Hall–Petch Relation

In 1951 E. O. Hall published measurements on mild steel showing that the lower yield point rose in proportion to the inverse square root of the ferrite grain size. Two years later N. J. Petch, working on cleavage fracture, found the same functional form. The relation has carried both names ever since, and it is the reason grain refinement occupies a special place in metallurgy: it is the only strengthening mechanism that raises strength and toughness at the same time. Solid solution, precipitation and cold work all buy strength by paying in ductility. Refining the grain buys both.

The mechanism is a traffic jam. Plastic flow is dislocations moving on slip planes, and a grain boundary is a discontinuity in crystal orientation that the slip plane cannot cross. Dislocations generated inside a grain glide until they reach the boundary and stop, piling up behind the leader. The pile-up concentrates stress at its tip — and the more dislocations in it, the greater the concentration — until the stress is enough to activate a source in the next grain over and carry the deformation onward. A large grain holds a long pile-up and concentrates stress efficiently, so it yields easily. A small grain holds a short one, and the applied stress must be higher to do the same job. Working the pile-up geometry through gives the stress concentration as proportional to the square root of the pile-up length, which is the grain diameter, and the −1/2 exponent falls out.

The exponent is a fit, not a theorem, and it inverts. Below roughly 20 nm of grain diameter the relation reverses: finer grains make a weaker metal, not a stronger one. This is the inverse Hall–Petch effect, and it is not a measurement artefact. A grain 10 nm across simply cannot contain a dislocation pile-up — there is not enough room, and the elastic field of one dislocation reaches the boundary on both sides. When the pile-up mechanism has nowhere to operate, deformation transfers to whatever else is available, which in nanocrystalline material is grain-boundary sliding and diffusion along the boundaries. Both get easier as grains get finer, because there is more boundary. The strength therefore peaks somewhere around 10 to 30 nm depending on the metal, and falls on either side. Any calculation this page returns for a grain size in the nanometre range is pointing the wrong way, and the page says so.

Two things about the constants deserve care. σ₀, the friction stress, is the yield strength the metal would have with no boundaries at all — a single crystal's lattice resistance. It is not a property of the element on its own: it carries solid-solution content, temperature and strain rate with it, so a σ₀ fitted at room temperature has no place in a hot-working calculation. k_y, the slope, describes how hard it is to transmit slip across a boundary. It rises with carbon content in steel because interstitial atoms segregate to boundaries and lock the sources there, and it is measurably different depending on whether you defined yielding as the upper yield point, the lower yield point or the 0.2 % proof stress. A k_y is only usable against the same definition it was measured with.

Two unit traps, both of which produce plausible answers. Grain size wants micrometres, and entering millimetres puts the strengthening term off by a factor of √1000 ≈ 31.6. And k_y is quoted variously in MPa·m^(1/2), MPa·mm^(1/2) and MPa·µm^(1/2) — again a factor of √1000 between each step. This site types k_y with the fracture toughness unit type, which will look strange on this page. The reason is dimensional and nothing else: k_y has units of Pa·√m, a stress intensity factor has units of Pa·√m, and MPa·√m converts to ksi·√in by identical arithmetic in both cases. The picker's label is wrong; the units are right. A Hall–Petch slope and a fracture toughness are unrelated quantities that happen to share a dimension, and that is the whole of the connection.

In practice the relation is what the entire discipline of thermomechanical controlled processing exists to exploit. Controlled rolling of microalloyed steel — niobium, titanium and vanadium in tenths of a percent — works by using carbonitride particles to pin austenite grain boundaries during hot rolling, so the austenite stays fine, recrystallises repeatedly, and transforms to a ferrite grain of a few micrometres instead of tens. That is where modern pipeline steel gets both its strength and its low-temperature toughness, and the arithmetic on this page is the reason it was worth the trouble.