In order for significant strain to occur, dislocations must overcome the obstacles to their motion — the irradiation-induced <a> loops. At low stresses this may happen by the process of dislocation climb, which means irradiation-induced point defects (PDs) diffuse to the dislocation and allow it to move around the obstacle. This is an important creep process, to be covered further in Section 4.6. At high stress or high strain rates, as in a power excursion, or at all practical strain rates out-of-reactor, the disloca­tions can actually interact with the <a> loop defects and remove them from the microstructure. In effect this creates a localized soft area, where addi­tional deformation tends to concentrate: this process is called dislocation channelling. The physical process is illustrated in Fig. 4.21. The long straight

white bands are dislocation channels in which irradiation damage (black areas and black spots) is removed. In zirconium alloys the channels are in the order of 0.01-0.30 pm wide depending on fluence and irradiation tem­perature, and each channel can accommodate large (50-300%) local strains. The channels intersect the surface to cause large protrusions or slip steps there (Adamson, 1968; Sharp, 1972).

In Zircaloy, the dislocation channels tend to form in a very a localized area called a deformation band. For a simple uniaxial tensile test specimen the sequence of formation is illustrated in Fig. 4.22. At point (A) the defor­mation band begins to form and is fully formed at (B). At point (B) a second deformation band forms perpendicular to the first, and the specimen frac­tures at point (D). Because virtually all the strain forms in the deformation band, there is little or no deformation in the rest of the specimen gauge length. A plot of measured strain along the length of a typical specimen is given in Fig. 4.23 . Since little plastic strain occurred outside deformation bands, the true gauge length of the specimen is much shorter than the nomi­nal specimen gauge length. Therefore, specimen geometry greatly influences reported strain values. The effect of test specimen geometry on failure strain is illustrated in Fig. 4.24 where the conventional value of uniform elongation (UE) is plotted against gauge length for different specimen geometries of

482 414 345 276 207 138 69

0.04

4.22 Engineering stress-strain curve for Zircaloy-2 sheet that had been irradiated at 280°C to a neutron fluence of 5 x 1020 n/cm2 and subsequently tested at 300°C. (Source: Reprinted, with permission, from Bement et al. (1965), copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428.)

basically the same material. These data show that strain values developed for use as failure criteria or strain limits are not real material properties, but are strongly influenced by the specimen design used to obtain the data.

For burst tests of irradiated materials at 350°C (623K) (Onimus et al, 2004) careful laser imaging measurements indicate that tubing deformed

homogeneously throughout the gauge section until strain concentrates in the burst region. However, this strain (i. e. the number of channels) is small compared to the burst strain.

A main reason that anisotropic deformation is decreased relative to unir­radiated material (Mahmood et al, 2000) is that the high stresses needed to reach the yield point activate alternate slip systems in irradiated Zircaloy. The primary slip plane in unirradiated Zircaloy is the prism plane, the so-called <1120>(1010) system. As the applied stress becomes high, both the pyramidal and basal planes can become active. Observations of prism plane dislocation channels have been well documented (Adamson et al., 1986; Bell, 1974; Adamson & Bell, 1986; Bourdiliau et al, 2010), but obser­vation of pyramidal and basal channels have also been reported (Bell, 1974; Fregonese et al, 2000; Regnard et al, 2001; Onimus et al, 2004, 2005; Bourdiliau et al., 2010). In fact the CEA group show with considerable data and justification that, for 350°C (623K) testing temperature, basal slip pre­dominates, but that may yet prove to be a function of irradiation and test­ing temperature, testing mode and impurity level (Bourdiliau et al, 2010). Dislocation channelling phenomena themselves and details about which channelling planes predominate are important when modelling crack prop­agation and material response to actual in-reactor loading patterns. Onimus et al. (2005) have made good progress in modelling the phenomena for the CEA conditions. The data is summarized in a ZIRAT 15 Annual Report (Adamson et al, 2010).

Specimen design plays a dual role, influencing ductility through both geometry and stress state. The type of plane stress specimens shown in Fig. 4.22 result in a ‘classical’ deformation band formation. The disloca­tion channels can freely extend from surface to surface. In the plane strain specimens of Fig. 4.25, the channels run into specimen regions where the stress is significantly lower before a free surface is reached, therefore pre­venting formation of a well-developed deformation band. The latter case, constrained plane strain, more realistically represents deformation in most reactor component situations.