Cladding fracture behavior is an important consideration, particularly in secondary damage of fuel cladding during service and during handling and storage of discharged fuel. A number of test techniques are available that approximate the stress-state experienced by the cladding for crack initiation and propagation in the axial direction (z) and thus provide a measure of the crack propagation resistance. However, the classical fracture mechanics procedure cannot be applied directly to the thin-walled cladding geometry. Thus, attempts to measure fracture toughness have been influenced not only by material characteristics but also by the technique used to measure it. A large scatter in the reported data exists. Crack propagation resistance in the radial direction (r) is even harder to quantify due to the small wall thickness. We report here on our collaborative round-robin exercise to measure and evaluate fracture toughness in unirradiated tubing at 20 and 300 degrees C, wherein seven laboratories participated in testing samples from the same set of materials. The samples were from RXA and SRA Zircaloy-4 cladding and an aluminum alloy tubing of dimensions same as the cladding. All three tubing materials were precharacterized using standard procedures for tensile property measurements. The K-IC for the aluminum alloy block material, from which the tubing was machined, was measured using standard CT (compact tension) testing. The relative toughness of the three materials is known to vary as aluminum alloy < SRA Zircaloy < RXA Zircaloy. The objective was to assess the various techniques (Pin-Loaded Tension, Vallecitos Embedded Charpy, X-Specimen, Internal Conical Mandrel, Double-Edge Notched Tension and Burst Test) for reproducibility of the results and their ability to discriminate between the material variants. Each laboratory pursued its own specific test technique and methodology of data evaluation under a mutually agreed upon set of common guidelines. Fracture characteristics of the materials from each of these seven techniques were evaluated. All the techniques except the Internal Conical Mandrel (ICM) and the Burst Test (BT) followed the conventional procedure of evaluating J values from load-displacement curves. Values for J were generated using a finite element simulation of crack initiation and propagation in the ICM and the stress intensity factor K-I calculated in the BT The paper includes data from various techniques and a comparative analysis that was performed. We conclude that the appropriate parameters for comparison purposes in these studies are J(0.2) and (dJ/da)(0.2). J(max) is less meaningful because of the extensive plasticity exhibited by the cladding material and the observation that crack extensions were far from comparable from different tests at maximum load. Each testing method was clearly able to distinguish the expected toughness order among the three materials. Reproducibility within each test method was very good compared to the scatter normally expected in fracture toughness testing. J(0.2) values, for SRA Zircaloy-4 at room temperature, fell into two groups; comparison of the toughness values among the various testing methods was surprisingly good, with standard deviations in the range, 5-17 %, although such an agreement was limited to techniques within each group. Reasons for the differences, such as loading at the crack tip, the methods used for measuring crack extension "Delta a," and the procedures adopted for analysis of the data were explored. It i clear that for thin-walled Zircaloy tubing no single value of fracture toughness exists. However, it does appear possible to obtain a useful toughness value that is appropriate for a specific application, if the technique (specimen geometry and local stress-strain conditions) closely models the application.
