p. G. KLEMENS 



mean free path, however, especially if the incident particle is heavy, it is 

 possible to get values of £" up to 10~- or 10~^eV, so that local melting may 

 occur. 



The situation is difTcrent in insulators, where local heating produces an 

 outward flow of lattice waves, whose mean free path / is of the order of 10~' 

 cm in typical cases. Substituting into (3), this would give values for E of the 

 order of 10~^eV for an incident electron, and I0~^ for a proton, so that 

 massive particles can produce local melting. 



While local melting along the track is probably an unusual occurrence in 

 metals and alloys, it is also possible to have spikes of spherical symmetry. 

 Once the energy of a particle has fallen below E^ (say, lO^eV), it does not 

 travel appreciably, and passes most of its energy, either directly or through 

 its progeny, to the lattice by collisions. This energy is distributed instan- 

 taneously within a volume l^, and even in a metal, where it is shared among 

 10^ to 10' atoms, local melting is possible. Such a spike occurs in the region 

 of a displacement cluster, and it was termed 'displacement spike' by 

 Brinkman". 



So far it has been tacitly assumed that all irradiation damage remains 

 locked in the lattice, but this is not invariably so. A displaced atom leaves 

 behind a vacancy and lodges itself in an interstitial position. However, 

 either the vacancy, or the interstitial, or both, may be mobile and recombine 

 by diffusion. Such recombination processes are favoured by the fact that a 

 large fraction of the displaced atoms are displaced with energy only slightly 

 above E^g, and do not travel far before being stopped, so that the resulting 

 interstitialcy-vacancy pair is initially of small separation. 



The mobility of these defects is a function of temperature and of the type 

 of material. In some cases, notably in non-metals, prolonged annealing is 

 necessary to remove the damage, in other cases, including many metals, 

 annealing is very pronounced even at room temperatures. In the latter 

 cases it is necessary to irradiate the specimen at low temperatures to retain 

 all radiation damage : in the case of copper even at liquid helium temperatures. 



There is a rough relationship between ease of annealing and ductility. 

 Plastic deformation may be pictured in terms of the motion of dislocations. 

 This motion will, in general, be impeded by obstacles, including other 

 dislocations, and the motion of dislocations is then no longer conservative, 

 but involves the creation and annihilation of vacancies, or possibly inter- 

 stitials. The details of these intei'action processes are not yet fully under- 

 stood, but it seems that the ability of point defects to diffuse plays a role in 

 the deformation process. Good ductility implies a high mobility of these 

 defects, and this implies easy annealing of at least part of the radiation 

 damage; hence the difference between metals and insulators in their ease 

 of annealing. 



It should be remembered that even in materials which do not anneal at 

 the temperature of irradiation, some self-annealing may occur diunng 

 irradiation because of the temperature increase in the thermal spikes. 



A different type of change occurs in polymers. Here the primary effect is 

 one of the breaking and reforming of bonds between randomly orientated 

 chains. This is an electronic effect, which can be produced by all types of 

 radiation. The effect of this is to produce increased cross-linking, and a 



275 



