The nuclear processes responsible for the CNO synthesis

The abundances of the isotopes of the elements C, N and O are mainly affected by the cold CNO cycles in non-explosive stellar situations, or by the hot CNO chains that can develop in certain explosive sites, like classical novae. Helium burning phases can modify the composition of the ashes of the CNO transmutations through several $\alpha$-capture reactions, the most famed one being 12C(a,g)16O. This contribution presents a short review of the purely nuclear physics limitations imposed on the accuracy of the predicted C, N and O yields from H-burning in non-explosive stars or novae. This analysis makes largely use of the NACRE compilation for the rates of the reactions on stable targets making up the cold CNO cycle. Some more recent rate determinations are also considered. The analysis of the impact of the rate uncertainties on the abundance predictions is conducted in the framework of a simple parametric astrophysical model. These calculations have the virtue of being a guide in the selection of the nuclear uncertainties that have to be duly analyzed in detailed model stars, particularly in order to perform meaningful confrontations between abundance observations and predictions. They are also hoped to help nuclear astrophysicists pinpointing the rate uncertainties that have to be reduced most urgently. A limited use of detailed stellar models is also made for the purpose of some specific illustrations.