grain size

If grains are the tiles of the mosaic, grain size just asks: how big is a typical tile? It is the single most reported number about a microstructure, because so much behaviour depends on it. But a grain is an irregular 3D blob, not a neat sphere, so its size is not one obvious number — it is an average of some measurable length, and you have to say which.

In practice grain size is quantified a few standard ways. You can report a mean grain diameter in micrometres; a mean linear intercept (the average length a random test line spends crossing one grain, from the linear-intercept method); or an ASTM grain-size number G, an integer on a logarithmic scale where bigger G means finer grains. These measures are related but not identical, so a reported grain size should always name its method. A real grain-size figure also comes with a spread — a grain-size distribution — because grains are never all the same size; the size is the average of that distribution.

Grain size matters because it directly tunes strength: the Hall-Petch relation says yield strength rises roughly as 1/sqrt(d), so halving the grain diameter noticeably strengthens a metal, while very fine grains also usually improve toughness. This is why grain refinement is one of the few ways to raise strength and toughness at the same time. Honest caveat: the grain size you read off one flat polished section is a biased estimate of the true 3D size, because a random cut rarely slices a grain through its widest part — stereology corrects for this.