A puzzle no one could draw
By the 1890s chemists had mastered ordinary valence: carbon makes four bonds, oxygen two, hydrogen one. The earlier rungs of this ladder gave you that whole machinery — fixed combining powers, tidy dot structures, octets. Then a family of cobalt compounds sat down on the table and refused to behave. Heat cobalt(III) chloride with ammonia and you can isolate several distinct, brightly coloured solids, all built from the same three ingredients in different amounts: CoCl3·6NH3 (orange), CoCl3·5NH3 (purple), CoCl3·4NH3 (green or violet). Cobalt's valence was supposedly satisfied at three chlorides. So why does it keep swallowing ammonia, a molecule whose own valences are already full?
The reigning explanation was the 'chain theory' of Sophus Mads Jørgensen, a brilliant experimentalist who pictured the extra ammonias hooked together in chains, much like carbon chains: -NH3-NH3-NH3- strings dangling off the cobalt, with chlorides clipped onto the ends. It was an honest attempt to reuse the only bonding picture available. But the chains had to be drawn differently for each compound, and an awkward experiment kept refusing to fit them.
The awkward experiment was a silver nitrate titration. Add Ag+ to a fresh solution of each compound and the chlorides that are loosely held — free to swim away as ions — instantly precipitate as white AgCl, while any chloride held tightly to the cobalt stays put and is not counted. The numbers came out startling. In CoCl3·6NH3, all three chlorides precipitate. In CoCl3·5NH3, only two do. In CoCl3·4NH3, only one. The third member loses a chloride at every step, exactly in step with losing an ammonia. Any honest chain theory predicting three identical chloride 'ends' could not explain why one chloride at a time quietly switched from ionic to silent.
Werner's leap: two kinds of valence
In 1893 a 26-year-old named Alfred Werner woke (by his own account) with the answer fully formed, and spent the rest of his life proving it. His [[werner-coordination-theory|coordination theory]] broke a single rigid 'valence' into two independent kinds. The first he called primary valence: the old, familiar combining power that we now read as the metal's charge — for cobalt(III) it is simply 3, satisfied by three units of negative charge somewhere in the compound. The second, and this was the genuinely new idea, he called secondary valence: a fixed number of groups the metal binds directly and holds close, regardless of charge. For these cobalt compounds the secondary valence is 6.
Picture the cobalt at the centre of a sphere with exactly six positions on its surface — Werner correctly guessed these six point to the corners of an octahedron. Those six positions are always filled. Ammonia molecules can fill them; so can chloride ions. What changes from compound to compound is how the six slots are shared between ammonia and chloride. A chloride sitting in one of those six inner positions is held tightly and does not precipitate with silver; a chloride that has been pushed out beyond the sphere floats free as an ordinary ion and falls out instantly as AgCl. That single distinction — inside the sphere versus outside it — dissolves the whole puzzle.
compound inside the sphere (held) outside (free ions) Cl- freed -------- ------------------------ ------------------- --------- CoCl3.6NH3 [Co(NH3)6]3+ 3 Cl- 3 CoCl3.5NH3 [Co(NH3)5Cl]2+ 2 Cl- 2 CoCl3.4NH3 [Co(NH3)4Cl2]+ 1 Cl- 1 CoCl3.3NH3 [Co(NH3)3Cl3]0 0 Cl- 0 secondary valence = 6 always; primary valence (Co charge) = 3 always
The modern picture: anatomy of a complex
Werner's two valences map cleanly onto today's vocabulary, and the immediately preceding guide already handed you the bonding mechanism: a complex is a Lewis acid–base adduct, the metal cation accepting lone pairs that the surrounding bases donate. A [[coordination-compound|coordination compound]] is built from four parts. At its core sits the central metal, almost always a cation in a definite oxidation state — Werner's primary valence, now read as charge. Around it cluster the [[inorg-ligand|ligands]]: the molecules or ions that donate electron pairs. Each ligand reaches in through a [[donor-atom|donor atom]] — the specific atom whose lone pair actually forms the bond (the N of ammonia, the O of water, the Cl of chloride). And the count of donor atoms bonded directly to the metal is the [[inorg-coordination-number|coordination number]] — Werner's secondary valence, made precise.
Werner's 'inside the sphere' has a modern name too: the coordination sphere, the metal plus everything bonded directly to it. We mark it off with square brackets. Everything inside the brackets travels as one rigid unit; everything outside is a free counter-ion. So [Co(NH3)6]Cl3 says cobalt with six ammonias forms one tight package carrying a 3+ charge, balanced by three chlorides loose outside. The charged package inside the brackets, [Co(NH3)6]3+, is the [[complex-ion|complex ion]]. Run the bookkeeping: cobalt is +3, the six neutral ammonias add nothing, so the complex ion is 3+; the three chloride counter-ions outside make the whole compound neutral. The two roles a chloride can play — tight ligand inside or free counter-ion outside — are exactly Werner's two valences wearing modern clothes.
Two honest cautions about the bookkeeping. First, [[oxidation-state|oxidation state]] is a counting device, not a literal charge sitting on the metal. Saying cobalt is '+3' does not mean three whole electrons have physically left it; it is a convention that assigns shared electrons to the more electronegative partner, useful precisely because it is consistent. The real electron density is smeared over the metal and its donor atoms. Second, the metal–ligand bond is genuinely partly covalent. Werner's picture and the simple Lewis picture both lean on an electrostatic, ionic image — and that image is a good first approximation — but later guides on crystal field and ligand field theory will have to add covalent character back in to explain colour and magnetism honestly.
Reading a complex, step by step
Hand someone the formula K3[Fe(CN)6] cold and it looks forbidding. Take it apart with Werner's logic and it falls open in four moves. The same routine works on almost any complex you will meet.
- Find the brackets. Whatever sits inside [ ] is the complex ion — the coordination sphere. Here that is [Fe(CN)6]. The three K+ outside the brackets are free counter-ions; they would precipitate or swap freely, they are not bonded to the iron.
- Count the coordination number. Six cyanide ligands bond to the iron, each through its carbon donor atom, so the coordination number is 6 — the metal sits at the centre of an octahedron, the most common geometry of all.
- Pin down the oxidation state. The whole compound is neutral; three K+ supply +3 outside, so the complex ion must be 3-. Each cyanide carries 1-, six of them give 6-, and the iron must therefore be +3 to land the bracketed charge at 3-. So this is iron(III): Fe3+.
- Picture the geometry. Coordination number 6 means an octahedron: the six donor carbons aim at the corners of an octahedron with iron dead centre. That shape — and the d-electron arrangement of the Fe3+ inside it — is what the crystal field guides will build on to explain why this particular salt is a deep, familiar red.
Coordination number 6 with its octahedron is far and away the most common, but it is not the only [[coordination-geometry|coordination geometry]]. Coordination number 4 appears as either a tetrahedron (common for small or d-electron-rich metals) or a square plane (favoured by metals like Pt2+ and Ni2+ in certain electron counts). Coordination number 2 gives a linear shape, classically for Ag+ in [Ag(NH3)2]+. Higher numbers — 7, 8, 9 — show up among the large lanthanide and actinide ions, which have room for more neighbours. The geometry is not random; it follows from the metal's size, its electron count, and how bulky the ligands are.
Why this idea sits at the centre
It is worth pausing on how much Werner's small move bought us. The instant you accept that a metal carries a secondary valence — a fixed inner cluster of donor groups — a whole continent of chemistry becomes describable. Colour: nearly every richly coloured solution and mineral you have ever seen, from the blue of copper sulfate to the green of nickel salts, is a complex ion absorbing light. Magnetism, catalysis, the way an enzyme grips a metal at its active site, the way a drug like cisplatin docks onto DNA, the pigments that run photosynthesis — all of them are coordination compounds, the same metal-wearing-ligands structure repeated with different parts.
One last honesty note, because the name itself misleads beginners. 'Inorganic' does not mean lifeless, and coordination chemistry is the proof: the iron in your blood, the magnesium at the heart of chlorophyll, the zinc in hundreds of your enzymes are all coordination compounds doing the work of life. Nor does 'inorganic' banish carbon — cyanide, carbon monoxide, and carbon-rich ligands are everywhere in this subject. Werner gave us a frame that holds the whole sprawling middle of chemistry, organic and biological and material alike. From here the ladder zooms in: next, the ligands themselves and the surprising extra grip that chelating ligands provide.