example 1 (foundational): comparing $\varepsilon$ for an octahedral vs. tetrahedral Co(II) complex. Aqueous $[\text{Co(H}_2\text{O})_6]^{2+}$ (octahedral, high-spin d⁷, $O_h$, centrosymmetric) shows its spin-allowed d-d band ($^4T_{1g}(F) \to {}^4T_{1g}(P)$) near 512 nm with $\varepsilon_{\max} \approx 5~\text{M}^{-1}\text{cm}^{-1}$. The tetrahedral complex $[\text{CoCl}_4]^{2-}$ ($T_d$, no inversion center) shows its corresponding d-d band ($^4A_2 \to {}^4T_1(P)$) near 690 nm with $\varepsilon_{\max}\approx 600~\text{M}^{-1}\text{cm}^{-1}$. Apply the reasoning developed above. In $[\text{Co(H}_2\text{O})_6]^{2+}$, the point group is (idealized) $O_h$, which contains $i$; the ground and excited terms are both built from $d^7$ configurations and so are gerade; the transition is Laporte-forbidden and can only acquire intensity vibronically, through admixture of ungerade vibrational character — consistent with the small observed $\varepsilon \approx 5$. In $[\text{CoCl}_4]^{2-}$, the point group is $T_d$, which does not contain $i$. There is no inversion operation to classify orbitals by parity in the first place; d and p atomic orbitals on cobalt transform under compatible irreducible representations of $T_d$ ($e$ and $t_2$ for d; $t_2$ for p), so metal d orbitals can mix directly with metal p character (and with ligand orbitals of $t_2$ symmetry) already at the equilibrium geometry. This static, symmetry-allowed d–p mixing switches on a transition dipole with no vibrational assistance required, giving an intensity roughly two orders of magnitude larger:

As a limiting-case check: if the tetrahedral complex's local site symmetry were artificially restored to have a center of inversion (which $T_d$ never does, but consider the thought experiment of forcing $C_i$ symmetry by symmetric ligand placement), the same derivation used above for $O_h$ would apply verbatim, and the ratio would collapse toward 1 — confirming that the entire ~120-fold enhancement traces to the absence of $i$, not to any difference in metal, oxidation state, or ligand field strength per se (both are high-spin Co(II)).

Example 2 (applied): identifying vibronic vs. static origin from a real spectrum. The $[\text{Mn(H}_2\text{O})_6]^{2+}$ ion (high-spin d⁵, $O_h$) has famously pale, almost colorless solutions, with the sharp, very weak bands (e.g. near 25,000 cm⁻¹, $^6A_{1g}\to {}^4E_g, {}^4A_{1g}$) showing $\varepsilon \approx 0.01$–$0.03~\text{M}^{-1}\text{cm}^{-1}$ — two to three orders of magnitude weaker than the $\text{Co}^{2+}$ example above, despite both being octahedral, high-spin, first-row transition metal aqua ions. The explanation requires stacking two independent selection rules. First, the Laporte rule applies exactly as before ($O_h$, d-orbitals all gerade): this alone would give $\varepsilon$ on the order of a few $\text{M}^{-1}\text{cm}^{-1}$, comparable to the $\text{Co}^{2+}$ case. But high-spin d⁵ Mn(II) has a $^6A_{1g}$ ground term (all five d electrons unpaired, maximum spin $S = 5/2$), and every excited term arising from the $d^5$ configuration at accessible energies has a different total spin ($S = 3/2$ or lower, since flipping any electron necessarily changes the total spin multiplicity from a maximally-spin-aligned starting configuration). The transitions are therefore also spin-forbidden ($\Delta S \neq 0$), which multiplies in a further intensity-reducing factor (spin-forbidden bands typically carry $10^{-2}$ to $10^{-3}$ of the intensity of a comparable spin-allowed band, because the spin-forbidden mechanism requires spin–orbit coupling to admix a small amount of different-multiplicity character, an effect that scales roughly as $(\zeta/\Delta E)^2$ with the spin–orbit coupling constant $\zeta$). Quantitatively, taking the vibronically-allowed, spin-allowed benchmark as $\varepsilon_0 \approx 5$–$10~\text{M}^{-1}\text{cm}^{-1}$ (the Co(II) aqua-ion order of magnitude) and a typical spin-forbiddenness attenuation factor of order $10^{-3}$ for a first-row 3d⁵ ion (light atom, small $\zeta$), the predicted order of magnitude is

matching, to within the expected order-of-magnitude scatter of such estimates, the observed $\varepsilon \approx 0.01$–$0.03~\text{M}^{-1}\text{cm}^{-1}$. As a limiting-case check: if spin–orbit coupling were switched off entirely ($\zeta \to 0$, an idealization sometimes approached in very light main-group-like systems), spin-forbidden transitions would have exactly zero intensity regardless of vibronic assistance, since the spin selection rule is independent of, and multiplicative with, the parity/vibronic mechanism — consistent with Mn(II) solutions being far paler than any other divalent first-row aqua ion.

Example 3 (exam/competition level): estimating the vibronic intensity ratio between two octahedral complexes with different M–L bond stiffness. Two octahedral d³ complexes, $[\text{Cr(NH}_3)_6]^{3+}$ and $[\text{Cr(CN)}_6]^{3-}$, both show the same type of Laporte-forbidden, spin-allowed $^4A_{2g}\to{}^4T_{2g}$ transition, vibronically activated through the same $t_{1u}$ asymmetric-stretch mode. Suppose the effective force constant for that mode is stiffer in the cyanide complex by a factor of 2.5 (cyanide is a much stronger-field, more covalently bonded ligand, giving shorter/stiffer Cr–C bonds), while the reduced mass for the mode is essentially the same in both cases (dominated by the light donor atom's motion relative to the heavy metal). Estimate the ratio of vibronic intensities expected at room temperature, in the classical (high-temperature) limit. From Step 3 of the temperature-dependence derivation, in the classical limit the mean-square vibrational displacement is

where $k_k = \mu_k\omega_k^2$ is the effective force constant of the mode (this is just the classical equipartition result for a harmonic oscillator, $\tfrac12 k_k\langle Q_k^2\rangle = \tfrac12 k_BT$). Since the vibronically borrowed intensity is proportional to $\langle Q_k^2\rangle$,

That is, the stiffer-bonded cyanide complex is predicted to have the weaker vibronically-allowed d-d band, by a factor of about 2.5, because a stiffer bond has a smaller thermal vibrational amplitude and therefore departs less far from the centrosymmetric equilibrium geometry. (In practice this specific comparison is complicated by $[\text{Cr(CN)}_6]^{3-}$ also having low-lying, allowed charge-transfer bands and a different set of vibronically-active mode frequencies, so this idealized ratio should be read as illustrating the mechanism, not as a literal spectroscopic prediction to be checked against tabulated $\varepsilon$ values.) As a limiting-case check: if the two complexes had identical force constants ($k_k$ ratio $=1$), the predicted intensity ratio correctly reduces to 1 — the vibronic mechanism contributes equally whenever the vibrational amplitudes (hence the degree of instantaneous symmetry-breaking) are equal, exactly as the derivation requires.

The Laporte rule is named for Otto Laporte, a German-American physicist who in 1924 analyzed the iron atomic emission spectrum and noticed that observed transitions always connected energy levels of opposite parity — a purely spectroscopic regularity he formalized before the group-theoretic underpinning (established shortly afterward via Wigner and others' application of quantum mechanics to selection rules) explained why it had to be true. Laporte's original context was atomic spectroscopy, not coordination chemistry; the extension of the rule to molecular and complex-ion d-d spectra, and the recognition that vibronic coupling could relax it, came later as ligand-field theory matured through the mid-20th century, with key contributions from spectroscopists systematically cataloguing transition-metal complex intensities and correlating them with point-group symmetry (a research program closely associated with the broader development of ligand-field and crystal-field theory by Bethe, Van Vleck, Griffith, and others building the theoretical scaffold onto which Laporte's empirical rule was mapped).