LAUSR

Abstract The origin of variations in 186 Os/ 188 Os ratios amongst mantle-derived basaltic and komatiitic lavas remains controversial, with opposing models arguing for deep core-mantle versus shallow mantle sources. Crustal contamination has generally not been favored due to the low Os contents of such sources, meaning that variations in 186 Os/ 188 Os would require involvement of extremely high proportions of crustal material. Here we re-examine crustal contamination as an effective means for generating significant 186 Os/ 188 Os variations in Earth materials. Using chromitites and peridotites from the Stillwater, Muskox and Rum layered intrusions, we show that radiogenic 186 Os/ 188 Os ratios are correlated with 187 Os/ 188 Os ratios and can only be explained by shallow-level mixing processes and crustal contamination. The samples have δ 186 Os ([{( 186 Os/ 188 Os sample[t] / 186 Os/ 188 Os PM(t) ) − 1} × 1000], where the modern primitive mantle [PM] 186 Os/ 188 Os is 0.1198388) values ranging between 0.04 to 0.15 for the ∼2.7 Ga Stillwater Igneous Complex, −0.05 to 0.17 for the ∼1.27 Ga Muskox Intrusion, and 0.02 to 0.13 for the ∼0.06 Ga Rum Layered Suite. The highly siderophile element (HSE: Os, Ir, Ru, Pt, Pd, Re) contents of the chromitites and peridotites can be modeled through high sulfide-melt partitioning (typically >8000) and emphasize the role of S-saturation and HSE scavenging. Considering the high sulfide-melt partitioning and accounting for high silicate melt to sulfide melt ratios ( R -factors), it is possible to explain the variations in 186 Os– 187 Os in layered intrusions using calculated Os isotope crustal evolution growth models. These calculations indicate that 186 Os is a possible cause for 186 Os/ 188 Os ratio variations observed in some komatiites. It is more difficult to explain radiogenic 186 Os/ 188 Os measured in Hawaiian lavas by crustal contamination processes. Instead, ancient high Pt/Os oceanic crust, shallow mantle sources such as metasomatic sulfide, or metal-rich large low-shear wave velocity provinces at the core-mantle boundary, all remain valid explanations.