Aims. Gravitational settling of 22Ne in cooling white dwarfs can affect the outcome of thermonuclear supernovae. We investigate how the supernova energetics and nucleosynthesis are in turn influenced by this process. We use realistic chemical profiles derived from state-of-the-art white dwarf cooling sequences. The cooling sequences provide a link between the white dwarf chemical structure and the age of the supernova progenitor system. Methods. The cooling sequence of a 1 M white dwarf was computed until freezing using an up-to-date stellar evolutionary code. We computed explosions of both Chandrasekhar mass and sub-Chandrasekhar mass white dwarfs, assuming spherical symmetry and neglecting convective mixing during the pre-supernova carbon simmering phase to maximize the effects of chemical separation. Results. Neither gravitational settling of 22Ne nor chemical differentiation of 12C and 16O have an appreciable impact on the properties of type Ia supernovae, unless there is a direct dependence of the flame properties (density of transition from deflagration to detonation) on the chemical composition. At a fixed transition density, the maximum variation in the supernova magnitude obtained from progenitors of different ages is ∼0.06 mag, and even assuming an unrealistically large diffusion coefficient of 22Ne it would be less than ∼0.09 mag. However, if the transition density depends on the chemical composition (all other things being equal) the oldest SNIa can be as much as 0.4 mag brighter than the youngest ones (in our models the age difference is 7.4 Gyr). In addition, our results show that 22Ne sedimentation cannot be invoked to account for the formation of a central core of stable neutron-rich Fe-group nuclei in the ejecta of sub-Chandrasekhar models, as required by observations of type Ia supernovae.