System response deconvolution and desorption model fitting to TPD tritium release from neutron irradiated Li ceramics

A procedure is evaluated for the analysis of the tritium release from neutron irradiated lithium ceramics by means of temperature ramps, and tritium detection using ionization chambers or proportional counters. A system with convolution time constants and delays of the order of 1 min., and heat rates of 1 to 10 K/min. is considered as a reference. The squared error minimization method is shown to be only a guide towards identifying the release mechanisms, to be suported by spectroscopic or other analysis, whenever more than one process contributes significantly to the overall release signal. Fourier transform deconvolution, assisted by low-pass data filtering, is shown to be possible for data with 0.5% of measurement noise. Numerical results are shown of the error incurred of the release peak maximum temperature if failing to deconvolute the system response prior to data analysis. These errors may induce the misidentification of a first order process as of second order, at high heating rates.