The 18-electron rule and organometallic bookkeeping

Main-group chemistry runs on the octet rule: fill an s and three p orbitals, get 8 electrons, stability follows. Transition metals have an additional five d orbitals to worry about, and the corresponding preferred count is 8 + 10 = 18. The 18-electron rule captures a huge fraction of organometallic reactivity: complexes with 18 valence electrons are stable and isolable, complexes with fewer are electron-deficient and hungry to add ligands, and complexes with more (rare) are electron-rich and prone to lose ligands. Learning to count valence electrons on a metal center takes 10 minutes and permanently upgrades how you read a metal complex.

Assume familiarity with d-orbital splitting from crystal field theory.

What counts as a valence electron?

For an organometallic complex, the valence-electron count includes:

1. d electrons on the metal, based on its oxidation state.
2. σ-donor electrons from each ligand — a pair of electrons donated from a lone pair on the ligand into an empty d orbital on the metal.
3. π-donor or π-acceptor electrons from each ligand where applicable — the ligand’s π-bonding or π-backbonding electrons.

Sum those three contributions, get the valence count, compare to 18. The three most useful ligand-counting conventions:

• σ-donor only ligands (2 electrons each): PR₃, NR₃, H₂O, NH₃, CO (contributes 2 σ-donor electrons).
• η² π-donor ligands (2 electrons): alkenes bound side-on, alkynes bound side-on.
• η³, η⁴, η⁵, η⁶ π-ligands (3, 4, 5, 6 electrons respectively): allyl (η³, 3e⁻), diene (η⁴, 4e⁻), cyclopentadienyl (η⁵, 5e⁻), arene (η⁶, 6e⁻).
• X-type ligands (1 electron each, in neutral method): H, halides, alkyls, alkoxides.

Two counting conventions dominate. The neutral (covalent) method counts each ligand contribution as if the ligand were neutral (so Cl contributes 1 e⁻ as a covalent bond, Cp contributes 5 e⁻ as a neutral radical). The ionic method counts the metal in its formal oxidation state and each ligand as an anion (Cl⁻ contributes 2 e⁻ as a σ-donor; Cp⁻ contributes 6 e⁻). Both give the same total. Crabtree Ch. 2 discusses when to use each; neutral is easier for σ-only complexes, ionic for anionic-ligand-heavy complexes.

The full electron-count checklist

For a complex ML_n with charge q, the neutral-method count is:

$\text{valence electrons} = (\text{group number of M}) + \sum_i (\text{electrons from ligand } i) - q$

Example: Cr(CO)₆. Cr is group 6 (6 d electrons in the neutral atom); each CO is a 2-electron σ-donor. Total = 6 + 6(2) − 0 = 18. Stable, isolable — the archetypal 18-electron complex, and the reason Cr(CO)₆ is the standard reference material in organometallic teaching.

Example: [Fe(CO)₄]²⁻. Iron is group 8 (8 d electrons); each CO donates 2 electrons; the −2 charge adds 2 more electrons to the total (because we’re counting from an anion). Total = 8 + 4(2) − (−2) = 8 + 8 + 2 = 18. Also 18. Stable as the sodium salt (Collman’s reagent).

Example: Cp₂Fe (ferrocene). Fe is group 8 (8 d electrons); each Cp (η⁵) donates 5 electrons in neutral counting. Total = 8 + 2(5) = 18. Sandwich-stable, aromatic, discovered in 1951 and one of the launch events of modern organometallic chemistry.

When the rule holds and when it fails

The rule holds cleanly for most 4d and 5d organometallic complexes (Rh, Ir, Pd, Pt, Ru, Os), which have larger Δ_o splittings and strongly prefer to fill their bonding orbitals. It also holds for most 3d complexes with strong-field π-acceptor ligands (CO, CN⁻, phosphine, arene).

The rule fails routinely for:

• d⁸ square-planar complexes (Rh(I), Ir(I), Pd(II), Pt(II), Ni(II) with strong-field ligands) — 16-electron count, not 18. The four in-plane σ-bonds fill four bonding orbitals; the fifth orbital (dz²) is high-lying and empty. Square-planar 16-electron complexes are stable because the dz² is empty — that vacancy is the site where oxidative addition happens.
• First-row metals with weak-field ligands — [Fe(H₂O)₆]²⁺ is d⁶ octahedral with 6+12 = 18 counting, but it’s kinetically and thermodynamically unstable because the weak-field H₂O ligands don’t lower the energy of the t₂g enough to overcome pairing.
• Cp complexes with d²-d³ counts — Cp₂TiCl₂ (d⁰ Ti(IV)) has 16 e⁻ and is quite stable; Cp₂Ti (d²) is 14 e⁻ and reactive.

The 16- and 18-electron pair explains a huge fraction of catalytic cycles: 16-electron complex + substrate → 18-electron intermediate → product + 16-electron complex. This is the Green-Davies-Mingos “16/18 rule” cycle at the core of Wilkinson’s-, Grubbs-, and countless other named catalysts.

Worked example — Wilkinson’s catalyst counting

Electron-count and reactivity predictions

Once you count, three predictions follow:

• 16-electron complexes react with donor ligands to give 18-electron adducts. Oxidative addition of H₂, HX, R-X, alkenes.
• 18-electron complexes are coordinatively saturated. They react by ligand dissociation — one ligand leaves to open a site, then a new ligand adds. The dissociation is the slow step; look for a labile ligand (weak field, high trans-influence position).
• 20-electron complexes are rare but not forbidden. Cp₂Ni is 20 e⁻ (d¹⁰ Ni(II) plus two Cp ligands giving 10 e⁻); the “extra” two electrons occupy an antibonding orbital, making the complex somewhat destabilized but still isolable.

For 4d and 5d catalysts, sticking to 16- and 18-electron intermediates during a cycle is a good check. If your proposed mechanism has a 14-electron intermediate, you likely missed a ligand or forgot a resonance.

Counting tricky ligands

• Bridging ligands (μ): a μ-CO between two metals donates 2 electrons total, split evenly (1 e⁻ to each metal in neutral counting). A μ-H donates 2 electrons total.
• Metal-metal bonds: each metal contributes 1 electron to the M-M single bond, so a metal in a bonded dimer has 1 more electron in its count than the “isolated” analog would.
• η² alkenes vs σ-bound alkyls: an η² alkene ligand is a 2-electron donor (the π-bond); an alkyl (σ-bound) is a 1-electron donor. If a ligand’s binding mode is ambiguous, X-ray structure or IR can resolve it.
• Hydrides (H⁻ or H·): in neutral counting, H is 1 e⁻; in ionic counting, H⁻ is 2 e⁻. Both work; be consistent throughout the count.

Miessler Ch. 13 has a canonical table of ligand contributions in both counting methods; keep it open when working through unfamiliar complexes.

Practice problems

Where to go next

• The Nernst equation calculator computes reduction potentials, useful for oxidative-addition energetics.
• The next inorganic topic building on electron counting is the catalytic cycle: how 16- and 18-electron intermediates chain into a mechanism (Crabtree Ch. 9-16 covers cross-coupling, olefin metathesis, hydrogenation).
• The polyatomic ions cheatsheet helps when assigning charges to complex ligands like acetate, nitrate, oxalate.

The 18-electron rule is not a law — it is a heuristic that works because bonding orbitals in an octahedral (or pseudo-octahedral) transition-metal complex hold 9 pairs. Complexes that satisfy the count exhibit filled bonding levels and empty antibonding levels — thermodynamically stable and kinetically inert. Complexes that miss are either hungry (< 18) or overfilled (> 18) and react accordingly. Learning to count is the entry ticket to organometallic chemistry.