The waterside corrosion kinetics of Zircaloy-4 are accelerated in pressurized water reactors (PWRs) in comparison with autoclaves. Beyond this comparison, an enhancement of oxidation rate - called phase III - can be observed from the third reactor cycle. This results in significant oxide thicknesses at high burnups. Several hypotheses have been devised to explain this phase III of Zircaloy-4 in PWRs, but none have been fully validated. In an attempt to better understand the oxidation acceleration phenomenon affecting Zircaloy-4 in PWRs, we decided to analyze the in-reactor corrosion of Zircaloy-4 by quantifying the acceleration factor KPWR. This was defined as the multiplication factor to be applied to the oxidation rate in an autoclave to obtain the kinetics in a PWR (with an equivalent metal-oxide interfacial temperature and taking into account both the power and thermal-hydraulic histories). This analysis was based on oxide thicknesses formed on Zircaloy-4 cladding containing UO2 or mixed-oxide fuel and having been irradiated for one to six cycles in French PWRs. This analysis enabled us to demonstrate the following:

1. KPWR is always >1, which clearly shows an acceleration in the Zircaloy-4 oxidation kinetics in a reactor.

2. KPWR is equivalent to [approximately]2 for rods having been subjected to one or two cycles.

3. Above two reactor cycles, KPWR increases with the level of irradiation and ends up reaching values close to 6. This KPWR increase is representative of phase III.

4. KPWR and its variations are not directly related to the increase in the fluence. Phase III is not associated with a burnup threshold.

5. Phase III seems to be related to a threshold that is a function of the oxide layer thickness.

6. The precipitation of hydrides could be used to define a threshold that is a function of the oxide layer thickness above which phase III occurs. This hypothesis is consistent with the thickness at which KPWR increases. Furthermore, phase III observed is consistent with the known increase in the oxidation kinetics of samples with hydride rims in an autoclave.

Therefore, acceleration of the oxidation kinetics in a reactor (compared with an autoclave) is not constant but does seem to be a complex function of different variables such as time, temperature, and both the thermal and neutron fluxes. Furthermore, the precipitation of hydrides seems to be a first-order factor triggering phase III of Zircaloy-4 in a reactor.