Even at low pressures, molecular iodine is an effective stress corrosion cracking (SCC) agent for zircaioy*-'. However, inside a reactor fuel pin, iodine is present mainly as cesium iodide. Also, metal iodides such as Fel-and A1I3 could attain partial pressures in the mtorr range. Cadmium is a fission product which thermodynamics suggests should exist inside a fuel pin in the metallic state. Liquid cadmium is known to cause severe embrittlement of zircaloy. In an inpile situation, however, the

more »

... ladding is in contact with the vapor phase with low metal partial pressure rather than with a condensed phase. Therefore, a study of the stress rupture behavior of zircaloy in the presence of All,(g), Fel2(g), Csl(g), and Cd(g) was undertaken. Tube specimens are internally pressurized with argon in order to produce a biaxial tensile stress state. Rupture lifetimes are measured as a function of stress, temperature, and partial pressure of the active chemical agent, which impinges as a molecular beam on a spot T-1 cm diameter. The fracture mode is also observed by SEM. Compared to control specimens at the same temperature and stress, failure times are greatly reduced by the presence of iron and aluminum iodides. For both iodine^ ) and these iodides, there exists a stress (376 MPa), beyond which "burst-type" failure occurs. Lower stresses produce "pinhole-type" failures. However, fractography of both types of failures shows cleavage characteristics which are very different from the ductile-dimple fracture surface found in the ahsence of a corrosive agent. These three types of stress-rupture failures are shown in Fig. l. All specimens tested under cadmium exhibited burst-type failure and -1-the fracture surface showed brittle-type cleavage features. For cadmium, the rupture times were shorter than those of the control specimens but exhibited more scatter than those in iodine or the metal iodides. Fracture occured randomly inside the cadmium-affected region on the zircaloy specimen, unlike iodine and the iodides, where failure always took place at the center of molecular beam spot. There may be some unknown surface inhomogenity controlling the cadmium embrittlement process which is not present in SCC by iodine or metal iodides. Specimens cut from different tubes gave widely varying rupture times in cadmium, but tests using the same zircaloy tubes gave consistent results. In cesium iodide tests, plots of time-to-failure versus stress did not show any decrease in comparison with control specimens. SEM fractography revealed completely ductile behavior, the same as that without a corrosive agent. Greatly reduced failure times and brittle fractography are the two main features of stress corrosion cracking. On both counts, iodine, iron iodide, and aluminum iodide are stress corrosion cracking agents for zircaloy. However, cesium iodide is not. This behavior is consist-(2) ent with the thermochemistry of the reactions 0.55 I2(g) + Zr(s) -Zrl^Cs) 0.37 Ml3Cg] + Zr(s; •* 0.37Ad (s) + Zrl^Cs) 0.55 Fel2(g) + Zr(s) -0.55 Fe(s) + Zrl^fs) 1.1 CsICg) + Ms) + 1.1 Cs(4) + Zrl^fs) The standard free energies suggest that the first three reactions are possible, but the last one is prohibited. This may be the reason why iodine, iron iodide and aluminium iodide, but not cesium iodide, can cause <00 K = " 37 -9 iG 600K = -n -° AG 600 K = _32 -7 iG 600K= 29 ' 4 kcal/mole (1) kcal/mole (2) kcal/mole (3) kcal/mole (4) -2-SCC of zircaloy. The rupture time data are correlated in terms of the crack growth model ^: || = A p n expC-E/RT] exp (BK/T) Where a is the crack length, p is the partial pressure of the SCC agent, K is the stress intensity factor, and T is the temperature. A and B are constants. When integrated from initial length a to final length a f at which the net section stress reaches the ultimate tensile stress, Eq. (5] gives the rupture time. The chemical or corrosion aspects of SCC is related to the reaction order n and activation energy E. These parameters are shown in Table 1 . The reaction orders for I-and Fel, are both unity, whereas that of All, is 2/3. These values reflect a rate-limiting surface chemical step with kinetics associated with the number of iodine atoms per molecule of SCC agent. Since it is unlikely that a nonlinear chemical reaction is involved in cadmium embrittlement (only alloying with the substrate metal is possible), the order for this species should be unity. This crack growth model also predicts SCC experiments with variable loading and surface roughness v .