Effects of quintessence on observations of Type Ia supernovae in the clumpy universe

We discuss the amplification dispersion in the observed luminosity of standard candles, such as supernovae (SNe) of Type Ia, induced by gravitational lensing in a universe with dark energy (quintessence). We derive the main features of the magnification probability distribution function (pdf) of SNe in the framework of on average Friedmann-Lemaitre-Robertson-Walker (FLRW) models for both lensing by large-scale structures and compact objects. Analytic expressions, in terms of hypergeometric functions, for luminosity distance-redshift relations in a flat universe with homogeneous dark energy have been corrected for the effects of inhomogeneities in the pressureless dark matter (DM). The magnification pdf is strongly dependent on the equation of state, ω Q , of the quintessence. With no regard to the nature of DM (microscopic or macroscopic), the dispersion increases with the redshift of the source and is maximum for dark energy with very large negative pressure; the effects of gravitational lensing on the magnification pdf, i.e. the mode biased towards de-amplified values and the long tail towards large magnifications, are reduced for both microscopic DM and quintessence with an intermediate ω Q . Different equations of state of the dark energy can deeply change the dispersion in amplification for the projected observed samples of SNe Ia by future space-borne missions. The 'noise' in the Hubble diagram due to gravitational lensing strongly affects the determination of the cosmological parameters from SNe data. The errors on the pressureless matter density parameter, Ω M , and on ω Q are maximum for quintessence with not very negative pressure. The effect of the gravitational lensing is of the same order as the other systematics affecting observations of SNe Ia. As a result of the lensing by large-scale structures, in a flat universe with Ω M = 0.4, at z = 1 a cosmological constant (ω Q = -1) can be interpreted as dark energy with ω Q < -0.84 (at 2a confidence limit).