Lanthanide Chemistry and Electronic Structure

A vial of Nd³⁺ solution glows purple-lavender under white light. Hold Er³⁺ next to it and it is a delicate salmon-pink. Neither color changes appreciably whether the ion sits in water, in a nitrate crystal, or clamped inside a macrocyclic MRI contrast agent. This stability would be inconceivable for a Cr³⁺ or Co²⁺ complex. Their colors swing wildly with ligand field strength. This immunity to environment is the single most diagnostic fact about the lanthanides. It follows directly from where the active electrons live. They are buried in the 4f shell, radially inside the filled 5s and 5p shells. This shields them from the chemical world outside the ion. The lanthanides — the fifteen elements from lanthanum (Z = 57) through lutetium (Z = 71) — matter far beyond being a curiosity row parked below the main periodic table. Neodymium magnets run every hard drive and electric motor built in the last thirty years. Europium and terbium phosphors make LED and plasma-display colors possible. Gadolinium complexes are the contrast agents in roughly a third of clinical MRI scans. Cerium oxide is the workhorse of catalytic converters and glass polishing. Understanding why these fifteen elements behave almost identically to each other is important. Yet they remain indispensable and non-interchangeable in technology. This requires taking the electronic structure of the 4f shell seriously. The big picture is a study in contrasts with the transition metals covered under Oxidation States and Redox Trends Across the d-Block. In the d-block, 3d orbitals are valence orbitals that mix strongly with ligand orbitals. This produces strong crystal-field splitting, intense charge-transfer bands, and oxidation-state versatility. The 4f orbitals, by contrast, are core-like. They are radially contracted, angular-momentum-rich (ℓ = 3, seven orbitals, up to fourteen electrons), and almost bonding-innocent. The dominant oxidation state across the whole series is +3. Coordination numbers routinely reach 8–12. Bonding is overwhelmingly ionic/electrostatic. Optical spectra consist of narrow, "atomic-like" f–f lines. These are governed by Russell–Saunders term symbols rather than broad crystal-field-split d–d bands. This article builds the electronic-structure picture from the aufbau anomalies of the 4f/5d/6s manifold. It explains why +3 dominates with controlled exceptions (Ce⁴⁺, Eu²⁺, Yb²⁺, Sm²⁺, Tb⁴⁺). It derives why f–f transitions are simultaneously parity-forbidden and yet weakly observed. It works through the coordination chemistry consequences: high coordination numbers, ionic radii control via the The Lanthanide Contraction and Its Consequences, and magnetism. This magnetism is dictated by spin–orbit coupled J states. This contrasts with quenched orbital angular momentum as in 3d ions. The 5f actinides are treated separately in Actinide Chemistry and the 5f Series. They will show what happens when this shielding logic partially breaks down.

Key takeaways

The lanthanides (La–Lu, Z = 57–71) fill the 4f subshell; 4f orbitals are radially contracted inside the filled 5s²5p⁶ shells, so they barely interact with ligands.
+3 is overwhelmingly the dominant oxidation state, set by the balance of ionization energies against lattice/hydration energy; +2 and +4 occur only where they reach or approach empty, half-filled, or filled 4f configurations (Eu²⁺ f⁷, Yb²⁺ f¹⁴, Ce⁴⁺ f⁰, Tb⁴⁺ f⁷).
Because ligand fields barely perturb 4f orbitals, spin-orbit coupling dominates over crystal-field splitting — energy levels are labeled by whole-atom term symbols ²ˢ⁺¹L_J, not by d-orbital-splitting diagrams.
f–f transitions are Laporte-forbidden (parity, Δℓ = 0 within f→f) yet observed as sharp, weak lines because of vibronic coupling and J-mixing — giving characteristic pastel colors and legendary optical purity (Nd:YAG, Er-doped fiber amplifiers).
Coordination numbers of 8–12 are routine because large, hard Ln³⁺ ions bond electrostatically to hard O/N donors with no orbital-directionality constraint.
Magnetism follows the Landé-g-factor/Hund's-rules J ground state (Van Vleck theory), not the spin-only formula used for most 3d ions — Gd³⁺ (f⁷, ⁸S₇/₂) is the outlier that behaves almost spin-only.
Key idea: the chemistry of the lanthanides is the chemistry of a nearly invariant +3 ionic core whose 4f electrons are spectroscopic spectators, shielded by outer 5s/5p shells, so periodicity here is written in ionic radius and f-electron count, not in orbital bonding.

Where this lives today

Neodymium and samarium-cobalt permanent magnets (Nd₂Fe₁₄B) exploit the huge, unquenched orbital angular momentum of the lanthanide 4f ions (the same L ≠ 0, large-J physics worked out for Dy³⁺ above) to generate enormous magnetocrystalline anisotropy — this is also why dysprosium and terbium are added to Nd-Fe-B magnets used in electric-vehicle motors and wind-turbine generators, to boost coercivity at operating temperature by exploiting their large orbital moments. Europium and terbium complexes are the red and green phosphors in older CRT/plasma displays and remain standard emissive dopants in LED phosphors and security inks, again directly exploiting the sharp, environment-insensitive f–f (and f–d, for Eu²⁺) emission lines derived above. Gadolinium(III) complexes (DTPA, DOTA-based chelates) are the paramagnetic contrast agents used in a large fraction of clinical MRI scans, chosen specifically because Gd³⁺'s ⁸S₇/₂ ground state gives the largest possible, nearly isotropic magnetic moment among stable trivalent ions with a long electronic relaxation time — a direct payoff of the L = 0 analysis in Example 2. Cerium oxide's facile Ce⁴⁺/Ce³⁺ (4f⁰/4f¹) redox cycle underlies its use as an oxygen-storage component in automotive three-way catalytic converters and as a chemical-mechanical polishing abrasive for silicon wafers and glass. Erbium-doped fiber amplifiers, which made modern long-haul fiber-optic internet infrastructure practical, rely on the sharp, well-defined ⁴I₁₃/₂ → ⁴I₁₅/₂ f–f transition of Er³⁺ sitting at almost exactly the 1550 nm minimum-loss window of silica fiber — a direct technological payoff of the narrow, host-insensitive spectral lines this article derives from parity-forbidden but vibronically-allowed f–f transitions. Lanthanide luminescent probes (Eu³⁺, Tb³⁺ chelates) are widely used in time-resolved fluorescence immunoassays precisely because their millisecond-scale excited-state lifetimes — orders of magnitude longer than ordinary organic fluorophores — follow directly from how strongly parity-forbidden the emissive f–f transition is.

Ask the tutor: How does the lanthanide contraction quantitatively change ionic radius and coordination number across the series, and why does it make the post-lanthanide 5d metals resemble their 4d congeners so closely?

The physical picture

Start with where the electrons actually sit. For a free lanthanide atom the ground-state configuration is, with a few irregularities, [Xe]4fⁿ6s² (n = 0 at La up to 14 at Lu), with cerium showing [Xe]4f¹5d¹6s² and a handful of others (Gd, at f⁷, and the heavy end near Lu) also borrowing a 5d electron because half-filled and filled f-shells are stabilized by exchange energy — precisely the same logic that gives Cr [Ar]3d⁵4s¹ and Cu [Ar]3d¹⁰4s¹ in the 3d row. Once you strip three electrons to make the Ln³⁺ ion — always losing 6s² and, where present, the 5d/extra 4f electron first — you get the clean, defining sequence:

\text{Ln}^{3+}: [\text{Xe}]4f^{n}, \quad n = 0\ (\text{La}^{3+}) \ldots 14\ (\text{Lu}^{3+})

The crucial structural point is radial: a 4f orbital, despite having principal quantum number 4, is more contracted than the 5s and 5p orbitals that are nominally "inner" by n. This is the lanthanide analogue of the orbital contraction that makes 3d fill before 4s empties efficiently — poor shielding by other f electrons (the 4f wavefunction has essentially no radial nodes and a probability density that peaks close to the nucleus) means each successive 4f electron feels a rapidly increasing effective nuclear charge, pulling the whole 4f shell in tight against the core. The result: the 4f density lies almost entirely inside the outer 5s²5p⁶ shell, which acts as an electrostatic Faraday cage. Ligand orbitals approaching the ion overlap with 5s/5p (and 6s/5d, used for bonding) density, not with 4f density. The 4f electrons are chemically inert spectators; they are, however, spectroscopically and magnetically everything, because their many-electron term structure — not any ligand field — sets the energy levels the ion can access.

Figure

Radial charge density: 4f is buried inside filled 5s/5p shells

Because 4f is shielded, the ligand field splitting of f-orbital energies in a real complex is tiny — typically 100 cm⁻¹ or less — while the free-ion electron-electron repulsion and spin–orbit coupling that generate Russell–Saunders terms are enormous by comparison, on the order of 10⁴ cm⁻¹ for interelectron repulsion and 10³ cm⁻¹ for spin-orbit coupling in the heavier lanthanides. This ordering — spin-orbit ≫ crystal field — is the exact opposite of the regime used for first-row transition metals (where crystal field ≫ spin-orbit), and it is why lanthanide spectroscopy is organized by atomic term symbols ²ˢ⁺¹L_J rather than by Oₕ/Td crystal-field splitting diagrams like eg/t2g.

Figure

Energy-scale hierarchy: d-block vs f-block

The consequence for oxidation states is more subtle than "the 4f shell is inert, so all lanthanides are +3." Ionization is a whole-ion energetic balance: the sum of the first three ionization energies, minus the lattice/hydration energy released on forming the salt or complex, decides which oxidation state is thermodynamically accessible. IE₁+IE₂+IE₃ is remarkably flat across the series (roughly 3400–3900 kJ/mol), while the huge charge density of the small, highly charged Ln³⁺ ion generates a correspondingly huge lattice or hydration energy that comfortably pays for three ionizations in virtually every lanthanide. Reaching +2 or +4 additionally requires either removing a fourth electron (feasible only when the resulting f-configuration is especially stable — f⁰, f⁷, f¹⁴) or failing to remove a third electron (stopping at f⁷ or f¹⁴ +2 configurations) — and even then, only under special conditions (aqueous reduction potentials, or solid-state/high-temperature synthesis) because the lattice-energy payoff for +3 is normally too large to give up.

The core ideas

1. The dominant ground-state pattern.

\text{Ln} \; ([\text{Xe}]4f^{n}5d^{0\text{ or }1}6s^2) \;\xrightarrow{-3e^-}\; \text{Ln}^{3+}\;([\text{Xe}]4f^{n})

where: n runs 0 (La) to 14 (Lu); the two 6s electrons and any 5d electron are always removed before any 4f electron, because 6s/5d lie at higher energy in the ion despite occasional 4f/5d near-degeneracy in the neutral atom. Reading: once ionized to +3, every lanthanide has an "inert-gas-plus-fⁿ" core; the electron count n is what changes from element to element, and it is n (through term symbols and radius) — not new bonding patterns — that drives the differences down the series.

2. Non-+3 oxidation states rationalized by special configurations.

\Delta E_{\text{access}} = \left(\sum \text{IE up to desired charge}\right) - \left(\text{lattice or solvation energy released}\right)

where: the oxidation state realized is the one that minimizes the net free energy of formation of the compound; deviations from +3 (Ce⁴⁺ = f⁰, Eu²⁺ = f⁷, Yb²⁺ = f¹⁴, Sm²⁺ ≈ f⁶, Tb⁴⁺ = f⁷) occur where the extra stability of an empty, half-filled, or filled 4f shell (via reduced interelectron repulsion / maximized exchange energy) lowers the relevant ionization energy enough, or raises the electron affinity enough, to tip the balance. Reading: this is a direct 4f-shell analogue of the Cr/Mn/Fe/Cu half-filled-shell stability arguments used for 3d ions — same physics (exchange energy maximized at half-filling), applied one shell further from the nucleus.

3. Term symbols and the ground-state J level (Hund's rules + spin-orbit coupling).

{}^{2S+1}L_J, \qquad J = |L-S| \; (\text{shell less than half-full}) \;\; \text{or} \;\; J = L+S \;(\text{shell more than half-full})

where: S and L are obtained by filling the seven 4f orbitals (mₗ = +3…−3) according to Hund's rules (maximize S first, then maximize L consistent with the Pauli principle), and J selects the spin-orbit-coupled level that is lowest in energy given the sign of the spin-orbit coupling constant (positive/less-than-half-full: J minimal; negative/more-than-half-full: J maximal). Reading: this single rule generates the entire "f-electron periodic table" of ground states — it is why Nd³⁺ (f³) is ⁴I₉/₂, Eu³⁺ (f⁶) is ⁷F₀, and Gd³⁺ (f⁷, half-filled, L = 0) is the pure-spin ⁸S₇/₂ that behaves almost like a 3d ion magnetically.

4. The Laporte/parity selection rule and how f–f transitions still appear.

\Delta \ell = 0 \; (\text{f} \to \text{f}) \Rightarrow \text{electric-dipole forbidden by parity (Laporte rule)}

where: an electric dipole operator has odd parity, so a transition between two states of the same parity (both 4f, both even) has zero electric-dipole matrix element in a centrosymmetric environment — yet real complexes show weak but clearly visible f–f bands (molar absorptivities ε ≈ 1–10 M⁻¹cm⁻¹, vs. 10³–10⁴ for fully allowed d–d or charge-transfer bands). Reading: the "forbidden but visible" paradox is resolved by vibronic coupling (odd-parity vibrations instantaneously break centrosymmetry, borrowing intensity from opposite-parity 5d states) and by any static lack of an inversion center in the real coordination geometry, exactly the same escape mechanisms used to explain weak d–d bands in centrosymmetric transition-metal complexes, but here even more pronounced because the "forbiddenness" is more rigorous (pure parity forbiddenness plus, for many transitions, spin-forbiddenness too).

Derivation AWhy +3 dominates: the thermodynamic tug-of-war between ionization energy and lattice energy

Goal. Show quantitatively why Ln³⁺ is thermodynamically preferred over Ln²⁺ or Ln⁴⁺ for a typical lanthanide, using a Born–Haber-type argument, and show why Eu and Yb are the recognized exceptions that support +2.

Step 1 — Set up the Born–Haber cycle for a generic lanthanide halide LnX₃(s). Formation of the solid from the elements can be split into: (i) atomization of the metal, (ii) atomization/dissociation of X₂ (or ½X₂ per X), (iii) ionization of Ln to Ln³⁺ (sum of IE₁, IE₂, IE₃), (iv) electron attachment to X to give X⁻, and (v) lattice assembly of Ln³⁺ and 3X⁻ into the solid, releasing lattice energy U. The overall enthalpy of formation is

\Delta H_f = \Delta H_{\text{atom,Ln}} + 3\Delta H_{\text{atom,X}} + (IE_1+IE_2+IE_3) + 3EA_X - U

Step 2 — Compare the energetic cost of stopping at Ln²⁺ instead of continuing to Ln³⁺. Forming LnX₂(s) instead removes IE₃ from the sum but also loses one Ln³⁺–X⁻ worth of lattice interactions (replaced by a lower-charge lattice with a smaller Madelung energy per formula unit, since the lattice energy scales roughly with the product of ionic charges, $U \propto z_+z_-/r$ ). Quantitatively, U(LnX₃) is typically 1.5–2× larger than U(LnX₂) because $z_+$ increases from 2 to 3 while r changes only modestly (lanthanide contraction is a small effect ion-to-ion). For a representative case, taking IE₃ for a mid-lanthanide (e.g., Nd) around 3900 kJ/mol and the extra lattice stabilization gained by going from a 2+ to a 3+ salt on the order of 4000–4500 kJ/mol (typical of small, highly charged M³⁺ halide lattices), the net balance:

\Delta(\Delta H_f)_{2^+\to 3^+} \approx \underbrace{-U_{\text{gain}}}_{\approx -4200\ \text{kJ/mol}} + \underbrace{IE_3}_{\approx +3900\ \text{kJ/mol}} \approx -300\ \text{kJ/mol}

is favorable (more negative ΔH_f) toward +3 for a typical lanthanide — the lattice-energy gain outweighs the cost of the third ionization.

Step 3 — Now do the same comparison for Eu, where IE₃ is anomalously high. Eu has configuration [Xe]4f⁷6s²; Eu²⁺ is [Xe]4f⁷, a half-filled, exchange-stabilized shell. Removing a further electron from this half-filled shell to reach Eu³⁺ ([Xe]4f⁶) costs substantially more energy than the "typical" IE₃ used above — experimentally IE₃(Eu) ≈ 4110 kJ/mol, distinctly higher than neighboring Sm (≈3900) or Gd (≈3550), because you are breaking into a stabilized half-filled shell. Re-running the Step 2 balance with this larger IE₃ and the same lattice-gain estimate:

\Delta(\Delta H_f)_{2^+\to 3^+}(\text{Eu}) \approx -4200 + 4110 \approx -90\ \text{kJ/mol}

— the margin favoring +3 shrinks dramatically. In practice this residual margin is still enough that Eu(III) is the stable aqueous and solid-state oxide/halide state, but it is thin enough that Eu²⁺ is readily accessible: EuCl₂ is a stable, isolable solid, Eu²⁺(aq) exists transiently and can be trapped in Sr²⁺/Ba²⁺ host lattices (isostructural substitution, similar ionic radius), and Eu²⁺ is a common activator ion in phosphors precisely because the 2+/3+ gap is anomalously small.

Step 4 — Confirm the trend with Yb (f¹⁴ target) and contrast with a "normal" ion like Nd. Yb²⁺ is [Xe]4f¹⁴ (filled shell, exchange-stabilized); IE₃(Yb) ≈ 4210 kJ/mol, again anomalously high (compare Tm ≈ 3550, Lu ≈ 4370 — note Lu is even higher because Lu³⁺ = f¹⁴ is already the stable target and removing a further electron from a filled shell to hypothetically reach Lu⁴⁺ would be prohibitive, which is why Lu never shows +2 or +4). Yb sits at the edge where +2 is a well-known, moderately stable oxidation state (YbCl₂, Yb metal itself is atypically volatile and divalent-like in some alloys), mirroring the Eu case. Nd (f³, no special stability at f² or f⁴) shows only trace, highly unstable Nd²⁺ under forcing non-aqueous conditions, consistent with a normal IE₃ and the Step 2 estimate.

Go deeper · Derivation B · 2 minDeriving the ground-state term symbol for Nd³⁺ (4f³) via Hund's rules and spin–orbit coupling

Derivation BDeriving the ground-state term symbol for Nd³⁺ (4f³) via Hund's rules and spin–orbit coupling

Goal. Work out the full Russell–Saunders ground term for Nd³�sup>3+ ([Xe]4f³) and show it matches the experimentally assigned ⁴I₉/₂ ground state used in Nd-laser spectroscopy.

Step 1 — Lay out the seven 4f orbitals by mₗ and place three electrons with Hund's first rule (maximize S). The 4f subshell has mₗ = +3, +2, +1, 0, −1, −2, −3. To maximize total spin S with only 3 electrons (out of 14 possible), put one electron in each of the three highest-mₗ orbitals, all with parallel spin:

m_l: +3\;(\uparrow), \quad +2\;(\uparrow), \quad +1\;(\uparrow)

giving total spin quantum number

S = \tfrac{1}{2}+\tfrac{1}{2}+\tfrac{1}{2} = \tfrac{3}{2} \quad\Rightarrow\quad 2S+1 = 4

Step 2 — Apply Hund's second rule (maximize L, subject to maximum S already fixed) by choosing the highest possible sum of mₗ. With the three electrons already forced into the three highest available mₗ slots to keep all spins parallel (Pauli forbids two parallel-spin electrons in the same mₗ orbital), the total

M_L^{\max} = 3+2+1 = 6 \quad\Rightarrow\quad L = 6

L = 6 corresponds to the spectroscopic letter I (S,P,D,F,G,H,I for L = 0,1,2,3,4,5,6). So the term is ⁴I (quartet I).

Step 3 — Determine J from the more-than/less-than half-filled rule. The 4f shell holds 14 electrons at most; 3 electrons is less than half-filled (half = 7). For a less-than-half-filled shell, the ground J is the minimum possible value,

J = |L - S| = \left|6 - \tfrac{3}{2}\right| = \tfrac{9}{2}

so the ground term is

{}^{4}I_{9/2}

Step 4 — Cross-check with the physical origin of the rule and connect to the well-known laser transition. The "minimize J when less than half full" rule comes from the sign of the one-electron spin-orbit constant ζ, which is positive for a shell that is less than half full (spin-orbit coupling then favors small J), and effectively negative (via the hole formalism) for more than half full, favoring maximal J — this is exactly analogous to how Hund's rules are applied to 3d ions like Ti³⁺ (d¹, ²D₃/₂) vs. Cu²⁺ (d⁹, hole formalism, ²D₅/₂). The result ⁴I₉/₂ for Nd³⁺ is the textbook ground state; the famous 1.064 μm Nd:YAG laser line is the ⁴F₃/₂ → ⁴I₁₁/₂ transition, a transition between two different excited/ground J-multiplets of the same fⁿ configuration, consistent with this whole free-ion term structure being only weakly perturbed by the YAG host crystal field (which merely splits each J level into a handful of closely spaced Stark sublevels, the source of the laser's fine vibronic structure, not a new set of "d-orbital-like" levels).

Worked examples

Example 1 (foundational): Ground-state configuration and term symbol of Eu³⁺. Eu has Z = 63, neutral configuration [Xe]4f⁷6s². Losing the 6s² electrons plus one further electron (removed from the exchange-stabilized 4f⁷ shell, as quantified in the Born–Haber derivation above) gives

\text{Eu}^{3+}: [\text{Xe}]4f^{6}

Apply Hund's rules to 4f⁶ (six electrons, less than half of 14): Maximize S: place electrons singly with parallel spin in the six highest mₗ slots (+3,+2,+1,0,−1,−2), all ↑:

S = 6 \times \tfrac12 = 3 \quad\Rightarrow\quad 2S+1 = 7

Maximize L:

M_L^{\max} = 3+2+1+0+(-1)+(-2) = 3 \quad\Rightarrow\quad L = 3 \;(\text{term letter F})

Determine J: 6 electrons is less than half-filled (half = 7), so J is minimized:

J = |L - S| = |3-3| = 0

Ground term: ⁷F₀. This is a famous, chemically consequential result: Eu³⁺ has a non-magnetic (J = 0) ground state at low temperature, so Eu³⁺ complexes show anomalously weak paramagnetism at low T relative to the naive spin-only or even full Hund's-rule expectation, and only become measurably paramagnetic as thermal population of the low-lying ⁷F₁ (J=1, only ~350–500 cm⁻¹ above ⁷F₀) sets in. This is a standard "gotcha" in lanthanide magnetism problems. Limiting-case check: apply the same recipe to n = 0 (La³⁺, 4f⁰): trivially S = 0, L = 0, J = 0, term ¹S₀ — correctly predicting La³⁺ is diamagnetic and colorless, consistent with it having no f electrons at all (and, at the other end, Lu³⁺, 4f¹⁴, is a filled shell, also ¹S₀, diamagnetic and colorless) — the theory correctly brackets both empty-shell termini of the series with the same trivial, verifiable term symbol.

Example 2 (applied): Why does GdCl₃ dissolve to give an ion with almost exactly the spin-only magnetic moment, while DyCl₃ does not? Gd³⁺ is 4f⁷. Apply Hund's rules: maximize S by placing one electron in each of the seven mₗ orbitals, all parallel:

S = 7 \times \tfrac12 = \tfrac72 \quad\Rightarrow\quad 2S+1=8

With all seven orbitals singly occupied, $M_L^{\max} = 3+2+1+0-1-2-3 = 0$ , so L = 0 (term letter S). Since L = 0, there is no orbital angular momentum to couple with spin, so J = S = 7/2 trivially, and the ground term is ⁸S₇/₂ — a pure-spin state, structurally identical in form to the high-spin d⁵ configuration of Mn²⁺/Fe³⁺ (also ⁶S₅/₂-type, pure-spin ground states with L=0). For a pure-spin state the Landé g-factor is exactly 2 (since orbital angular momentum contributes nothing), and the effective magnetic moment reduces to the same spin-only formula used for high-spin 3d ions:

\mu_{\text{eff}} = g\sqrt{J(J+1)}\,\mu_B = 2\sqrt{\tfrac72\cdot\tfrac92}\,\mu_B = 2\sqrt{15.75}\,\mu_B \approx 7.94\ \mu_B

which matches the experimental moment of Gd³⁺ (≈7.9–8.0 μB) essentially exactly — Gd³⁺ is the one lanthanide ion for which the simple 3d-style spin-only picture works, because L = 0 removes the usual complication. Contrast with Dy³⁺, 4f⁹ (more than half-filled, since half = 7). Nine electrons: fill all seven mₗ orbitals singly (↑) using 7 electrons (S from those 7 = 7/2), then the remaining 2 electrons must pair up in the two highest-mₗ orbitals (+3, +2), each now doubly occupied, which cancels 2 of the 7 spins:

S = \tfrac72 - 2\times\tfrac12 = \tfrac52 \quad (2S+1 = 6)

M_L^{\max} = (3+3) + (2+2) + 1+0-1-2-3 = 6+4+1+0-1-2-3 = 5 \quad (L = 5,\ H)

Since 4f⁹ is more than half-filled, J is maximized:

J = L+S = 5+\tfrac52 = \tfrac{15}{2}

Ground term ⁶H₁₅/₂, with large L = 5 — substantial orbital angular momentum survives, unquenched (unlike in 3d ions, where the ligand field usually quenches orbital angular momentum for anything but a perfectly octahedral/tetrahedral T-term case). The full Landé formula is required:

g_J = 1 + \frac{J(J+1)+S(S+1)-L(L+1)}{2J(J+1)}

g_J = 1 + \frac{\tfrac{15}{2}\cdot\tfrac{17}{2} + \tfrac{5}{2}\cdot\tfrac{7}{2} - 5\cdot6}{2\cdot\tfrac{15}{2}\cdot\tfrac{17}{2}} = 1 + \frac{127.5+8.75-30}{127.5} \approx 1 + \frac{106.25}{127.5} \approx 1.833

\mu_{\text{eff}} = g_J\sqrt{J(J+1)}\,\mu_B = 1.833\sqrt{\tfrac{15}{2}\cdot\tfrac{17}{2}}\,\mu_B = 1.833\sqrt{63.75}\,\mu_B \approx 1.833\times7.98 \approx 14.6\ \mu_B

This matches the experimental Dy³⁺ moment (≈10.5–10.7 μB is the often-quoted lab value, but the free-ion Landé calculation above of ≈14.6 μB substantially overshoots — an important, commonly tested discrepancy). The resolution: this Landé free-ion formula assumes the full multiplet is thermally accessible and undistorted; in practice, low-symmetry crystal-field splitting of the ⁶H₁₅/₂ multiplet into Stark sublevels, combined with a J-mixing correction, brings the experimental room-temperature moment down toward the more commonly tabulated ≈10.6 μB. This is precisely why Dy³⁺ (and Tb³⁺, Ho³⁺ similarly) are treated as "anomalous" in undergraduate magnetism tables relative to the clean agreement seen for Gd³⁺: L ≠ 0 ions are always somewhat sensitive to their environment even though the sensitivity is orders of magnitude weaker than for a 3d ion. Limiting-case check: setting L = 0 in the general Landé formula recovers exactly g_J = 2 and the spin-only result used for Gd³⁺ above:

g_J\Big|_{L=0} = 1 + \frac{J(J+1)+S(S+1)-0}{2J(J+1)} = 1 + \frac{S(S+1)}{J(J+1)} = 1 + \frac{S(S+1)}{S(S+1)} = 2

(using J = S when L = 0), confirming the Dy-style general formula correctly collapses to the simple Gd-style spin-only case.

Example 3 (exam/competition level): Predict whether CeO₂ or Ce₂O₃ is the thermodynamically stable cerium oxide in air, and rationalize using 4f⁰/4f¹ configurations. Cerium: [Xe]4f¹5d¹6s². Ce³⁺ = [Xe]4f¹ (one f electron); Ce⁴⁺ = [Xe]4f⁰ (empty shell, exceptionally stable — same "special configuration" logic as Eu²⁺/f⁷ and Yb²⁺/f¹⁴, but here favoring the higher, not lower, oxidation state because emptying rather than half-filling/filling is the stabilizing target). Set up the comparison as an oxidation half-reaction competition in air (O₂ present, oxidizing conditions):

\text{Ce}^{3+} \rightarrow \text{Ce}^{4+} + e^- \qquad E^\circ \approx +1.72\ \text{V (vs. SHE, aqueous)}

This is unusually accessible for a lanthanide (compare: no other Ln³⁺→Ln⁴⁺ couple is anywhere near this reachable in aqueous solution outside Ce and, with much more difficulty, Pr/Tb under forcing conditions) — because Ce⁴⁺ reaches the especially stable empty 4f⁰ shell, sharply lowering IE₄ relative to a "generic" fourth ionization. In the solid state, CeO₂ (fluorite structure, Ce⁴⁺, colorless-to-pale-yellow, 4f⁰) is the thermodynamically stable oxide formed when cerium metal or Ce₂O₃ is heated in air/O₂, because:

2\text{Ce}_2\text{O}_3(s) + \text{O}_2(g) \rightarrow 4\text{CeO}_2(s), \qquad \Delta G^\circ \ll 0

is strongly exergonic — the lattice-energy gain from the smaller, more highly charged Ce⁴⁺ ion (fluorite CeO₂, an exceptionally robust lattice used industrially as an oxygen-storage material in three-way catalytic converters, exploiting the facile CeO₂ ⇌ Ce₂O₃ + O₂ redox cycle) combined with the 4f⁰ stabilization outweighs the extra ionization cost, mirroring the logic of the Born–Haber derivation above but with the sign flipped relative to Eu/Yb: here the special stability lies on the higher, not lower, oxidation state side. Cross-check against the periodic trend: every other lanthanide except Pr and Tb (which show +4 only in solid-state oxides like PrO₂/Tb₄O₇ under forcing conditions, never as clean aqueous Ln⁴⁺(aq)) is exclusively +3 in its common oxide (Ln₂O₃), confirming that Ce's behavior is the exception explained by, and only by, the 4f⁰ target — not a general lanthanide tendency to form +4 oxides.

Limiting cases, scope & common traps

Situation	What happens	Why
n = 0 (La³⁺) or n = 14 (Lu³⁺)	Colorless, diamagnetic, ¹S₀	No unpaired f electrons / no term structure possible; both bracket the series correctly
n = 7 (Gd³⁺)	Nearly pure spin-only magnetism, L = 0	Half-filled shell → maximal S, but M_L sums to zero, killing orbital angular momentum
Ce, Pr, Tb	Can reach +4	Ce³⁺→Ce⁴⁺ reaches 4f⁰; Pr, Tb⁴⁺ approach f⁰/f⁷ analogues, but much less easily (solid-state only)
Eu, Yb, (Sm, Tm weakly)	Can reach +2	Eu²⁺ = f⁷, Yb²⁺ = f¹⁴ are exchange/shell-closure stabilized; Sm²⁺, Tm²⁺ are far weaker analogues
All Ln³⁺ in ligand fields	f–f bands barely shift with ligand	4f shielded by 5s/5p; crystal-field splitting ~10² cm⁻¹, negligible next to spin-orbit/repulsion
Coordination number	6–12, commonly 8–9	Ionic, non-directional bonding; steric packing of hard O/N donors around a large ion, not orbital hybridization limits
As n increases across series (fixed +3)	Ionic radius shrinks smoothly (lanthanide contraction)	Imperfect 4f shielding of increasing nuclear charge — sets up the radius-driven periodicity discussed under the lanthanide contraction

Common trapCommon misconception 2: "The lanthanides show a full spread of oxidation states like the transition metals, from +2 to +4 more or less freely."

Only a handful of lanthanides access +2 or +4 at all, and only because the resulting configuration is f⁰, f⁷, or f¹⁴ — Ce(IV), Eu(II), Yb(II), and (much more weakly) Sm(II), Tm(II), Pr(IV), Tb(IV). Every other lanthanide is essentially locked into +3 because there is no configuration-based stabilization to offset the unfavorable ionization/lattice-energy balance; this is qualitatively different from the d-block, where entire families of stable oxidation states span a genuine continuous range (e.g., Mn from +2 to +7) because d-orbitals participate directly in covalent bonding and are stabilized by ligand-field effects that vary continuously with ligand and geometry.

The chemist behind it

Origins · 1 minGeorges Urbain and how this was discovered

Georges Urbain

1872–1938 · French chemist, rare-earth separation and spectroscopy, Paris

Urbain spent decades on the brutally repetitive fractional crystallizations needed to separate chemically near-identical lanthanides from mixed ore concentrates — the direct experimental consequence of the very shielding argument this article develops: with no orbital-driven chemical differentiation available, only tiny, gradual ionic-radius and solubility differences (the lanthanide contraction in action) allow separation at all. He is credited with isolating and correctly characterizing lutetium (1907), settling a priority dispute with Carl Auer von Welsbach, and his optical/emission spectroscopy of purified rare-earth salts provided some of the sharpest early evidence that these ions carry characteristic, largely ligand-independent line spectra — a direct precursor to the term-symbol picture used throughout modern lanthanide spectroscopy.

Urbain's career is a useful correction to any impression that lanthanide chemistry is "boring" because it's all +3: the entire nineteenth- and early twentieth-century struggle to even isolate individual lanthanides — some fifteen elements initially lumped together as "didymium," "erbia," and other composite mixtures — was a direct consequence of how chemically similar the shielded 4f electrons make these ions. The eventual clean separation techniques (ion-exchange chromatography, developed decisively during the Manhattan Project era for fission-product lanthanides and adapted for stable rare-earth purification) exploit exactly the small, monotonic ionic-radius trend across the series, the same trend responsible for coordination-number and hydration-number changes down the row.

Origins · 1 minhow this was discovered

History

- 1794 — Johan Gadolin identifies "yttria," the first rare-earth mixture, from a Swedish mineral (ytterbite/gadolinite).

- 1839–1843 — Carl Gustaf Mosander separates "cerium," "lanthanum," and "didymium" from early rare-earth mixtures, foreshadowing decades of fractional-crystallization separations.

- 1885 — Carl Auer von Welsbach splits "didymium" into praseodymium and neodymium.

- 1907 — Georges Urbain and (independently) Carl Auer von Welsbach isolate lutetium, completing the identification of the stable lanthanide series.

- 1937–1950s — Quantum-mechanical Russell–Saunders term-symbol treatments (Racah, Judd–Ofelt theory in the late 1950s/1962) formalize why f–f transitions are weak yet observable, completing the electronic-structure picture used today.

Why is +3 so overwhelmingly dominant across the whole lanthanide series?

The sum of the first three ionization energies stays roughly flat across the series (about 3400–3900 kJ/mol), while forming a small, highly charged Ln³⁺ ion releases a very large lattice or hydration energy — large enough, for almost every lanthanide, to more than pay for three ionizations. Reaching +2 means giving up that large lattice-energy payoff in exchange for avoiding one ionization, and reaching +4 means paying for a very costly fourth ionization; both are unfavorable unless the resulting fⁿ configuration (f⁰, f⁷, or f¹⁴) is specially stabilized by exchange energy or shell closure, which happens only for a few elements (Ce, Eu, Yb prominently; Pr, Sm, Tb, Tm weakly).

Why do lanthanide complexes have such high coordination numbers compared to 3d transition-metal complexes?

Lanthanide bonding to ligands is overwhelmingly ionic/electrostatic rather than covalent, because the 4f orbitals are shielded and do not project outward to overlap with ligand orbitals the way 3d orbitals do. Bonding is therefore governed mainly by electrostatics and packing geometry around a large ion (ionic radii around 0.86–1.16 Å for Ln³⁺, larger than typical 3d M²⁺/M³⁺ radii), not by orbital hybridization schemes that cap coordination number at 4 or 6. As a result, coordination numbers of 8, 9, and even up to 12 (as in some nitrate complexes) are common and often fluxional in solution.

Are f–f transitions really "forbidden," and if so, why can we see color at all?

Yes — f–f transitions are Laporte (parity) forbidden because both initial and final states have the same (even/gerade) parity, and electric dipole transitions require a parity change. They are frequently spin-forbidden too. They become weakly visible because odd-parity vibrations transiently distort the local symmetry away from a perfect inversion center, mixing in a small amount of opposite-parity 5d character (vibronic coupling), and because real coordination environments are rarely perfectly centrosymmetric anyway. The result is molar absorptivities of only 1–10 M⁻¹cm⁻¹ (versus 10³–10⁴ for allowed transitions), which is why lanthanide colors are pale but still visible.

How is lanthanide magnetism fundamentally different from 3d transition-metal magnetism?

For most 3d ions, the ligand field is strong enough to quench orbital angular momentum almost completely, so the simple spin-only formula μ = 2√[S(S+1)] usually works well. For lanthanides the ligand field is a tiny perturbation, so the full atomic angular momentum J survives essentially intact, and the correct formula uses the Landé g-factor: μ = g_J√[J(J+1)]. Only Gd³⁺ (L = 0, pure spin state) happens to reduce to the simple spin-only-looking result; every other lanthanide (e.g. Dy³⁺, Ho³⁺) requires the full J-based Landé treatment, and even that needs crystal-field corrections for close quantitative agreement.

Why can cerium reach +4 so much more easily than any other lanthanide?

Ce³⁺ has configuration [Xe]4f¹; removing that single, relatively loosely bound f electron produces Ce⁴⁺ with configuration [Xe]4f⁰ — an empty shell, which is unusually stable in the same way that noble-gas configurations are stable. This lowers the effective fourth ionization energy dramatically relative to other lanthanides, and the resulting Ce⁴⁺ ion forms an especially robust oxide lattice (CeO₂, fluorite structure). The Ce³⁺/Ce⁴⁺ couple has a standard aqueous potential around +1.72 V, mild enough to be used routinely as an oxidant in analytical redox titrations (cerimetry).

Do lanthanides ever form significantly covalent bonds, or is the ionic picture always right?

The ionic/electrostatic picture is an excellent zeroth-order description and dominant for the vast majority of lanthanide chemistry, especially with hard O and N donors (aqua, carboxylate, polyaminocarboxylate, oxide, halide ligands). However, measurable covalent character appears with soft donors (some Ln–S, Ln–P organometallic bonds) and is systematically somewhat larger for the early lanthanides (larger, more diffuse 5d/6s orbitals available for mixing) than the late ones. This small but real covalency gradient is one of the features that distinguishes lanthanide bonding from the still-more-ionic actinide-like extreme and from the substantially more covalent early actinides.

How does this electronic structure connect to the lanthanide contraction?

Because 4f electrons shield each other poorly, effective nuclear charge felt by outer electrons increases steadily across the series, and the ionic radius of Ln³⁺ shrinks smoothly and monotonically from La³⁺ to Lu³⁺ despite the increasing atomic number. This "lanthanide contraction" is the same imperfect-shielding physics discussed here, applied to explain radius rather than oxidation state or spectra; it is responsible for the near-identical (and sometimes smaller) radii of post-lanthanide 5d transition metals compared to their 4d congeners, a major periodicity effect explored separately.