At the low temperatures of interstellar dust grains, it is well established that surface chemistry proceeds via diffusive mechanisms of H atoms weakly bound (physisorbed) to the surface. Until recently, however, it was unknown whether atoms heavier than hydrogen could diffuse rapidly enough on interstellar grains to react with other accreted species. In addition, models still require simple reduction as well as oxidation reactions to occur on grains to explain the abundances of various molecules. In this paper we investigate O-atom diffusion and reactivity on a variety of astrophysically relevant surfaces (water ice of three different morphologies, silicate, and graphite) in the 6.5-25 K temperature range. Experimental values were used to derive a diffusion law that emphasizes that O atoms diffuse by quantum mechanical tunnelling at temperatures as low as 6.5 K. The rates of diffusion on each surface, based on modelling results, were calculated and an empirical law is given as a function of the surface temperature. The relative diffusion rates are kH2Oice > ksil > kgraph ≫ kexpected. The implications of efficient O-atom diffusion over astrophysically relevant time-scales are discussed. Our findings show that O atoms can scan any available reaction partners (e.g., either another H atom, if available, or a surface radical like O or OH) at a faster rate than that of accretion. Also, as dense clouds mature, H2 becomes far more abundant than H and the O:H ratio grows, and the reactivity of O atoms on grains is such that O becomes one of the dominant reactive partners together with H. This journal is

Any evolving system can change state via thermal mechanisms (hopping a barrier) or via quantum tunneling. Most of the time, efficient classical mechanisms dominate at high temperatures. This is why an increase of the temperature can initiate the chemistry. We present here an experimental investigation of O-atom diffusion and reactivity on water ice. We explore the 6-25 K temperature range at submonolayer surface coverages. We derive the diffusion temperature law and observe the transition from quantum to classical diffusion. Despite the high mass of O, quantum tunneling is efficient even at 6 K. As a consequence, the solid-state astrochemistry of cold regions should be reconsidered and should include the possibility of forming larger organic molecules than previously expected.