A reinvestigation of debris disc halos

Context. Scattered-light images reveal that a significant fraction of debris discs consist of a bright ring beyond which extends a wide halo. This halo is expected and should be made of small grains collisionally produced in the ring of parent bodies (PBs) and pushed on high-eccentricity orbits by radiation pressure. It has been shown that, under several simplifying assumptions, the surface brightness (SB) of this halo should radially decrease as r(-3.5) in scattered light Aims. We aim to revisit the halo phenomenon and focus on two unexplored issues: (1) how the unavoidable presence of small unbound grains, non-isotropic scattering phase functions (SPFs), and finite instrument resolution affect scattered-light SB profiles, and (2) how the halo phenomenon manifests itself at longer wavelengths in thermal emission, both on resolved images and on system-integrated spectral energy distributions (SEDs). Methods. We use a collisional evolution code to estimate the size-dependent spatial distribution of grains in a belt+halo system at steady state. We use the GRaTeR radiative-transfer code to derive synthetic images in scattered light and thermal emission, as well as SEDs. Results. We find that unbound grains account for a significant fraction of the halo's luminosity in scattered light, and can significantly flatten the SB radial profile for the densest and brightest discs. Because halos are strongly size-segregated with radial distance, realistic size-dependent SPFs also have an effect, resulting here again in shallower SB profiles. For edge-on discs, non-resolving the vertical profile can also significantly flatten the projected SB profile. We show that roughly half of the observationally derived halo profiles found in the literature are compatible with our new results, and that roughly half of the remaining systems are probably shaped by additional processes (planets, stellar companions, etc.). We also propose that, in future observational studies, the characteristics of the PB belts and the halos should be fitted separately. In thermal emission, we find that wide halos should remain detectable up to the far-infrared (far-IR) and that, with the exception of the similar to 8-15 mu m domain, the halo accounts for more than half of the system's total flux up to lambda similar to 80-90 mu m. The contribution from the halo strongly decreases in the submm to mm but still represents a few percent of the system's luminosity at lambda similar to 1 mm. For unresolved systems, the presence of a halo can also affect the determination of the radius of the disc from its SED.