The Chicxulub impact is commonly believed to have caused the Cretaceous–Tertiary boundary mass extinction and a thin impact spherule layer in the North Atlantic and Caribbean is frequently cited as proof. We evaluated this claim in the seven best North Atlantic and Caribbean Cretaceous–Tertiary boundary sequences based on high-resolution biostratigraphy, quantitative faunal analyses and stable isotopes. Results reveal a major Cretaceous–Tertiary boundary unconformity spanning most of Danian subzone P1a(1) and Maastrichtian zones CF1–CF2 (~400 ka) in the NW Atlantic Bass River core, ODP Sites 1049A, 1049C and 1050C. In the Caribbean ODP Sites 999B and 1001B the unconformity spans from the early Danian zone P1a(1) through to zones CF1–CF4 (~3 Ma). Only in the Demerara Rise ODP Site 1259B is erosion relatively minor and restricted to the earliest Danian zone P0 and most of subzone P1a(1) (~150 ka). In all sites examined, Chicxulub impact spherules are reworked into the early Danian subzone P1a(1) about 150–200 ka after the mass extinction. A similar pattern of erosion and redeposition of impact spherules in Danian sediments has previously been documented from Cuba, Haiti, Belize, Guatemala, south and central Mexico. This pattern can be explained by intensified Gulf stream circulation at times of climate cooling and sea level changes. The age of the Chicxulub impact cannot be determined from these reworked impact spherule layers, but can be evaluated based on the stratigraphically oldest spherule layer in NE Mexico and Texas, which indicates that this impact predates the Cretaceous–Tertiary boundary by about 130–150 ka.