Ligand Field Theory — What It Explains That Crystal Field Theory Cannot

By Aniket Bhardwaj · 9 September 2026 · Advanced Chemistry

Crystal field theory (CFT) takes you a long way. It gives you Δ_o, high spin versus low spin, CFSE, and the Jahn–Teller effect. But CFT is built on a lie that everybody agrees to tell in the second year: it treats ligands as structureless negative point charges and assumes the metal–ligand interaction is purely electrostatic. Ligand field theory (LFT) is what you get when you drop that assumption and rebuild the same picture using molecular orbital theory, keeping the useful CFT parameters as adjustable quantities fitted to experiment. This article shows exactly where CFT breaks, how the MO picture repairs it, and the numbers you are expected to handle at CSIR-NET, GATE and MSc level.

Four experiments CFT cannot explain

The spectrochemical series is in the wrong order for an electrostatic model. If ligands were point negative charges, field strength should track charge and size. It does not. OH⁻ (charged) produces a weaker field than neutral H₂O. CO and CN⁻ sit at the very top even though CO is neutral and has a tiny dipole.
The nephelauxetic effect. The interelectron repulsion parameters (the Racah parameters B and C) measured for a complex are always smaller than for the free gaseous ion. The d electrons behave as if their cloud has expanded. A point-charge model has nothing to expand into — the d orbitals in CFT are pure metal orbitals.
Direct spectroscopic evidence of ligand character in "metal" orbitals. The EPR spectrum of [IrCl₆]²⁻ shows hyperfine coupling of the unpaired electron to the chlorine nuclei. An electron in a pure metal d orbital could not couple to a ligand nucleus at all. It is delocalised onto the ligands.
Charge-transfer bands. The intense colours of MnO₄⁻ and CrO₄²⁻ are not d–d transitions — d⁰ ions have no d electrons to excite. They are ligand-to-metal charge transfer, which only exists if metal and ligand orbitals are mixed.

The σ-only octahedral MO scheme

Start with an ML₆ octahedron in the O_h point group and let each ligand supply one σ-donor orbital pointing at the metal. The six ligand σ orbitals form symmetry-adapted linear combinations (SALCs) that transform as a_1g + e_g + t_1u. The metal valence orbitals transform as:

Metal orbitals	Symmetry in O_h	σ partner among ligand SALCs?
(n−1)d_xy, d_yz, d_xz	t_2g	No — they point between the ligands
(n−1)d_z², d_x²−y²	e_g	Yes
ns	a_1g	Yes
np_x, np_y, np_z	t_1u	Yes

So six of the nine metal valence orbitals find a symmetry match and split into six bonding and six antibonding MOs; the t_2g set is left strictly non-bonding in a σ-only complex. Reading the MO diagram from the bottom, the energy order is:

a_1g < t_1u < e_g (bonding, mostly ligand) < t_2g (non-bonding, pure metal d) < e_g* (antibonding, mostly metal d) < a_1g* < t_1u*

The twelve electrons donated by the six ligands exactly fill a_1g, t_1u and e_g. The metal's own d electrons then occupy t_2g and e_g*. And here is the payoff:

Δ_o = E(e_g*) − E(t_2g)

This is the same Δ_o you used in CFT, occupying the same place in every formula — but it is now the gap between a non-bonding level and a σ-antibonding level, not the result of electrostatic repulsion. Stronger σ donation raises e_g* further and increases Δ_o. That single change of interpretation is the heart of LFT.

Worked example 1 — electron bookkeeping in [Co(NH₃)₆]³⁺.
Co(III) is d⁶. Six NH₃ ligands donate 2 electrons each = 12 electrons.
Bonding MOs filled: a_1g² t_1u⁶ e_g⁴ = 12 electrons — the ligand lone pairs.
Metal d electrons: 6, and because Δ_o > P (pairing energy) for ammonia on Co(III), they occupy t_2g⁶ e_g*⁰.
Total valence electron count = 12 + 6 = 18, diamagnetic, low spin — the same answer CFT gives, reached by counting molecular orbitals instead of point charges.

π bonding — the part CFT simply does not have

Most ligands also have orbitals of π symmetry with respect to the M–L axis, and in O_h the ligand π set spans t_1g + t_1u + t_2g + t_2u. Only the t_2g component matches the metal d_xy/d_yz/d_xz orbitals — the very set that was non-bonding a moment ago. Two cases follow, and between them they explain the spectrochemical series.

Ligand type	Relevant ligand π orbitals	Effect on metal t_2g	Effect on Δ_o	Examples
π donor	Filled p (or π) orbitals below the metal d level	Pushed up; t_2g becomes π-antibonding	Decreases — weak field, favours high spin	I⁻, Br⁻, Cl⁻, F⁻, OH⁻, O²⁻, RS⁻
σ only	None of t_2g symmetry	Unchanged, non-bonding	Intermediate	NH₃, en, H⁻, CH₃⁻
π acceptor	Empty π* orbitals above the metal d level	Pushed down; t_2g becomes π-bonding	Increases — strong field, favours low spin	CO, CN⁻, NO⁺, PR₃, bipy, phen

Now the spectrochemical series reads as chemistry rather than as a list to memorise:

I⁻ < Br⁻ < S²⁻ < SCN⁻ < Cl⁻ < F⁻ < OH⁻ < C₂O₄²⁻ < H₂O < NCS⁻ < NH₃ < en < bipy < phen < NO₂⁻ < PPh₃ < CN⁻ < CO
← π donors ····· σ-only ····· π acceptors →

OH⁻ below H₂O stops being a paradox: hydroxide is the better π donor of the two (it has more available lone pairs and a full negative charge to push), so it lowers Δ_o more. CO tops the series not because of charge but because its π* orbitals are a superb acceptor. The same π back-bonding is what makes metal carbonyls stable and lowers their C–O stretching frequency in the IR.

The angular overlap model — LFT with numbers

The angular overlap model (AOM) parameterises each metal–ligand interaction by a σ term e_σ and a π term e_π. For a regular octahedron:

Δ_o = 3e_σ − 4e_π
e_π > 0 for a π donor, e_π < 0 for a π acceptor, e_π = 0 for a σ-only ligand.

Worked example 2 — how π character moves Δ_o.
(a) A σ-only ligand with e_σ = 7500 cm⁻¹, e_π = 0:
Δ_o = 3(7500) − 4(0) = 22 500 cm⁻¹.

(b) A π donor with the same σ strength, e_σ = 7500, e_π = +1500 cm⁻¹:
Δ_o = 3(7500) − 4(1500) = 22 500 − 6000 = 16 500 cm⁻¹.

(c) A π acceptor, e_σ = 7500, e_π = −1500 cm⁻¹:
Δ_o = 22 500 + 6000 = 28 500 cm⁻¹.

Identical σ donation, and yet the field strength spans 16 500 to 28 500 cm⁻¹ purely through π effects. This is why CFT, which contains no π term at all, can never reproduce the series.

Worked example 3 — converting Δ_o into the units an examiner asks for.
[Co(NH₃)₆]³⁺ has Δ_o ≈ 22 900 cm⁻¹. The conversion constant is 1 cm⁻¹ = hcN_A = 11.9627 J mol⁻¹.
To kJ mol⁻¹: 22 900 × 11.9627 = 273 946 J mol⁻¹ = 273.9 kJ mol⁻¹.
To wavelength: λ = 1/(22 900 cm⁻¹) = 4.367 × 10⁻⁵ cm = 437 nm.
Cross-check on the colour: absorbing near 437 nm (blue-violet) leaves the transmitted light yellow-orange, and [Co(NH₃)₆]³⁺ is indeed a yellow-orange ion. The arithmetic agrees with the bottle.

The nephelauxetic effect — measuring covalency

"Nephelauxetic" means cloud expanding. Because the d electrons are partly delocalised onto the ligands, they are on average further apart than in the free ion, so they repel each other less. The Racah parameter B shrinks, and the ratio is the nephelauxetic parameter:

β = B(complex) ⁄ B(free ion) (β < 1 always)

Worked example 4 — reading β as a covalency scale.
The free Cr³⁺ ion has B tabulated at about 918 cm⁻¹. For [Cr(H₂O)₆]³⁺, spectral fitting gives B ≈ 725 cm⁻¹.
β = 725 ⁄ 918 = 0.790, i.e. about a 21% reduction in interelectron repulsion.
A purely ionic bond would give β = 1. The measured 0.79 is a direct experimental statement that roughly a fifth of the repulsion has been lost to delocalisation — which is exactly the quantity CFT sets to zero by construction.

Ordering ligands by how much they reduce B gives the nephelauxetic series, and it is not the spectrochemical series:

F⁻ < H₂O < NH₃ < en < C₂O₄²⁻ < NCS⁻ < Cl⁻ < CN⁻ < Br⁻ < I⁻
increasing covalency, decreasing β

Note where the halides sit. Iodide is near the bottom of the spectrochemical series (weak field) but near the top of the nephelauxetic series (very covalent). The two series measure different things: one measures orbital splitting, the other measures delocalisation. Confusing them is one of the commonest errors in NET answers.

What each theory is actually good for

Question	Crystal field theory	Ligand field theory
d-orbital splitting pattern, high/low spin, CFSE	Correct and fast	Same answers, more work
Magnetic moments, Jahn–Teller activity	Correct	Same
Why the spectrochemical series has that order	Fails	π donor / acceptor argument
Nephelauxetic effect, Racah B reduction	Fails (no covalency)	Natural consequence
Charge-transfer bands, colour of MnO₄⁻	Fails	LMCT / MLCT between MOs
Metal carbonyl stability, ν(CO) lowering	Fails	π back-bonding into CO π*
18-electron rule, organometallic bonding	Not addressed	Direct MO electron count

Mistakes that cost marks

Saying "LFT replaces Δ_o". It does not. Δ_o survives unchanged as a parameter; only its physical origin changes, from electrostatic repulsion to a σ-antibonding/non-bonding gap modified by π interactions.
Forgetting that t_2g is non-bonding only in the σ-only case. Once any π-active ligand is present, t_2g is either π-bonding or π-antibonding. Half of the exam questions on this topic turn on exactly that sentence.
Getting the direction of the π effect backwards. π donor pushes t_2g up and shrinks Δ_o; π acceptor pulls t_2g down and grows Δ_o. A quick sanity check: CO is the best acceptor and sits at the strong-field end, so acceptors must increase Δ_o.
Mixing up the spectrochemical and nephelauxetic series. I⁻ is weak field and highly covalent. There is no contradiction — different property, different ordering.
Assigning MnO₄⁻'s colour to a d–d transition. Mn(VII) is d⁰. Any "d–d" answer here is automatically wrong.
Quoting β > 1. Delocalisation can only reduce interelectron repulsion, so β is always less than 1. A value above 1 means you inverted the ratio.
Using the sign convention of e_π carelessly. In Δ_o = 3e_σ − 4e_π, an acceptor carries a negative e_π, so the minus sign becomes a plus. State your convention in the answer.

Where this appears in the exam

Exam	Typical demand
CSIR-NET Chemical Sciences	Explain a spectrochemical-series anomaly; identify π donor vs acceptor from a ligand list; calculate or interpret β; assign LMCT vs MLCT
GATE Chemistry	Symmetry labels of the σ SALCs; which metal orbital is non-bonding; effect of π bonding on Δ_o and spin state; AOM expression
IIT-JAM / CUET-PG	Ordering ligands by field strength and justifying it; total valence-electron counts
MSc coursework	Constructing the full O_h MO diagram including π; relating ν(CO) in carbonyls to back-bonding; Racah parameters from a d–d spectrum

Get the units right before you compare any two ligand fields. Ligand-field problems switch between cm⁻¹, kJ mol⁻¹, eV and nanometres inside a single question, and that is where marks are lost — not in the symmetry. The ABC Chemistry Calculator Suite keeps a unit converter, the scientific constants and a full scientific calculator on one page so you can check each conversion as you go.

Open the ABC Chemistry Calculator Suite →

Preparing for CSIR-NET, GATE, IIT-JAM or CUET-PG? ABC Chemistry runs dedicated competitive-exam batches at the coaching centre and online across India — details at abcchemistry.in.