During the 1970s, reactor safety authorities developed increasing interest in methods for accurately predicting the extent of hazards associated with severe accidents in light water reactors (LWRs). In response to these concerns, out-of-pile experimental projects were initiated by the U.S. Nuclear Regulatory Commission and the French Nuclear Protection and Safety Institute, at Oak Ridge National Laboratory (ORNL) and the Commissariat à l’Energie Atomique (CEA), respectively. Both experimental efforts were designed for source term characterization of the fission products (FPs) released from LWR fuel samples under test conditions representative of severe accidents, i.e., in oxidizing or reducing atmospheres at temperatures up to 2700 K (at ORNL) and 2570 K (at CEA). The experimental devices, procedures, and parameters are described. The combined database of available results is summarized and related to experimental conditions. Using Booth diffusion theory, diffusion coefficients of the FPs were calculated, and their evolution with temperatures in the 1070 to 2700 K range were plotted. The results show the good agreement between the independently determined ORNL and CEA FP diffusion coefficient values. By plotting the data in Arrhenius fashion, it has been possible to do the following:

  1. quantify the thermal activation coefficients for both volatile and low volatile FPs
  2. identify classes of FPs whose release behavior does not follow a purely diffusional mechanism, but rather depends on chemical interactions with the environment; i.e., they exhibit mixed diffusional and transport-dependent mechanisms.
A model supported by these experimental results and based on the CORSOR-Booth code is proposed and compared with CORSOR calculations. To facilitate the comparison, a model based on FP diffusion mechanisms that was developed at ORNL and adapted with CEA experimental data is proposed. This CEA model, as well as CORSOR-Booth calculations are compared with the ORNL experimental data in a blind test. Additional experimental work is in progress to determine the releases of nonvolatile FPs and transuranics at temperatures up to fuel melting as well as FP deposition mechanisms under both oxidizing and reducing conditions. An effort is made to quantify the dependence of the temperature transient δT/δt as well as the oxygen potential effects on the release kinetics. Those data will allow further verification and extension of the field of application for our model.