Crossing the border into organometallic chemistry
You arrive here already fluent in coordination chemistry. You can read brackets, count donor atoms for a coordination number, pin down the metal's oxidation state by balancing charges, and use crystal and ligand field theory to predict color and magnetism. Almost every complex you have met so far binds the metal through nitrogen, oxygen, or a halide — donor atoms with lone pairs to spare. Now we change one thing, and it changes everything: we bond the metal directly to carbon. A compound with at least one genuine [[metal-carbon-bond|metal-carbon bond]] is called organometallic, and the moment carbon sits in the coordination sphere, the chemistry tilts. The bonds become more covalent, the metals favor low oxidation states, and a whole repertoire of reactions opens up that builds and breaks carbon frameworks on demand.
Why does carbon make such a difference? Carbon is far less electronegative than the oxygen or fluorine donors of classical complexes, so the metal-carbon bond is genuinely covalent — electrons are shared, not handed over. More importantly, many carbon-based ligands carry empty orbitals of just the right shape to drink electron density *back* from the metal, the pi back-donation you will meet in detail in the carbonyl guide. That two-way traffic — ligand gives a lone pair to the metal, metal pushes its d electrons back into the ligand — stabilizes the unusually low oxidation states (zero, even negative) that classical chemistry rarely sees. This is why organometallic chemistry is its own continent, bridging the inorganic and organic worlds.
Classifying the ligands: how many electrons, and how attached
Before you can count electrons you must know what each ligand brings to the table. The first axis of ligand classification is how many electrons a ligand donates. An ammonia, a phosphine, a chloride lone pair, or a carbonyl each typically offer two electrons — these are the bread-and-butter two-electron donors. A bridging or doubly bonded group can offer four; a face-capping one, six. The second axis is [[hapticity|hapticity]], written with the Greek letter eta: it counts how many contiguous atoms of a single ligand actually touch the metal. A chloride bound through its one atom is eta-1; an ethylene molecule lying sideways so both of its carbons touch the metal is eta-2; the flat five-carbon cyclopentadienyl ring sitting face-on, all five carbons engaged, is eta-5.
Carbon ligands span a glorious range. A simple alkyl or aryl group, like the methyl in a metal-CH3 bond, hangs on through one sigma bond. Carbon monoxide, the workhorse of the metal carbonyls you will study next, donates the lone pair on its carbon. A flat aromatic ring like cyclopentadienyl lies face-on and shares the whole pi cloud of its five carbons at once. The single skill that lets you treat all of these on the same footing — the alkyl, the carbonyl, the sideways alkene, the face-on ring — is electron counting, and it is the organometallic chemist's very first move on any new compound.
Counting valence electrons: two conventions, one answer
The [[valence-electron-count|valence electron count]] is the total number of electrons in the metal's valence shell once you add up everything the metal brings plus everything the ligands donate. It governs stability the way the octet does for light main-group atoms, except the magic number here is usually 18 — a filled set of one s, three p, and five d orbitals. That guideline is the [[eighteen-electron-rule|18-electron rule]], and we will lean on it in a moment. But first the count itself. There are two bookkeeping schemes for getting it, and the crucial thing — the thing that trips up every beginner — is that a *correctly applied* count gives the same total either way. The two conventions are the ionic (also called the donor-pair) method and the neutral (also called the covalent or radical) method.
In the ionic method you imagine breaking every metal-ligand bond *heterolytically* — both bonding electrons go to the ligand. That leaves the metal as an ion in a definite oxidation state, so you take its d-electron count (the d-electron count you already master) and add the electrons donated by each ligand treated as a closed-shell anion or neutral: a methyl becomes CH3- giving 2, a chloride is Cl- giving 2, a carbonyl is neutral giving 2, a cyclopentadienyl is Cp- giving 6. In the neutral method you imagine breaking every bond *homolytically* — one electron to each side. The metal keeps its full neutral group-number of valence electrons (iron, group 8, brings 8), and each ligand is counted as a neutral radical: a methyl radical gives 1, a chlorine atom gives 1, a carbonyl gives 2, a neutral Cp radical gives 5. The accounting differs but the books must balance.
Worked example: ferrocene, counted both ways
Watch the two methods agree on ferrocene, Fe(eta-5-Cp)2 — the sandwich molecule that is the headline of the next-but-one guide. By the ionic method: the molecule is neutral and each Cp ring carries a formal -1, so iron is Fe2+, a d6 ion, contributing 6; each Cp- donates 6, and there are two of them, contributing 12. Total: 6 + 12 = 18. By the neutral method: neutral iron is group 8, contributing 8; each neutral Cp radical donates 5, two of them contributing 10. Total: 8 + 10 = 18. Same 18, reached down two different staircases. The trick to never mixing them up: the ionic method needs you to assign an oxidation state and pull the matching d-count; the neutral method ignores oxidation state entirely and just uses the metal's group number. Pick one lane and stay in it for a given count.
Ferrocene Fe(eta5-Cp)2 target: 18 e- IONIC (donor-pair) method NEUTRAL (covalent) method Fe2+ is d6 ............ 6 Fe(0), group 8 ........ 8 2 x Cp- @ 6 e- ....... 12 2 x Cp(.) @ 5 e- ..... 10 ---- ---- total ............... 18 total ................ 18 remember: charge on the whole ion shifts the metal contribution (for a cation, subtract; for an anion, add)
Why the count is your first move
Here is the practical recipe you will run almost reflexively on every new compound.
- Pick a convention — ionic or neutral — and commit to it for the whole calculation.
- Read each ligand's hapticity (the eta value) and look up how many electrons it donates under your chosen convention.
- Get the metal's contribution: its d-electron count in its oxidation state (ionic), or its group number (neutral).
- Fold any net charge on the complex into the metal's contribution — once, and only once.
- Add it all up and compare the total to 18: a value of 18 flags a likely stable, coordinatively saturated species.
What does the answer buy you? A count near 18 says the complex is coordinatively saturated — its bonding orbitals are full, so it tends to be stable and unreactive toward picking up more ligands. A count of 16 or fewer flags an unsaturated species with a vacant site, hungry to bind something. That single reading predicts reactivity. In a catalytic cycle, which this rung builds toward, the metal breathes between counts: it loses a ligand to fall to 16 (opening a site), binds a substrate back to 18, rearranges, and releases product. The whole machine is a choreography of electron counts ticking 18, 16, 18, 16, and reading those numbers is how you follow the dance.
The 18 is a guideline, not a law
Be honest about how far the rule reaches. The 18-electron rule works best for the middle of the d block with strong-field, pi-accepting ligands like carbonyls, where the splitting between metal-based orbitals is large enough that filling exactly the bonding levels really is the favored state — the same large-splitting logic you met in ligand field theory. But there are whole families of stable exceptions. The clearest is the d8 square-planar complexes of the late metals — platinum(II), palladium(II), nickel(II), rhodium(I) — where a famous 16-electron count is perfectly stable, the square-planar exception that lies right at the heart of cross-coupling catalysis. Early transition metals, short on d electrons, often settle below 18; some bulky-ligand complexes do too because there is simply no room for more.
Keep three honesties pinned up as you go. First, oxidation state in the ionic method is a bookkeeping device, exactly as it was back in the redox rung — Fe2+ in ferrocene does not mean the iron really carries a clean double-plus charge; the bond is genuinely covalent and the electrons are shared. Second, the count is a robust *number* even though the orbital story behind it is a simplification; do not over-read it. Third, an 18-electron count predicts thermodynamic and kinetic *robustness*, but as the mechanisms rung stressed, thermodynamic stability and kinetic lability are independent ideas — an 18-electron complex can still react quickly if a low-energy pathway exists. Treat 18 as a strong default expectation, then enjoy the exceptions for what they teach.