There exists strong circumstantial evidence from their eccentric orbits that most of the known extra-solar planetary systems are the survivors of violent dynamical instabilities. Here we explore the effect of giant planet instabilities on the formation and survival of terrestrial planets. We numerically simulate the evolution of planetary systems around Sun-like stars that include three components: (i) an inner disk of planetesimals and planetary embryos; (ii) three giant planets at Jupiter-Saturn distances; and (iii) an outer disk of planetesimals comparable to estimates of the primitive Kuiper belt. We calculate the dust production and spectral energy distribution of each system by assuming that each planetesimal particle represents an ensemble of smaller bodies in collisional equilibrium. Our main result is a strong correlation between the evolution of the inner and outer parts of planetary systems, i.e. between the presence of terrestrial planets and debris disks. Strong giant planet instabilities -that produce very eccentric surviving planetsdestroy all rocky material in the system, including fully-formed terrestrial planets if the instabilities occur late, and also destroy the icy planetesimal population. Stable or weakly unstable systems allow terrestrial planets to accrete in their inner regions and significant dust to be produced in their outer regions, detectable at mid-infrared wavelengths as debris disks. Stars older than ∼100 Myr with bright cold dust emission (in particular at λ ∼ 70 μm) signpost dynamically calm environments that were conducive to efficient terrestrial accretion. Such emission is present around ∼16% of billion-year old Solar-type stars. Our simulations yield numerous secondary results: 1) the typical eccentricities of as-yet undetected terrestrial planets are ∼0.1 but there exists a novel class of terrestrial planet system whose single planet undergoes large amplitude oscillations in orbital eccentricity and inclination; 2) by scaling our systems to match the observed semimajor axis distribution of giant exoplanets, we predict that terrestrial exoplanets in the same systems should be a few times more abundant at ∼0.5 AU than giant or terrestrial exoplanets at 1 AU; 3) the Solar System appears to be unusual in terms of its combination of a rich terrestrial planet system and a low dust content. This may be explained by the weak, outward-directed instability that is thought to have caused the late heavy bombardment.Key words. planets and satellites: formation -methods: numerical -planets and satellites: dynamical evolution and stabilitycircumstellar matter -infrared: planetary systems -astrobiology IntroductionCircumstellar disks of gas and dust are expected to produce three broad classes of planets in radially-segregated zones (Kokubo & Ida 2002). The inner disk forms terrestrial (rocky) planets because it contains too little solid mass to rapidly accrete giant planet cores, which are thought to form preferentially beyond the snow line where the surface density in solids is ...

We present Herschel far-infrared and submillimeter maps of the debris disk associated with the HR 8799 planetary system. We resolve the outer disk emission at 70, 100, 160 and 250 µm and detect the disk at 350 and 500 µm. A smooth model explains the observed disk emission well. We observe no obvious clumps or asymmetries associated with the trapping of planetesimals that is a potential consequence of planetary migration in the system. We estimate that the disk eccentricity must be < 0.1. As in previous work by Su et al. (2009), we find a disk with three components: a warm inner component and two outer components, a planetesimal belt extending from 100 -310 AU, with some flexibility (±10 AU) on the inner edge, and the external halo which extends to ∼ 2000 AU. We measure the disk inclination to be 26 ± 3 • from face-on at a position angle of 64 • E of N, establishing that the disk is coplanar with the star and planets. The SED of the disk is well fit by blackbody grains whose semi-major axes lie within the planetesimal belt, suggesting an absence of small grains. The wavelength at which the spectrum steepens from blackbody, 47 ± 30 µm, however, is short compared to other A star debris disks, suggesting that there are atypically small grains likely populating the halo. The PACS longer wavelength data yield a lower disk color temperature than do MIPS data (24 and 70 µm), implying two distinct halo dust grain populations.