ambipolar diffusion

In a metal you can move one kind of atom on its own. In an ionic ceramic you cannot: every atom is a charged ion, and if the positive cations tried to march one way while the negative anions stayed put, a huge electric charge would build up almost instantly and slam the door. So in ceramics, cation and anion must move together, in the right ratio to keep every region electrically neutral. That coupled, charge-balanced motion of the ions is ambipolar diffusion.

Here is the mechanism. Suppose you want MgO to transport a unit of 'MgO' from one place to another (as happens when it sinters or reacts). One Mg2+ and one O2- must both migrate. They generally have very different intrinsic mobilities — in most oxides the oxygen ion is far more sluggish than the metal ion, or vice versa. If the faster ion races ahead, it leaves a tiny charge imbalance that sets up an internal electric field; that field pulls the fast ion back and drags the slow ion forward until both drift at the same compromise rate. The net result is an effective, coupled diffusion coefficient — the ambipolar D — that is a blend of the two, and it comes out close to the slower ion's value. The couple can move no faster than its laggard.

This is one of the deepest facts in ceramic kinetics: the rate of sintering, of solid-state reaction, and of oxidation is set not by the fastest ion but by the slowest one along its fastest available path. It reframes every 'how do I speed this up?' question: you must accelerate the rate-limiting ion (often by doping to raise its defect population, or by exploiting a fast grain-boundary path), because speeding the already-fast partner does nothing. The neutrality constraint is absolute — you cannot pump net charge into a growing grain or a reaction layer — so the two species are permanently yoked together.