When a metal is plastically deformed (cold worked), its dislocation density increases dramatically — from roughly 10^6 cm/cm^3 in an annealed metal to 10^10 or higher after heavy deformation. These dislocations interact, tangle, and pin one another, raising the stress required for further deformation. This is work hardening (strain hardening), and it is why a bent paperclip becomes stiffer at the bend. Cold working simultaneously increases yield strength and hardness while decreasing ductility and electrical conductivity. The increased dislocation density also stores elastic strain energy in the lattice, making the material thermodynamically unstable. Recovery is the first stage of restoration: when a cold-worked metal is heated to a moderate temperature, dislocation rearrangement occurs — dislocations of opposite sign annihilate, and remaining dislocations organize into lower-energy configurations (subgrain boundaries or cell walls) through polygonization. Recovery reduces residual stresses and partially restores electrical conductivity without significantly changing grain structure or the overall strength.
Plot yield strength, ductility, and electrical conductivity versus percent cold work to see the trade-offs quantitatively. Then overlay the effects of annealing temperature to distinguish the recovery, recrystallization, and grain growth stages. Examining TEM micrographs of dislocation tangles (cold-worked) versus organized cell structures (recovered) makes the microstructural changes concrete.
From your study of plastic deformation mechanisms, you know that plastic flow in metals occurs by dislocation motion — dislocations glide through the crystal lattice on slip planes, and the cumulative offset of millions of glide events is what we observe as permanent shape change. You also know from strengthening mechanisms that anything that obstructs dislocation glide raises the yield strength. Work hardening is the most dramatic demonstration of this principle: every dislocation that moves and gets tangled creates new obstacles for every subsequent dislocation. The material hardens itself as it deforms.
The quantitative story is striking. A well-annealed copper crystal might contain 10⁶ dislocation segments per square centimeter of cross-section. After heavy cold rolling or drawing, that density can reach 10¹⁰–10¹² cm/cm². The spacing between dislocations shrinks from microns to nanometers. At that density, dislocations encounter each other before they can travel far: they form dislocation tangles, forest dislocations that must be cut through (requiring extra stress), and Lomer-Cottrell locks — sessile configurations that cannot glide in any direction. Each of these interactions raises the flow stress. This is why a wire becomes stiffer and harder to bend the more you bend it back and forth, and why a copper pipe that has been extensively worked in manufacturing requires less force to start deforming but fails after less total deformation than annealed copper.
The trade-off shows up directly in the stress-strain curve. As a metal is cold-worked, its yield strength climbs and its uniform elongation drops. You can plot these against percent cold work to see the exchange: at 30% cold work, copper might have twice the yield strength of its annealed state but only half the elongation before necking. This trade-off is exploited commercially — spring temper, half-hard, and full-hard designations for copper and brass specify how much cold work has been applied to achieve a target strength level. When ductility is the priority (for deep drawing sheet metal into cups, for instance), the material must be in the annealed condition.
The stored elastic strain energy from all those tangled dislocations makes cold-worked metal thermodynamically unstable — it wants to reduce its energy. Recovery is the first step when you supply thermal activation by heating. At temperatures typically 30–50% of the melting point (in Kelvin), dislocations of opposite sign can annihilate each other, and remaining dislocations rearrange by climb and glide into ordered low-energy configurations called subgrain boundaries in a process called polygonization. The grain structure itself is unchanged — you are rearranging defects within existing grains. The effects are modest but practically important: residual stresses (from the deformation process) are substantially reduced, and electrical conductivity is largely restored. Strength drops only slightly. Recovery is used industrially when you need to relieve process-induced stresses (preventing stress-corrosion cracking in copper alloys, for example) without sacrificing the work-hardened strength that was the point of the cold-working process. The next stage — recrystallization — nucleates entirely new strain-free grains and nearly eliminates all the work hardening, which is why the distinction matters.