In case of a hypothetical severe accident in a nuclear LWR (light water reactor), the high radiation fields reached in the reactor containment building due to the release of fission products from the reactor core could induce air radiolysis. The air radiolysis products could, in turn, oxidise gaseous molecular iodine into aerosol–borne iodine–oxygen–nitrogen compounds. Thereby, this reaction involves a change of iodine speciation and a decrease of iodine volatility in the reactor containment atmosphere. Kinetic data were produced within the PARIS project on the air radiolysis products formation and destruction, and on their reaction with molecular iodine, with the objective of developing and validating existing kinetic models. The current paper includes the non-iodine tests of the PARIS project whose objective was to determine the rates of formation and destruction of air radiolysis products in the presence of both structural containment surfaces (decontamination coating (“paint”) and stainless steel), aerosol particles such as silver rich particles (issued from the control rods) in boundary conditions representative for LWR or PHEBUS facility containments. It is found that the air radiolysis products concentration increases with dose and tend to approach saturation levels at doses higher than about 1 kGy. This behaviour is more evident in oxygen/steam atmospheres, producing ozone, than in air/30% (v/v) steam atmospheres, the latter favouring the model-predicted on-going production of nitrogen dioxide even at very high doses. No significant effect of temperature, dose rate and hydrogen addition (4%, v/v) was observed. Furthermore, the inserted surfaces do not exhibit significant effects on the air radiolysis concentrations. However, these “non-noticeable influence” could be due to a masking of small effects by the appreciable scattering of the experimental air radiolysis product concentrations. The PARIS results are then analysed using two different kinetic models, an empirical and a mechanistic one. The kinetic constants within an empirical model including formation and destruction of air radiolysis products, derived from PARIS results, are in reasonable agreement with those determined previously for lower steam fractions. From the mechanistic model IODAIR-IRSN, it is concluded that ozone is the predominant air radiolysis product at low doses in air/steam atmospheres. At doses higher than 1 kGy, nitrogen dioxide becomes increasingly important, both due to an increase in its concentration and due to a simultaneous decrease in ozone concentration.