Metal Oxidation State, d-Electron Count, and Coordination Number

Electron counting via O.S./dⁿ/C.N. is the daily working language of catalyst design: a chemist screening a new hydroformylation or cross-coupling catalyst first checks whether a proposed intermediate is a plausible 16- or 18-electron species with a sensible d-count, before ever running a reaction. In homogeneous catalysis, square-planar $d^8$ Rh(I) and Pd(II) 16-electron complexes are deliberately chosen because their coordinative unsaturation (C.N. = 4, two open sites relative to 18-electron saturation) permits substrate binding and Oxidative Addition — a step whose feasibility is judged directly from the metal's d-count and available coordination sites. Wilkinson's catalyst, RhCl(PPh₃)₃, is analyzed exactly this way: Rh(I), $d^8$, C.N. 4 (square planar), 16-electron, primed for oxidative addition of H₂. In bioinorganic chemistry, assigning the oxidation state of iron in heme proteins (Fe(II) in deoxyhemoglobin vs. Fe(III) in methemoglobin) determines predicted spin state and O₂-binding affinity via straightforward $d^6$ vs. $d^5$ crystal-field reasoning, directly informing drug design for conditions involving oxygen transport and cytochrome P450 metabolism (where Fe(IV)=O "Compound I" intermediates require careful non-innocent-ligand-style electron accounting). In materials chemistry, mixed-valence oxides (e.g., magnetite Fe₃O₄, containing both Fe²⁺ and Fe³⁺) are classified and their electronic/magnetic behavior predicted using exactly this oxidation-state bookkeeping, extended to solid-state coordination environments. And in the metal-carbonyl and organometallic synthesis, the reasoning above is the everyday first filter used before consulting or predicting structures in the study of Metal Carbonyl Complexes and Structures — checking C.N., dⁿ, and the 18-electron total is the first move any organometallic chemist makes when handed an unfamiliar formula.

:::quiz[What is the oxidation state of Mn in [MnO₄]⁻ (permanganate)?]

• +5
• +6
• +7 ✓
• +4

Ask the tutor: How does the oxidation state and d-count of a metal change during oxidative addition of H₂ to a square-planar d⁸ complex, and why does that make the reaction geometrically favorable?

Consider a transition-metal ion sitting at the center of a set of ligands. Physically, the bonding is a continuum: some M–L bonds are highly polar (nearly ionic, as in a fluoride or oxide ligand) and some are largely covalent (a CO or phosphine ligand donating a lone pair, or an alkyl group sharing an electron pair). Oxidation state assignment forces a choice: pretend every bond is 100% ionic, and assign both electrons of each bond to whichever atom is formally more electronegative. This is exactly analogous to assigning formal charges in organic chemistry, except here the "electronegativity ranking" is fixed by convention (ligand atom wins unless it is a bare metal–metal bond or a genuinely more electropositive fragment). Once the ionic split is made, the metal's formal charge is the oxidation state. The d-electron count is then obtained by remembering that a neutral metal atom in group n has n valence electrons distributed over (n−2) s²d^(n−2) or similar configurations — but for bookkeeping purposes we simply use the periodic table group number as the total valence electron count of the neutral atom, and subtract the oxidation state to account for electrons formally removed by ionic bond assignment. Coordination number is a purely geometric/topological count: how many donor atoms surround the metal in the primary coordination sphere. For simple monodentate ligands, C.N. equals the number of ligands. For polyhapto ligands like η⁵-C₅H₅⁻ or η⁶-C₆H₆, the entire ring is one ligand but effectively occupies several coordination sites — conventions differ on whether to count C.N. as "5" for cyclopentadienyl or as "1" (treating the whole ring as a single locant). Both conventions are used in the literature; what must stay internally consistent is the electron count that goes with each choice.

:::svg[Ionic electron counting for Fe(CO)₅ vs. ionic counting for [Fe(CN)₆]⁴⁻: same recipe, different ligands]

::: The figure makes the essential point vivid: two very different ligand environments, neutral π-acceptor carbonyls versus anionic cyanides, both resolve to an 18-electron Fe center, but via completely different oxidation states and d-counts (Fe⁰, d⁸ versus Fe²⁺, d⁶). The coordination geometry differs too (trigonal bipyramidal vs. octahedral) because C.N. changes from 5 to 6. Nothing here is arbitrary: it all follows from correctly classifying each ligand's donor type and charge before doing the arithmetic.

The entire scheme rests on three linked definitions. First, oxidation state via the ionic convention:

where: the sum runs only over X-type (anionic, covalent-radical) ligands, each assigned its "closed-shell" ionic charge (Cl⁻ = −1, O²⁻ = −2, CH₃⁻ = −1, H⁻ = −1); L-type (neutral two-electron donors like CO, PR₃, NH₃) and Z-type (acceptor) ligands contribute zero to this sum because they are assigned no formal charge transfer.

Reading: this equation says the metal soaks up whatever charge is left over after every ligand has been assigned its "natural" ionic charge. It is a conservation law for formal charge, not a statement about where electron density actually sits. Second, the d-electron count:

where: $N_g$ is the periodic table group number of the metal (using the convention that Sc–Zn are groups 3–12, so Fe is group 8, Co is group 9, etc.), and O.S. is the oxidation state just computed.

Reading: removing O.S. electrons (as positive charge) from a neutral atom with $N_g$ valence electrons leaves $d^n$ electrons in the d-manifold; this works because for transition metals in oxidation states ≥ 0 all valence electrons removed upon ionization come from s before d, but by O.S. ≥ +1 the atom is treated as purely $d^{N_g - \text{O.S.}}$, with no separate s electrons tracked — a convention, not a claim about atomic spectroscopy. Third, the total valence electron count (TVEC), the number the Ionic and Covalent Electron Counting Methods both must reproduce identically:

where: $e_i$ is the number of electrons donated by ligand $i$ under the ionic convention (X-type ligands donate 2, as X⁻; L-type donate 2; Z-type donate 0), and $d^n$ is computed as above.

Reading: this is the 18-electron rule's raw ingredient. TVEC = 18 for a saturated, "electronically happy" 18-electron complex; TVEC < 18 signals coordinative unsaturation, and TVEC > 18 is rare and usually signals a very electron-rich late-metal complex or an error in ligand classification.

Example 1 (Foundational): [Co(NH₃)₆]³⁺, a classic Werner complex. Ammonia is a neutral, L-type, two-electron donor (lone pair on N; no formal charge). The overall complex charge is +3, and there are no X-type ligands to absorb any of it (NH₃ contributes zero to the O.S. sum).

Cobalt is group 9, so

Coordination number: six NH₃ ligands, each monodentate, so C.N. = 6, and the geometry is octahedral (confirmed experimentally, and rationalized by $d^6$ crystal-field stabilization in the low-spin case, which this famously diamagnetic complex adopts). Check via TVEC: $d^6 + 6\times 2(\text{NH}_3) = 6+12 = 18$. The complex is an 18-electron, coordinatively saturated, substitutionally inert (low-spin $d^6$, large CFSE) species — consistent with its historical role as one of the first kinetically inert complexes Werner used to establish octahedral coordination chemistry.

Example 2 (Applied): CpFe(CO)₂Cl, a mixed-ligand piano-stool complex relevant to catalyst design. Ligand inventory: one η⁵-C₅H₅ (cyclopentadienyl, treated as Cp⁻, an X₅-type ligand — formally anionic, donating 6 electrons under the ionic method: X-type but polyhapto, contributing 5 covalent/ionic single-bond equivalents worth 6 electrons total via the standard $6e^-$, $L_2X_3$-equivalent counting, though the simplest bookkeeping treats Cp⁻ as a 6-electron, charge $-1$ donor); two CO (neutral L-type, 2e⁻ each); one Cl (anionic X-type, charge −1, 2e⁻ donor). Overall complex charge = 0 (neutral, isolable compound). Oxidation state:

Iron is group 8:

Coordination number: here conventions genuinely diverge. If Cp is counted as occupying one coordination site (common shorthand for "piano-stool" geometry counting), C.N. = 1 (Cp) + 2 (CO) + 1 (Cl) = 4, giving a pseudo-tetrahedral "three-legged piano stool." If each Cp carbon is counted individually, C.N. = 5 + 2 + 1 = 8, which is almost never used in practice for geometry description (though it matters for steric cone-angle style discussions). The chemically standard description is the piano-stool picture: C.N. = 4 (treating Cp as a single polyhapto ligand occupying one leg-equivalent locus, but really representing 3 legs + Cp cap), i.e., a "three-legged piano stool" with Cp as the seat. TVEC check: $d^6 + 6(\text{Cp}) + 2(2)(\text{CO}) + 2(\text{Cl}) = 6+6+4+2 = 18$. This confirms the complex is 18-electron saturated, rationalizing its isolability as a stable, air-stable-ish (kinetically robust) 18-electron piano-stool complex widely used as a synthetic precursor in organometallic catalysis (e.g., for ligand substitution chemistry at Fe).

Example 3 (Exam/competition level): [Mn(CO)₅]⁻ versus Mn₂(CO)₁₀ — metal–metal bonds and the "extra" electron. Consider [Mn(CO)₅]⁻ first. Five CO ligands, all L-type, neutral, 2e⁻ each; overall charge −1.

Manganese is group 7:

C.N. = 5 (five CO, monodentate), and TVEC = $8 + 5(2) = 18$ — an 18-electron, closed-shell anion, isoelectronic with Fe(CO)₅ (which is neutral, $d^8$, C.N. 5) — both trigonal bipyramidal $d^8$ ML₅ fragments, exactly the pattern predicted by isoelectronic/isolobal reasoning. Now consider the dimer Mn₂(CO)₁₀, which contains a genuine Mn–Mn single bond. Each Mn center is bonded to five CO (all L-type) and to the other Mn atom. The subtlety: a metal–metal bond is neither X-type nor L-type in the usual ligand sense — it is treated specially. In the ionic method, homonuclear M–M bonds are not assigned any formal charge transfer (neither Mn atom is "more electronegative" than the other), so the bond contributes zero to either metal's oxidation state, but contributes one electron to each metal's count (a covalent single bond, one electron from each partner, exactly like the covalent treatment of any homolytic single bond).

TVEC per Mn center: $d^7 + 5(2)_{\text{CO}} + 1_{(\text{M–M bond, one electron})} = 7 + 10 + 1 = 18$. C.N. per Mn: 5 (CO) + 1 (the other Mn) = 6, octahedral-ish local geometry at each metal (five CO plus one Mn–Mn bond arranged roughly octahedrally). This example is the classic proof that homolytic M–M bond cleavage (Mn₂(CO)₁₀ → 2 •Mn(CO)₅) generates two 17-electron $d^7$ radicals, each one electron short of 18 — precisely the reactive fragments invoked in radical-chain mechanisms for metal-carbonyl substitution, and a direct segue into how Oxidative Addition and related bond-activation steps are analyzed by the same electron-counting logic.

The bookkeeping formalized here traces its lineage to Alfred Werner, whose coordination theory (1893) first proposed that a metal ion has a fixed "primary valence" (oxidation state, in modern language) and a separate "secondary valence" (coordination number), resolving decades of confusion about compounds like CoCl₃·6NH₃. Werner's insight — that these are two independent numbers describing the same metal center — is exactly the conceptual split this article formalizes rigorously with modern electron-counting rules. His careful resolution of the optical isomers of [Co(en)₂Cl₂]⁺-type complexes provided the experimental proof that octahedral geometry, not some alternative hexagonal-planar arrangement, was correct — an early triumph of reasoning from coordination number to structure.