Hydride related problems are viewed as an issue to be considered for clad failure. The main sources of hydrogen for the clad are: the corrosion reaction of metal with water, hydrogen released by radiolysis of water and hydrogen gas that is added in the coolant to keep the oxygen potential low.72 The defects present in the clad (such as manufacturing defect, PCI crack, debris fretting, etc.) can aid the pick-up of hydrogen and the coolant could surge in through these defects when they grow through-thickness and form steam. The steam reacts with the fuel and hydrogen is released. When the hydrogen-to-steam ratio crosses a critical value (steam starva­tion), the growing oxide layer on the ID of the tube finally breaks down and the hydrogen diffuses into the matrix of the tube. The hydrogen thus picked up can reduce the toughness of the zirconium matrix in three ways: (i) hydride reorientation, (ii) delayed hydrogen cracking and (iii) forma­tion of a hydride blister.

The solubility limit of hydrogen in zirconium at the reactor operat­ing temperature is about 100 ppm. When the temperature is reduced (for instance, during reactor shutdown), the excess hydrogen precipitates in the form of hydride. The hydride precipitates along the radial direction of the tube owing to the texture and the hoop stress in the tube. The hoop stress required for reorientation (in an unirradiated and recrystallized Zircaloy-2) is about 80 MPa.73 The differential temperature between ID and OD of the clad wall (either during service or at wet repository) drives the hydrogen to the OD side which is at a lower temperature. The concentration of hydrides found in the tube after irradiation is higher near the water side than at the fuel side which is attributed to the corrosion reaction between the clad OD and the coolant.74 The threshold stress for failure of irradiated and hydride-reoriented spent fuel cladding is significantly higher than the stress due to the internal pressure of the fuel rod. The degree of oxidation and hydriding in the more advanced fuel claddings commonly used these days in LWRs, such as low-Sn Zircaloy-4 (Sn content around 1.3 wt.%), optimized Zircaloy-4, Zirlo (a Zr-1Sn-1Nb-0.1Fe alloy), M5 (a Zr-1Nb alloy) and opti­mized Zircaloy-2, is relatively low even at high burnup.

Frequently, a hydride blister is produced when a fuel rod that contains spalled oxide is operated continuously to high burnup. During steam star­vation, hydrogen ingress is faster than its diffusion into the tube matrix. This leads to excess amounts of hydrogen getting localized at the inner wall of the clad tube forming a large hydride called a blister. The hydrogen atoms, diffusing down the temperature gradient, form radial hydrides in a sun-burst pattern.

Delayed hydrogen cracking (DHC) is important for spent fuel in either wet or dry repository and is a two-step process. Hydrogen migrates up the stress gradient towards a stressed crack-tip and precipitates as hydride that cracks and extends further. There is an incubation time for the hydrogen to arrive and the concentration to build to the required level, so that the solubility limit is exceeded for the hydrides to form and grow, before the crack extends further. The crack-tip can undergo a corrosion reaction and the hydrogen released can either be absorbed by the matrix close to the crack-tip or the hydrogen can diffuse through the matrix to the crack-tip. The former is called corrosion hydrogen cracking and the latter is known as DHC. Knowing the crack velocity enables prediction of the failure time of the tube. Since there is no way to measure the crack velocity in the reactor, it is assumed that the crack starts at the centre of the tube and proceeds in both directions with a velocity in the range of 2.5 x 10-7-6.6 x 10-7 m/s which is determined from out-of-pile unirradiated laboratory specimens.75 It is possible that the velocity may be much higher in reactor as the stress state is more severe. The stress arises partly from the increased pellet volume because of increased temperature (due to reduced thermal conductivity as the pellet cracks up with burnup) and partly from the increased volume of the oxide layer of the Zr-liner on the interior of the tube. The increase in stress due to this volume expansion is faster than the creep relaxation by the clad.