Crystal field theory fundamentals

Intuition [Beginner]

Transition metals form coloured, magnetic compounds because their d-electrons are split into different energy levels by the surrounding ligands. Crystal field theory explains how this splitting works.

The five d-orbitals on a free metal ion all have the same energy — they are degenerate. When six ligands surround the metal in an octahedral arrangement (along the x, y, and z axes), two orbitals ($d_{z^2}$ and $d_{x^2-y^2}$) point directly at the ligands and feel strong electron-electron repulsion. The other three ($d_{xy}$, $d_{xz}$, $d_{yz}$) point between the ligands and feel less repulsion. The result: the five d-orbitals split into a higher pair ($e_g$) and a lower trio ($t_{2g}$), separated by the crystal field splitting energy ${\Delta }_o$.

The magnitude of ${\Delta }_o$ depends on the ligand. The spectrochemical series ranks ligands from weak-field (small ${\Delta }_o$, like I${}^{-}$ and Br${}^{-}$) to strong-field (large ${\Delta }_o$, like CN${}^{-}$ and CO). Weak-field ligands tend to give high-spin complexes (many unpaired electrons); strong-field ligands give low-spin complexes (fewer unpaired electrons).

Crystal field stabilisation energy (CFSE) measures how much the d-electron configuration lowers the total energy relative to the spherically symmetric free ion. Each electron in the $t_{2g}$ set stabilises by $-0.4{\Delta }_o$; each in the $e_g$ set destabilises by $+0.6{\Delta }_o$. The total CFSE is the sum of all contributions.

Visual [Beginner]

Worked example [Beginner]

For [Fe(H${}_2$O)${}_6$]${}^{2+}$ (high-spin d${}^6$) vs [Fe(CN)${}_6$]${}^{4-}$ (low-spin d${}^6$), calculate CFSE and predict magnetic moments.

[Fe(H${}_2$O)${}_6$]${}^{2+}$. Water is a weak-field ligand, ${\Delta }_o<P$ (pairing energy). High-spin filling: $t_{2g}^4{e}_g^2$. Four electrons in $t_{2g}$ (three unpaired, one paired), two in $e_g$ (one each).

CFSE = $4\times (-0.4{\Delta }_o)+2\times (+0.6{\Delta }_o)=-1.6{\Delta }_o+1.2{\Delta }_o=-0.4{\Delta }_o$.

Unpaired electrons: $t_{2g}^4=\uparrow \downarrow ,\uparrow ,\uparrow$ (one pair, two singles) and $e_g^2=\uparrow ,\uparrow$. Total unpaired: 2 + 2 = 4.

Magnetic moment (spin-only): $\mu =\sqrt{4(4+2)}{\mu }_B=\sqrt{24}{\mu }_B\approx 4.9{\mu }_B$.

[Fe(CN)${}_6$]${}^{4-}$. CN${}^{-}$ is a strong-field ligand, ${\Delta }_o>P$. Low-spin filling: $t_{2g}^6{e}_g^0$.

CFSE = $6\times (-0.4{\Delta }_o)=-2.4{\Delta }_o$ (minus $2P$ for two extra pairings).

Unpaired electrons: 0 (all paired). Diamagnetic.

The same metal ion gives opposite magnetic behaviour depending on the ligand — one of the cleanest confirmations of crystal field theory.

Additional worked example: d${}^4$ high-spin vs low-spin.

For d${}^4$ in octahedral geometry, the high-spin configuration is $t_{2g}^3{e}_g^1$ and the low-spin is $t_{2g}^4{e}_g^0$.

High-spin: CFSE = $3(-0.4{\Delta }_o)+1(+0.6{\Delta }_o)=-0.6{\Delta }_o$. Unpaired electrons: 4.

Low-spin: CFSE = $4(-0.4{\Delta }_o)=-1.6{\Delta }_o$, minus one pairing energy $P$ (the fourth electron pairs in $t_{2g}$ instead of going into $e_g$). Unpaired electrons: 2.

Low-spin is favoured when $-1.6{\Delta }_o-P<-0.6{\Delta }_o$, i.e., ${\Delta }_o>P$. Cr${}^{2+}$ (d${}^4$) is high-spin with water ligands (${\Delta }_o\approx 14,000$ cm${}^{-1}$, $P\approx 20,000$ cm${}^{-1}$) but low-spin with CN${}^{-}$ (${\Delta }_o\approx 26,000$ cm${}^{-1}$). This prediction is confirmed experimentally: [Cr(H${}_2$O)${}_6$]${}^{2+}$ has four unpaired electrons ($\mu \approx 4.9{\mu }_B$) and is paramagnetic, while [Cr(CN)${}_6$]${}^{4-}$ has two unpaired electrons ($\mu \approx 2.8{\mu }_B$).

Check your understanding [Beginner]

Formal definition [Intermediate+]

A transition-metal ion in the gas phase has five degenerate d-orbitals. When the ion is placed at the centre of an arrangement of ligands, each ligand (modelled as a point charge or point dipole) produces an electrostatic potential at the metal that perturbs the d-orbital energies. Crystal field theory is the model that treats ligands as classical point charges and computes the resulting splitting pattern from the symmetry of the arrangement alone.

Octahedral splitting. For six identical ligands at the vertices of a regular octahedron (along $\pm x$, $\pm y$, $\pm z$), the five d-orbitals decompose under the O${}_h$ point group into two sets. The pair $d_{z^2}$ and $d_{x^2-y^2}$ (the $e_g$ set, dimension 2) have lobes pointing directly at the ligands and are destabilised. The trio $d_{xy}$, $d_{xz}$, $d_{yz}$ (the $t_{2g}$ set, dimension 3) have lobes directed between the ligands and are stabilised relative to the free-ion average. The energy gap between the two sets is the crystal field splitting parameter ${\Delta }_o$ (also written $10Dq$). The detailed group-theoretic derivation of this decomposition from the O${}_h$ character table is given in unit 16.03.02.

The barycentre rule constrains the energies: the weighted average of the split levels must equal the free-ion energy, which is taken as the zero of energy. With two $e_g$ orbitals at energy $E(e_g)$ and three $t_{2g}$ orbitals at energy $E(t_{2g})$:

$$2E(e_g)+3E(t_{2g})=0,E(e_g)-E(t_{2g})={\Delta }_o.$$

Solving simultaneously gives $E(e_g)=+\frac{3}{5}{\Delta }_o$ and $E(t_{2g})=-\frac{2}{5}{\Delta }_o$. Each $t_{2g}$ electron lowers the energy by $0.4{\Delta }_o$ and each $e_g$ electron raises it by $0.6{\Delta }_o$.

CFSE formula. For an octahedral complex with $n_{t_{2g}}$ electrons in the $t_{2g}$ set and $n_{e_g}$ electrons in the $e_g$ set, the crystal field stabilisation energy is:

$$\text{CFSE}=n_{t_{2g}}\times (-0.4{\Delta }_o)+n_{e_g}\times (+0.6{\Delta }_o)-n_P\times P$$

where $P$ is the mean electron-pairing energy and $n_P$ is the number of electron pairs beyond the minimum required by Hund's rule. When CFSE is stated without pairing correction, the $-n_{P}P$ term is omitted and the quantity is described as "CFSE before pairing penalty."

Tetrahedral splitting. Four identical ligands at the vertices of a regular tetrahedron produce a splitting pattern that is the inverse of the octahedral case: a lower-energy $e$ pair and a higher-energy $t_2$ trio. The tetrahedral splitting ${\Delta }_t$ is related to the octahedral splitting for the same metal-ligand pair by:

$${\Delta }_t=\frac{4}{9}{\Delta }_o.$$

The reduction factor arises from two geometric effects. First, a tetrahedron has four ligands rather than six, reducing the total electrostatic perturbation by a factor of $\frac{2}{3}$. Second, no ligand sits directly on a Cartesian axis; the ligand directions approach the d-orbital lobes at an angle, reducing the overlap. The combined effect gives the $\frac{4}{9}$ ratio. Because ${\Delta }_t$ is always less than half ${\Delta }_o$, tetrahedral complexes of first-row transition metals are essentially always high-spin.

High-spin vs low-spin. For configurations d${}^4$ through d${}^7$ in octahedral geometry, two electron configurations are possible. The high-spin configuration places electrons in the $e_g$ set before pairing in $t_{2g}$, maximising the number of unpaired electrons. The low-spin configuration pairs electrons in $t_{2g}$ before occupying $e_g$, minimising the number of unpaired electrons. The competition is between the pairing energy cost $P$ and the crystal field splitting ${\Delta }_o$:

• High-spin is favoured when ${\Delta }_o<P$ (the energy cost of promoting to $e_g$ is less than the cost of pairing).
• Low-spin is favoured when ${\Delta }_o>P$ (pairing in $t_{2g}$ is energetically cheaper than occupying $e_g$).

Spin-only magnetic moment. For $n$ unpaired electrons, the spin-only magnetic moment is:

$${\mu }_{\text{s.o.}}=\sqrt{n(n+2)}{\mu }_B$$

where ${\mu }_B$ is the Bohr magneton ($9.274\times 10^{-24}$ J T${}^{-1}$). This formula neglects orbital angular momentum contributions, which are partially or fully quenched in many octahedral complexes by the crystal field. Deviations from the spin-only value provide evidence for orbital contributions and are treated at Master tier.

Spectrochemical series (selected):

$${\text{I}}^{-}<{\text{Br}}^{-}<{\text{Cl}}^{-}<{\text{F}}^{-}<{\text{OH}}^{-}<{\text{H}}_2\text{O}<{\text{NH}}_3<\text{en}<{\text{NO}}_2^{-}<{\text{CN}}^{-}<\text{CO}$$

Factors affecting ${\Delta }_o$:

1. Nature of the ligand. The spectrochemical series orders ligands by the ${\Delta }_o$ they produce. Pi-acceptor ligands (CO, CN${}^{-}$) give the largest ${\Delta }_o$; pi-donor ligands (I${}^{-}$, Br${}^{-}$) give the smallest.

2. Charge on the metal. Higher charge draws ligands closer, increasing the metal-ligand electrostatic interaction and raising ${\Delta }_o$. Compare: [Fe(H${}_2$O)${}_6$]${}^{2+}$ (${\Delta }_o\approx 10,400$ cm${}^{-1}$) vs [Fe(H${}_2$O)${}_6$]${}^{3+}$ (${\Delta }_o\approx 14,000$ cm${}^{-1}$).

3. Principal quantum number. ${\Delta }_o$ increases down a group because the 4d and 5d orbitals extend further from the nucleus and overlap more with the ligands. This is why second- and third-row transition metals are almost always low-spin: even weak-field ligands produce a ${\Delta }_o$ larger than the pairing energy.

Counterexamples to common slips

• CFSE is not the only contribution to complex stability. The Irving-Williams series (${\text{Mn}}^{2+}<{\text{Fe}}^{2+}<{\text{Co}}^{2+}<{\text{Ni}}^{2+}<{\text{Cu}}^{2+}>{\text{Zn}}^{2+}$) reflects CFSE plus metal-specific effects like Jahn-Teller distortion and covalent contributions beyond the point-charge model.

• The point-charge model underestimates ${\Delta }_o$ by roughly an order of magnitude. The purely electrostatic calculation gives ${\Delta }_o$ values far too small to match experiment. Covalent bonding (ligand field theory) is needed for quantitative agreement. CFT correctly predicts splitting patterns and qualitative trends; it does not predict magnitudes.

• High-spin and low-spin are not terms for all d-electron counts. Configurations d${}^0$–d${}^3$ and d${}^8$–d${}^{10}$ have only one possible arrangement in octahedral geometry. The spin-state choice exists only for d${}^4$ through d${}^7$.

• "Strong-field" and "weak-field" describe the ligand, not the complex. A weak-field ligand can still produce a low-spin complex if the metal is from the second or third transition series (large radial extension of d-orbitals increases ${\Delta }_o$). Conversely, a strong-field ligand on a first-row metal in a low oxidation state may still give a high-spin complex if the metal-ligand distance is large.

Key theorem with proof [Intermediate+]

Theorem (Barycentre conservation). In a crystal field of any symmetry, the weighted sum of d-orbital energies after splitting equals the barycentre (mean energy) of the degenerate free-ion d-orbitals. For the octahedral case, this constraint together with the definition of ${\Delta }_o$ uniquely determines the energies: $E(e_g)=+\frac{3}{5}{\Delta }_o$ and $E(t_{2g})=-\frac{2}{5}{\Delta }_o$.

Proof. The crystal field acts as a perturbation on the degenerate free-ion d-orbitals. Because the crystal field Hamiltonian ${\hat{V}}_{\text{CF}}$ is traceless — it redistributes energy among the five d-orbitals but does not add net energy — the sum of the five diagonal matrix elements vanishes:

$$\sum _{i=1}^5⟨d_i|{\hat{V}}_{\text{CF}}|d_i⟩=0.$$

Setting the free-ion barycentre as the zero of energy, the perturbed orbital energies are ${\epsilon }_i=⟨d_i|{\hat{V}}_{\text{CF}}|d_i⟩$, so the trace condition gives:

$$\sum _{i=1}^5{\epsilon }_i=0.$$

In O${}_h$ symmetry, the five d-orbitals split into the doubly-degenerate $e_g$ set at energy $E(e_g)$ and the triply-degenerate $t_{2g}$ set at energy $E(t_{2g})$. Substituting:

$$2E(e_g)+3E(t_{2g})=0.$$

The crystal field splitting parameter is defined as ${\Delta }_o=E(e_g)-E(t_{2g})$, so $E(e_g)=E(t_{2g})+{\Delta }_o$. Substituting into the barycentre condition:

$$2(E(t_{2g})+{\Delta }_o)+3E(t_{2g})=0⟹5E(t_{2g})+2{\Delta }_o=0,$$

which gives $E(t_{2g})=-\frac{2}{5}{\Delta }_o$ and correspondingly $E(e_g)=+\frac{3}{5}{\Delta }_o$. The same trace argument applies to any point group: the weighted sum of orbital energies after splitting must vanish. For T${}_d$ symmetry, the result is ${\Delta }_t=\frac{4}{9}{\Delta }_o$ with the splitting inverted (the $e$ set is now lower). $\square$

Bridge. The barycentre theorem is the foundational reason that the $-0.4/+0.6$ coefficients in the CFSE formula are fixed by symmetry alone, independent of the metal or ligand identity. This is exactly the algebraic constraint that makes CFSE a universal bookkeeping device across the entire d-block. The theorem builds toward 16.03.02 pending, where the group-theoretic machinery of the O${}_h$ character table provides a rigorous derivation of which orbitals fall into $e_g$ and $t_{2g}$, and appears again in the non-octahedral geometries treated at Master tier, where the same trace condition applied to D${}_{4h}$ and T${}_d$ produces the splitting patterns for square-planar and tetrahedral complexes. The central insight is that the trace of the crystal field perturbation vanishes by construction; the bridge is between this algebraic fact and the experimentally measurable splitting parameter ${\Delta }_o$.

Exercises [Intermediate+]

Ligand field theory and the spectrochemical series [Master]

Crystal field theory treats ligands as point charges. It correctly predicts splitting patterns and qualitative trends but underestimates ${\Delta }_o$ by an order of magnitude. The reason is that covalent metal-ligand bonding — orbital overlap and electron sharing — contributes substantially to the splitting. Ligand field theory (LFT) extends CFT by including covalent bonding through the molecular orbital framework while retaining the symmetry analysis of CFT.

Three classes of ligand behaviour explain the spectrochemical series:

Sigma-donor ligands (NH${}_3$, H${}_2$O) interact with the metal through sigma bonding only. The ligand donates electron density into the metal $e_g$ orbitals (which point along the metal-ligand axis), raising their energy. The $t_{2g}$ orbitals are non-bonding. The result is a moderate ${\Delta }_o$, with NH${}_3$ producing a larger ${\Delta }_o$ than H${}_2$O because ammonia is a stronger sigma donor.

Pi-donor ligands (halides, OH${}^{-}$) have filled p-orbitals that overlap with the metal $t_{2g}$ set. This interaction donates additional electron density into the $t_{2g}$ orbitals, raising their energy and reducing ${\Delta }_o$. The more effective the pi donation, the smaller the splitting. This explains why I${}^{-}$ (large, polarisable, strong pi donor) produces a smaller ${\Delta }_o$ than F${}^{-}$ (small, hard, weak pi donor). The halide ordering in the spectrochemical series (I${}^{-}<$ Br${}^{-}<$ Cl${}^{-}<$ F${}^{-}$) tracks pi-donor ability.

Pi-acceptor ligands (CO, CN${}^{-}$, NO${}_2^{-}$) have empty pi-star (${\pi }^{\ast }$) orbitals that accept electron density from the filled metal $t_{2g}$ set. This back-bonding interaction lowers the energy of the $t_{2g}$ orbitals (they acquire bonding character with respect to the metal-ligand pi interaction) while the $e_g$ set is unaffected. The result is a large ${\Delta }_o$. Carbon monoxide is the archetypal pi-acceptor ligand; the strength of its back-bonding interaction makes CO complexes among the most strongly split.

The spectrochemical series is thus explained by the progression from pi-donor (small ${\Delta }_o$, raises $t_{2g}$) through sigma-only (moderate ${\Delta }_o$, $t_{2g}$ non-bonding) to pi-acceptor (large ${\Delta }_o$, lowers $t_{2g}$). This progression is developed in quantitative detail through molecular orbital calculations in unit 16.03.02.

The quantitative data confirm the progression. Representative ${\Delta }_o$ values (in cm${}^{-1}$) for first-row transition-metal complexes:

Complex	d${}^n$	${\Delta }_o$ (cm${}^{-1}$)
[V(H${}_2$O)${}_6$]${}^{3+}$	d${}^2$	18,500
[Cr(H${}_2$O)${}_6$]${}^{3+}$	d${}^3$	17,400
[Cr(NH${}_3$)${}_6$]${}^{3+}$	d${}^3$	21,500
[Cr(CN)${}_6$]${}^{3-}$	d${}^3$	26,600
[Co(H${}_2$O)${}_6$]${}^{3+}$	d${}^6$	18,600
[Co(NH${}_3$)${}_6$]${}^{3+}$	d${}^6$	22,900
[Co(CN)${}_6$]${}^{3-}$	d${}^6$	34,800
[Fe(H${}_2$O)${}_6$]${}^{2+}$	d${}^6$	10,400
[Fe(CN)${}_6$]${}^{4-}$	d${}^6$	33,800
[Ni(H${}_2$O)${}_6$]${}^{2+}$	d${}^8$	8,500
[Ni(NH${}_3$)${}_6$]${}^{2+}$	d${}^8$	10,800

Three trends are immediately visible. For the same metal and oxidation state, ${\Delta }_o$ increases across the spectrochemical series (compare Cr${}^{3+}$ with H${}_2$O, NH${}_3$, CN${}^{-}$). For the same ligand, ${\Delta }_o$ increases with oxidation state (compare Fe${}^{2+}$ vs Fe${}^{3+}$ with the same ligands). For the same ligand and oxidation state, ${\Delta }_o$ increases going down a group (4d > 3d, 5d > 4d), which is why second- and third-row transition metals are essentially always low-spin.

Tanabe-Sugano diagrams and quantitative d-d spectroscopy [Master]

A Tanabe-Sugano diagram plots the energy of each electronic excited state (as $E/B$, where $B$ is the Racah parameter measuring interelectronic repulsion) against the crystal field strength ${\Delta }_o/B$ for a given d${}^n$ configuration. These diagrams are the primary tool for interpreting d-d absorption spectra quantitatively. They were introduced by Tanabe and Sugano in 1954 ^{[Tanabe Sugano 1954]} and remain the standard reference for extracting crystal field parameters from electronic spectra.

Each line on a Tanabe-Sugano diagram represents a term symbol (${}^4{T}_{2g}$, ${}^4{T}_{1g}$, etc.) derived from the free-ion Russell-Saunders term after splitting in the O${}_h$ field. The ground state is plotted as the horizontal axis ($E=0$). For configurations with a spin crossover (d${}^4$ through d${}^7$), the diagram has a discontinuity at the crossover ${\Delta }_o/B$ value where the ground state changes from high-spin to low-spin; lines change slope at this point.

The observed absorption bands correspond to vertical transitions from the ground state to excited states at the fixed ${\Delta }_o/B$ of the complex. The positions of these bands encode both ${\Delta }_o$ and the Racah parameter $B$. The procedure for extracting these parameters is best illustrated with a concrete example.

The d${}^3$ case (Cr${}^{3+}$). The free-ion ground term for d${}^3$ is ${}^{4}F$, which splits in O${}_h$ into ${}^4{A}_{2g}$ (ground), ${}^4{T}_{2g}$, and ${}^4{T}_{1g}(F)$. The first excited free-ion term ${}^{4}P$ gives ${}^4{T}_{1g}(P)$. The three spin-allowed transitions are:

$${}^4{A}_{2g}\to {}^4{T}_{2g},{}^4{A}_{2g}\to {}^4{T}_{1g}(F),{}^4{A}_{2g}\to {}^4{T}_{1g}(P).$$

The lowest-energy transition (${}^4{A}_{2g}\to {}^4{T}_{2g}$) directly gives ${\Delta }_o$ because in the Tanabe-Sugano diagram this transition energy equals ${\Delta }_o$ at all field strengths. The two higher transitions depend on both ${\Delta }_o$ and $B$; their spacing provides the Racah parameter.

For [Cr(H${}_2$O)${}_6$]${}^{3+}$, the three observed bands are at approximately 17,400 cm${}^{-1}$, 24,600 cm${}^{-1}$, and 38,000 cm${}^{-1}$. The first band gives ${\Delta }_o=17,400$ cm${}^{-1}$ directly. The position of the second band (${}^4{T}_{1g}(F)$) relative to ${\Delta }_o$ gives $B\approx 918$ cm${}^{-1}$, compared to the free-ion value $B_0\approx 1030$ cm${}^{-1}$. The reduction $B/B_0\approx 0.89$ is the nephelauxetic effect — interelectronic repulsion is reduced in the complex relative to the free ion, a consequence of covalent electron delocalisation discussed in the next section.

The d${}^6$ case and spin crossover. The Tanabe-Sugano diagram for d${}^6$ illustrates the spin-crossover phenomenon directly. At low ${\Delta }_o/B$ the ground state is the high-spin ${}^5{T}_{2g}$ term ($t_{2g}^4{e}_g^2$). As ${\Delta }_o/B$ increases, the low-spin ${}^1{A}_{1g}$ term ($t_{2g}^6{e}_g^0$) drops below it, and at a critical value (${\Delta }_o/B\approx 20$ for typical $C/B$ ratios) the ground state switches. The crossover point marks the boundary between high-spin and low-spin behaviour for d${}^6$ complexes. Real complexes near this crossover — such as [Fe(phen)${}_2$(NCS)${}_2$] — can be thermally switched between spin states, a phenomenon exploited in spin-crossover materials for molecular electronics and sensors.

The nephelauxetic effect and covalent delocalisation [Master]

The Racah parameter $B$ measured in a complex is always smaller than the free-ion value $B_0$. This reduction, quantified by $\beta =B/B_0<1$, is called the nephelauxetic effect (from Greek "nephelauxesis," cloud-expanding). Covalent bonding delocalises the d-electrons over a larger volume encompassing both the metal and the ligands, reducing the average interelectronic repulsion. The more covalent the metal-ligand bonding, the greater the delocalisation and the smaller $B$ becomes.

The nephelauxetic series ranks ligands by the magnitude of the $B$-reduction they produce:

$${\text{F}}^{-}>{\text{H}}_2\text{O}>{\text{NH}}_3>\text{en}>{\text{Cl}}^{-}>{\text{CN}}^{-}>{\text{Br}}^{-}>{\text{I}}^{-}$$

Ligands on the left produce $\beta$ close to 1 (nearly ionic bonding, minimal delocalisation); ligands on the right produce $\beta$ as low as 0.6–0.7 (substantial covalent character). The ordering correlates with, but is not identical to, the spectrochemical series: both track covalent interaction strength, but the nephelauxetic series measures the total covalent delocalisation while the spectrochemical series specifically measures the $t_{2g}$-$e_g$ splitting interaction.

The nephelauxetic effect also depends on the metal. For a given ligand, $\beta$ decreases (more covalent delocalisation) going down a group: 5d < 4d < 3d. This is the same trend that makes second- and third-row transition metals prefer low-spin configurations — the larger, more diffuse d-orbitals overlap more effectively with the ligand orbitals, producing both larger ${\Delta }_o$ and smaller $B$.

The physical origin of the nephelauxetic effect can be understood through the molecular orbital picture. In a purely ionic complex, the d-electrons are confined to the metal and experience the full interelectronic repulsion of the free ion. As covalent bonding develops, the d-electrons acquire some ligand-orbital character — the d-electron cloud expands onto the ligand framework, increasing the average interelectron distance and reducing the Coulomb repulsion integrals that contribute to $B$. The Racah parameter $B$ is proportional to the Slater-Condon integrals $F^2$ and $F^4$, which decrease as the radial extent of the d-electron distribution increases.

In practice, the nephelauxetic effect is measured by extracting $B$ from the electronic spectrum (via the Tanabe-Sugano analysis described above) and comparing it to the free-ion value. For [Ni(H${}_2$O)${}_6$]${}^{2+}$, $B\approx 940$ cm${}^{-1}$ vs $B_0\approx 1084$ cm${}^{-1}$, giving $\beta =0.87$. For [Ni(NH${}_3$)${}_6$]${}^{2+}$, $B\approx 892$ cm${}^{-1}$, giving $\beta =0.82$. For [Ni(CN)${}_4$]${}^{2-}$, $B\approx 750$ cm${}^{-1}$, giving $\beta =0.69$. The progression tracks the increasing covalent character of the metal-ligand bond.

Selection rules and spectroscopic intensities [Master]

The d-d transitions that give transition-metal complexes their colours are subject to two selection rules that severely limit their intensity. Understanding these rules and the mechanisms by which they are relaxed is essential for interpreting electronic spectra quantitatively.

The Laporte (parity) selection rule forbids transitions between states of the same parity. In a centrosymmetric point group like O${}_h$, all d-orbitals have $g$ (gerade, even) parity, so any $g\to g$ transition is Laporte-forbidden. The d-d transitions are all $g\to g$ and are therefore formally forbidden. They gain intensity only through mechanisms that break the centre of symmetry:

1. Vibronic coupling. Molecular vibrations instantaneously distort the complex away from perfect O${}_h$ symmetry. During an asymmetric vibration (one that removes the inversion centre), the $g$/$u$ distinction is temporarily lost, and the transition acquires a small allowed character. The intensity of the d-d band is borrowed from an allowed (charge-transfer or ligand-centred) transition through this vibronic coupling mechanism.

2. Static distortion. Complexes with permanent Jahn-Teller distortions (e.g., Cu${}^{2+}$, d${}^9$) lack a centre of symmetry and have intrinsically more intense d-d bands. The distortion lowers the symmetry from O${}_h$ to D${}_{4h}$ (or lower), partially lifting the Laporte restriction.

The spin selection rule ($\Delta S=0$) forbids transitions between states of different spin multiplicity. Spin-allowed transitions (same multiplicity) are orders of magnitude more intense than spin-forbidden ones. The visible colour of most transition-metal complexes arises from spin-allowed, Laporte-forbidden d-d bands. The typical molar absorptivity ranges are:

• Spin-allowed, Laporte-forbidden (d-d): $ϵ\approx 10$–100 M${}^{-1}$cm${}^{-1}$
• Spin-forbidden (d-d): $ϵ<1$ M${}^{-1}$cm${}^{-1}$
• Charge-transfer (Laporte-allowed): $ϵ\approx 10^3$–$10^5$ M${}^{-1}$cm${}^{-1}$

The intensity ordering reflects the hierarchy of selection-rule relaxation. Charge-transfer bands are fully allowed and are the most intense features in the spectrum, but they typically lie in the ultraviolet and do not contribute to the visible colour of the complex. The visible absorption — and hence the perceived colour — comes from the weak d-d bands.

Charge-transfer transitions involve electron movement between the metal and the ligand. In ligand-to-metal charge transfer (LMCT), an electron from a ligand-based orbital is promoted to a metal-based orbital. In metal-to-ligand charge transfer (MLCT), an electron from a metal d-orbital is promoted to a ligand-based ${\pi }^{\ast }$ orbital. Because these transitions involve a change in orbital parity ($g\to u$ or $u\to g$), they are Laporte-allowed and intensely coloured. The intense purple of [MnO${}_4$]${}^{-}$ (permanganate, LMCT) and the deep red of [Fe(phen)${}_3$]${}^{2+}$ (MLCT) are charge-transfer colours, not d-d colours.

The interplay between d-d and charge-transfer bands makes electronic spectroscopy a powerful tool for characterising transition-metal complexes. The position of d-d bands gives ${\Delta }_o$ and $B$; the presence of charge-transfer bands confirms the metal-ligand covalent interaction; and the intensity pattern constrains the symmetry of the complex.

Non-octahedral geometries and the Jahn-Teller theorem [Master]

The barycentre conservation theorem applies to any point-group symmetry. The splitting pattern depends on the symmetry of the ligand arrangement, and the same trace-zero constraint produces different energy-level diagrams for different geometries.

Tetrahedral geometry (T${}_d$). Four ligands at the vertices of a regular tetrahedron produce a splitting pattern that is the inverse of the octahedral case: the $e$ pair ($d_{z^2}$, $d_{x^2-y^2}$) is lower and the $t_2$ trio ($d_{xy}$, $d_{xz}$, $d_{yz}$) is higher. The inversion occurs because in T${}_d$ symmetry the ligands approach along the body diagonals of the cube, which are closer to the $t_2$ orbital lobes (which point toward the cube edges) than to the $e$ orbital lobes (which point along the cube axes). The splitting is ${\Delta }_t=\frac{4}{9}{\Delta }_o$ for the same metal-ligand pair. Tetrahedral complexes are named without the $g$ subscript ($e$ and $t_2$ rather than $e_g$ and $t_{2g}$) because the tetrahedron lacks a centre of inversion.

Square-planar geometry (D${}_{4h}$). The square-planar splitting is derived from the octahedral diagram by removing the two axial ligands. The effect is dramatic: the $d_{z^2}$ orbital drops significantly because it no longer experiences axial repulsion, while $d_{x^2-y^2}$ rises to become the highest orbital because all four remaining ligands are in the xy plane and interact maximally with this orbital. The ordering is:

$$d_{x^2-y^2}\gg d_{xy}>d_{z^2}>d_{xz},d_{yz}$$

The very high energy of $d_{x^2-y^2}$ explains why d${}^8$ metals (Pt${}^{2+}$, Pd${}^{2+}$, Au${}^{3+}$) strongly prefer square-planar geometry: the d${}^8$ configuration leaves the destabilising $d_{x^2-y^2}$ orbital empty while fully occupying the four lower orbitals. The splitting between $d_{x^2-y^2}$ and $d_{xy}$ is approximately $1.3{\Delta }_o$ for a typical Pt(II) complex — larger than the original octahedral splitting because the in-plane ligands are drawn closer when the axial ligands are removed. The preference for square-planar over tetrahedral geometry in d${}^8$ metals increases down a group: Ni${}^{2+}$ forms both square-planar and tetrahedral complexes depending on the ligand, while Pt${}^{2+}$ is almost exclusively square-planar.

Tetragonal distortion and the Jahn-Teller theorem. The Jahn-Teller theorem states that any non-linear molecular system in an orbitally degenerate electronic state will distort to lift that degeneracy. The theorem does not predict the direction or magnitude of the distortion, only that it must occur. For octahedral transition-metal complexes, the distortions are most significant when the degeneracy occurs in the $e_g$ set (strong Jahn-Teller effect) because the $e_g$ orbitals are sigma-antibonding and interact directly with the ligands. Degeneracy in the $t_{2g}$ set produces only a weak Jahn-Teller effect because $t_{2g}$ orbitals are pi-nonbonding or weakly pi-bonding.

The configurations subject to strong Jahn-Teller distortion in octahedral geometry are d${}^4$ high-spin ($t_{2g}^3{e}_g^1$, uneven $e_g$ occupation) and d${}^9$ ($t_{2g}^6{e}_g^3$, uneven $e_g$ occupation). For d${}^9$ (Cu${}^{2+}$), the dominant distortion is axial elongation: the two axial bonds lengthen while the four equatorial bonds shorten. This distortion splits the $e_g$ pair into $b_{1g}$ ($d_{x^2-y^2}$, now highest) and $a_{1g}$ ($d_{z^2}$, lowered), and splits the $t_{2g}$ trio into $b_{2g}$ ($d_{xy}$) and $e_g$ ($d_{xz}$, $d_{yz}$). The Jahn-Teller stabilisation energy — the energy gained by the distortion — is typically 1–2 eV for Cu${}^{2+}$ complexes, which is large enough to be a dominant factor in the stereochemistry of copper(II) chemistry. Virtually every known Cu${}^{2+}$ complex exhibits a measurable tetragonal distortion.

Five-coordinate geometries. Square-based pyramidal and trigonal bipyramidal geometries (coordination number 5) have intermediate splitting patterns between octahedral and tetrahedral. These geometries are encountered in five-coordinate intermediates during ligand substitution reactions at octahedral metal centres and in certain catalytically active species. The splitting pattern depends sensitively on the geometry: a trigonal bipyramidal complex splits the five d-orbitals into three sets ($e^{\prime }+e^{\prime \prime }+a_1^{\prime }$ in D${}_{3h}$), while a square-based pyramidal complex splits them into four or five levels depending on the degree of distortion from C${}_{4v}$.

Geometry preferences across the d-block. The preferred coordination geometry of a transition-metal ion correlates strongly with its d-electron count and the resulting CFSE. Ions with zero CFSE (d${}^0$, d${}^5$ high-spin, d${}^{10}$) have no crystal-field preference and adopt the geometry determined by ligand-ligand repulsions alone (typically octahedral for six-coordinate species, tetrahedral for four-coordinate). Ions with large octahedral CFSE (d${}^3$, d${}^6$ low-spin, d${}^8$) strongly prefer octahedral or square-planar geometries. Ions with large Jahn-Teller distortions (d${}^4$ high-spin, d${}^9$) adopt elongated octahedral geometries. This correlation between d-electron count and preferred geometry is one of the most successful predictive applications of crystal field theory: knowing the d${}^n$ configuration and the position of the ligand in the spectrochemical series, one can predict the most likely coordination geometry with high reliability.

The geometry-CFSE correlation also explains the kinetic inertness of certain complexes. d${}^3$ (Cr${}^{3+}$) and low-spin d${}^6$ (Co${}^{3+}$) complexes are kinetically inert — they undergo ligand substitution slowly — because any transition state leading to substitution (typically five-coordinate or seven-coordinate) has a substantially different splitting pattern with less CFSE. The loss of CFSE in the transition state creates an activation energy barrier that slows the reaction. Conversely, d${}^0$, d${}^5$ high-spin, and d${}^{10}$ complexes are kinetically labile because the transition state incurs no CFSE penalty. This kinetic consequence of CFSE underlies the widespread use of Cr${}^{3+}$ and Co${}^{3+}$ complexes in coordination chemistry: they are stable enough to isolate and characterise, yet kinetically accessible through controlled conditions.

CFSE and thermodynamic properties [Master]

The double-humped CFSE curve (maxima at d${}^3$ and d${}^8$, minima at d${}^0$, d${}^5$, d${}^{10}$) has direct thermodynamic consequences that are among the most compelling experimental evidence for crystal field theory. Three thermodynamic quantities — hydration enthalpies, lattice energies, and complex-formation constants — all show deviations from the smooth trends expected from nuclear charge alone, and the deviations match the CFSE predictions quantitatively.

Hydration enthalpies. The enthalpy of hydration $\Delta H_{\text{hyd}}$ for the reaction M${}^{2+}$(g) $\to$ M${}^{2+}$(aq) becomes more exothermic across the first-row transition metals from Ca${}^{2+}$ to Zn${}^{2+}$ as the increasing nuclear charge draws water ligands closer. If CFSE played no role, the trend would be approximately linear. In reality, the experimental values show the characteristic double-humped deviation: the hydration enthalpies for V${}^{2+}$ (d${}^3$, CFSE = $-1.2{\Delta }_o$) and Ni${}^{2+}$ (d${}^8$, CFSE = $-1.2{\Delta }_o$) are more exothermic than the linear baseline by approximately 40–60 kJ/mol, while Mn${}^{2+}$ (d${}^5$, CFSE = 0) and Zn${}^{2+}$ (d${}^{10}$, CFSE = 0) fall on or near the baseline. The deviations are quantitatively consistent with the CFSE values calculated from the spectroscopically measured ${\Delta }_o$ for each hexaaqua complex, confirming that the CFSE contribution to the hydration enthalpy is real and measurable.

The Irving-Williams series. The stability constants for complexes of divalent first-row transition-metal ions with a given ligand follow the ordering:

$${\text{Mn}}^{2+}<{\text{Fe}}^{2+}<{\text{Co}}^{2+}<{\text{Ni}}^{2+}<{\text{Cu}}^{2+}>{\text{Zn}}^{2+}$$

This ordering holds for virtually all ligands and reflects the combined effects of increasing nuclear charge across the series and the CFSE variation. The stability peaks at Cu${}^{2+}$ (d${}^9$) rather than at Ni${}^{2+}$ (d${}^8$, maximum CFSE) because Cu${}^{2+}$ benefits from an additional Jahn-Teller stabilisation that further lowers its energy in most coordination geometries. The fact that Zn${}^{2+}$ (d${}^{10}$, zero CFSE) forms less stable complexes than Cu${}^{2+}$ despite having a higher nuclear charge is direct evidence that CFSE — not just electrostatics — governs transition-metal complex stability.

Spinel site preferences. Spinels are mixed-metal oxides with the general formula AB${}_2$O${}_4$. In a normal spinel, the A${}^{2+}$ ions occupy tetrahedral sites and the B${}^{3+}$ ions occupy octahedral sites. In an inverse spinel, half the B${}^{3+}$ ions swap with the A${}^{2+}$ ions, giving B(AB)O${}_4$. The preference for normal vs inverse structure is determined by the CFSE of the metal ions in tetrahedral vs octahedral sites. For example, NiFe${}_2$O${}_4$ is an inverse spinel because Ni${}^{2+}$ (d${}^8$) gains $-1.2{\Delta }_o$ in octahedral sites but only a smaller stabilisation in tetrahedral sites (where ${\Delta }_t=\frac{4}{9}{\Delta }_o$). The octahedral preference energy (CFSE${}_{\text{oct}}$ minus CFSE${}_{\text{tet}}$) predicts the site occupancy for a wide range of spinel compositions, providing another quantitative test of the crystal field model.

Lattice energies. The lattice energies of the dihalides MX${}_2$ across the first-row transition metals show the same double-humped deviation as the hydration enthalpies. The deviation is most pronounced for fluorides (small anion, more ionic bonding, CFSE is a larger fraction of the total lattice energy) and least pronounced for iodides (large anion, more covalent character, CFSE is a smaller fraction). The systematic variation of the deviation magnitude with halide identity is consistent with the expectation that CFSE depends on the metal-ligand distance: fluoride (short M-F distance) produces a larger ${\Delta }_o$ and therefore a larger CFSE contribution than iodide (long M-I distance).

Connections [Master]

• Symmetry and group theory in chemistry 16.02.01 (pending). The $E_g\oplus T_{2g}$ decomposition of the d-orbital basis under O${}_h$ is derived from the O${}_h$ character table using the projection-operator machinery of finite group representation theory. The barycentre theorem proved in this unit is the algebraic consequence of the trace-free crystal field perturbation; the symmetry analysis in 16.02.01 provides the rigorous group-theoretic foundation for why the d-orbitals split into exactly two sets in octahedral symmetry.

• Periodic trends quantified 16.01.01 pending (pending). The d-orbital electron configurations that determine CFSE — the filling order, the role of effective nuclear charge in controlling orbital energies, and the periodic trends in orbital radii — are a direct application of the periodic-trend framework. The increase in ${\Delta }_o$ down a group and across a period follows from the systematic trends in d-orbital radial extent and effective nuclear charge developed in 16.01.01.

• Coordination chemistry 16.04.01 (proposed). Coordination numbers, preferred geometries, isomerism, and thermodynamic stabilities of transition-metal complexes are partly governed by CFSE. The preference of d${}^8$ metals for square-planar geometry, the Jahn-Teller distortions in d${}^9$ Cu${}^{2+}$ complexes, and the Irving-Williams stability series all connect the crystal field splitting patterns developed here to the macroscopic chemistry of coordination compounds.

• Organometallic chemistry 16.05.01 (proposed). The 18-electron rule for organometallic compounds is the closed-shell criterion in the ligand-field MO framework. The strong-field ligands common in organometallic chemistry (CO, phosphines, cyclopentadienyl) produce large ${\Delta }_o$ values, and the electron counting that predicts stability is a direct extension of the CFSE bookkeeping developed here. The spectrochemical series rankings of these ligands explain why organometallic complexes are almost universally low-spin.

Historical & philosophical context [Master]

Hans Bethe introduced crystal field theory in 1929 ^{[Bethe 1929]} in his paper "Termaufspaltung in Kristallen" (Term Splitting in Crystals), published in Annalen der Physik 3, 133–208. Bethe's analysis was purely theoretical: he calculated how the energy levels of an ion with partially filled d- or f-shells split in the electrostatic field produced by the surrounding ions in a crystal lattice, using the symmetry of the lattice to determine the splitting pattern. The paper derived the $e_g$/$t_{2g}$ splitting for octahedral coordination and the corresponding patterns for tetrahedral, cubic, and other symmetries, establishing the complete framework that chemists still use.

John Hasbrouck Van Vleck extended Bethe's crystal field theory to magnetism in the early 1930s ^{[Van Vleck 1932]}, providing the theoretical explanation for the magnetic properties of transition-metal salts. Van Vleck showed that the quenching of orbital angular momentum in octahedral fields — the $t_{2g}$ electrons have no matrix elements coupling to the $e_g$ set through the orbital angular momentum operator — explains why the magnetic moments of many octahedral complexes are well approximated by the spin-only formula. His 1932 monograph The Theory of Electric and Magnetic Susceptibilities (Oxford University Press) became the standard reference for magnetic measurements in coordination chemistry.

Lars Erik Orgel and Leslie E. Orgel (not related) independently developed ligand field theory in the mid-1950s, incorporating covalent metal-ligand bonding into the crystal field framework. Leslie Orgel's 1960 monograph An Introduction to Transition-Metal Chemistry: Ligand-Field Theory ^{[Orgel 1960]} (Methuen) was the first textbook to present LFT as the natural extension of CFT, showing how the spectrochemical series, the nephelauxetic effect, and the failure of the point-charge model all point toward covalent bonding as an essential component.

Yukito Tanabe and Satoru Sugano published their systematic energy-level diagrams in 1954 ^{[Tanabe Sugano 1954]} in the Journal of the Physical Society of Japan 9, 753–766. These diagrams provided the first practical tool for extracting quantitative crystal field parameters from electronic spectra, transforming transition-metal spectroscopy from a qualitative art into a quantitative science. The Tanabe-Sugano approach remains the standard method for analysing d-d spectra in inorganic chemistry.

Bibliography [Master]

@article{Bethe1929,
  author = {Bethe, Hans},
  title = {Termaufspaltung in Kristallen},
  journal = {Annalen der Physik},
  volume = {3},
  pages = {133--208},
  year = {1929},
}

@book{VanVleck1932,
  author = {Van Vleck, J. H.},
  title = {The Theory of Electric and Magnetic Susceptibilities},
  publisher = {Oxford University Press},
  year = {1932},
}

@book{Orgel1960,
  author = {Orgel, L. E.},
  title = {An Introduction to Transition-Metal Chemistry: Ligand-Field Theory},
  publisher = {Methuen},
  year = {1960},
}

@article{TanabeSugano1954,
  author = {Tanabe, Yukito and Sugano, Satoru},
  title = {On the Paramagnetism of Complex Ions},
  journal = {J. Phys. Soc. Japan},
  volume = {9},
  pages = {753--766},
  year = {1954},
}

@book{Ballhausen1962,
  author = {Ballhausen, C. J.},
  title = {Introduction to Ligand Field Theory},
  publisher = {McGraw-Hill},
  year = {1962},
}

@book{HousecroftSharpe2018,
  author = {Housecroft, C. E. and Sharpe, A. G.},
  title = {Inorganic Chemistry},
  edition = {5th},
  publisher = {Pearson},
  year = {2018},
}

@book{CottonWilkinson1999,
  author = {Cotton, F. A. and Wilkinson, G.},
  title = {Advanced Inorganic Chemistry},
  edition = {6th},
  publisher = {Wiley},
  year = {1999},
}