The independence of temperature shown by the strain hardening coefficient of face centred cubic metals appears to be at variance with several other features of this strain hardening that depend on temperature. However, on the forest theory of this hardening, thermal activation is expected to have opposite effects on two underlying aspects. First, it reduces the applied stress required to enable glide dislocations to cut through forest obstacles. Second, it increases the density of the secondary dislocations that form these obstacles so that the forest becomes harder to penetrate.
The problemThe coefficient of strain hardening, d/d, where and are shear stress and strain, respectively, of a Face Centred Cubic (FCC) metal crystal, in stage II, has a value of about /300, where is the shear modulus, observed to be independent of the temperature of straining, at least over a wide low-temperature range, e.g. [1]. It is often supposed that this indicates that in this stage the hardening is athermal, for example, due to long-range elastic interactions between glide dislocations. But the effect is surprising because several other features of such strain hardening show clear temperature dependencies. Thus, the flow stress for a given dislocated state falls with rising temperature, e.g. by about a third for aluminium, from zero to room temperature, and the temperature-dependent part of this stress remains strictly proportional to the total flow stress over a wide range of strains. Again, stage III of the hardening, sets in increasingly early at higher temperatures. From these thermal dependencies it might be expected that, by first straining at a low temperature and then warming up so as to reduce the flow stress, a lower rate of strain hardening would be found at this higher temperature. However, this argument is wrong because it has been shown that the sequence of strain hardened states (i.e. dislocation structures) is unique for each temperature and cannot be obtained from any other temperature (assuming, which we shall do throughout, that the strain rate is constant) [2], although this uniqueness makes the constancy of /300 at various temperatures even more surprising.