2013MNRAS.431...63T

I show that the temperature of nuclear star clusters, starburst clusters in M82, compact high-z galaxies and some globular clusters of the Galaxy likely exceeded the ice-line temperature (TIce ~ 150-170K) during formation for a time comparable to the planet formation time-scale. The protoplanetary discs within these systems will thus, not have an ice line, decreasing the total material available for building protoplanetary embryos, inhibiting the formation of gas- and ice-giants if they form by core accretion, and prohibiting habitability. Planet formation by gravitational instability is similarly suppressed because Toomre's Q > 1 in all but the most massive discs. I show that cluster irradiation can in many cases dominate the thermodynamics and structure of passive and active protoplanetary discs for semi-major axes larger than ∼ 1-5au. I discuss these results in the context of the observed lack of planets in 47 Tuc. I predict that a similar search for planets in the globular cluster NGC 6366 ([Fe/H] = -0.82) should yield detections, whereas (counterintuitively) the relatively metal-rich globular clusters NGC 6440, 6441 and 6388 should be devoid of giant planets. The characteristic stellar surface density above which TIce is exceeded in star clusters is ∼ 6x103M☉pc- 2f- 1/2dg,MW, where fdg,MW is the dust-to-gas ratio of the embedding material, normalized to the Milky Way value. Simple estimates suggest that ∼ 5-50 per cent of the stars in the universe formed in an environment exceeding this surface density. Future microlensing planet searches that directly distinguish between the bulge and disc planet populations of the Galaxy and M31 can test these predictions. Caveats and uncertainties are detailed.