nonstoichiometry

Chemistry teaches you that compounds have fixed, whole-number formulas: water is exactly H2O, iron(II) oxide is FeO. But press on a real sample of 'FeO' and you find it is never quite FeO — its true composition hovers around Fe0.95O, with measurably less iron than oxygen. Compounds whose composition is not a simple whole-number ratio, and which can vary continuously over a range, are called nonstoichiometric. They quietly break the law of definite proportions that beginning chemistry holds sacred.

How can a crystal lose some of its iron and still hold together? The key is that the transition metal can adopt more than one oxidation state. In iron(II) oxide, for every Fe2+ that is missing, the crystal compensates by converting two neighbouring Fe2+ ions into Fe3+. The total positive charge still balances the oxide ions, so the crystal stays neutral — but it now has cation vacancies and a mixture of Fe2+ and Fe3+ in place of pure Fe2+. The formula drifts away from 1:1 because the iron sublattice has holes in it. This is only possible for elements with accessible variable oxidation states, which is why nonstoichiometry is overwhelmingly a phenomenon of transition-metal and lanthanide compounds (oxides, sulfides, hydrides) and almost never of, say, sodium chloride.

Nonstoichiometry is not a curiosity; it is the engine of much functional materials chemistry. The mixed oxidation states and mobile vacancies make these compounds electronic and ionic conductors, catalysts, and the active components of batteries (lithium slides in and out of nonstoichiometric LixCoO2 as the cell charges) and of high-temperature superconductors (whose oxygen content is deliberately made nonstoichiometric). The honest framing is that the textbook law of definite proportions is an excellent rule for molecular and simple ionic compounds but a genuine approximation for extended solids of variable-valence elements.