but a well-studied exception is polycrystalline Ni3Al, which suffers from intergranular weakness. It has been found, however, that small amounts of boron will segregate to the grain boundaries, strengthen these interfaces, and thus restore a level of ductility commensurate with that of the crystalline state. The effect of boron on the ductility and strength of polycrystalline Ni3Al (24 atomic percent aluminum) is shown in Figure 29.54 There is increasing evidence that the beneficial influence of boron is not one of simply displacing a harmful impurity (such as sulfur) from the grain boundaries but rather is a matter of enhancing the electronic bonding across the interface. For this reason the role of boron seems to be quite specific: it works superbly only on the nickel-rich side of Ni3Al stoichiometry, but not on the aluminum-rich side, and it is not necessary for grain boundary strengthening in numerous other L12 alloys. In fact, the interfacial strength and hence intergranular ductility have been shown to increase with the valence difference between the two constituent atoms in a wide variety of L12-type A3B phases.55 This idea has been further substantiated by corresponding studies of ternary additions to Ni3Al alloys.56
Many L12 intermetallic phases exhibit anomalous strengthening with increasing test temperature, often up to nearly the disordering temperature; Figure 30 shows several examples.57 This strange mechanical behavior raises lively questions regarding both its origin and its potential use for high-temperature service. It should be emphasized, however, that the anomalous temperature dependence may be displayed to only a minor degree, or even not at all, by some L12 intermetallic phases. Nevertheless, we can expect that diffusion-dependent creep will be significantly retarded because of the relatively low atomic mobility typical of ordered lattices, as discussed below.