The origin of platinum group minerals from the freetown intrusion, Sierra Leone, inferred from osmium isotope systematics

The origin of certain alluvial platinum-group mineral (PGM) grains is contr oversial, with two major proposed models. One involves primary formation wi thin ultramafic intrusions, followed by weathering into laterites with no f urther alteration prior to alluvial concentration. The other involves in ad dition, solution of the platinum-group elements (PGE) during weathering of the ultramafic source rocks, probably by organic reactions, followed by the supergene growth of the macroscopic grains within laterites. In order to provide further evidence concerning the formation of macroscopi c (similar to mm in size) platinum-group mineral grains, the authors have s tudied the osmium isotope ratios in PGM grains from the Freetown peninsula in Sierra Leone using an ion microprobe. Indirect isotopic evidence and modeling are required to infer their origin as no PGM grains larger than similar to 10 mu m have ever been discovered i n situ within the source rocks. Unlike virtually all other alluvial PGM gra ins that have been studied, the Freetown ones are distinguished by Os-187/O s-188 ratios that are much higher than canonical mantle values and vary con siderably between different grains (0.12 less than or equal to Os-187/Os-18 8 less than or equal to 0.28). The Os-187/Os-188 ratios are, with one excep tion, the same for individual osmium-rich inclusions within the platinumiro n alloy host of each grain, although different grains have significantly di fferent Os-187/Os-188 ratios. In the one significant exception, there are t wo Os-Ru-Ir alloy inclusions within a single platinum-iron alloy host which have Os-187/Os-188 ratios that are different from each other. Various models are considered to explain these observations: (1) the elevat ed osmium isotope ratios could be primary and derived from a mantle inhomog eneity such as a mantle plume; (2) the intruded material with 8 canonical m antle signature may have reacted with the host rocks at the time of the int rusion or at a later stage during postulated metamorphic or hydrothermal ac tivity, or (3) the supergene remobilization could have modified the osmium isotope signature of the PGM. Interpretation of the osmium isotopes is by n o means unambiguous. The range of Os-187/Os-188 ratios is far higher than k nown mantle sources which, combined with the grain to grain variation, seem s to rule out the primary formation model. Sr, Nd and O isotope evidence in dicates no reaction with the host rocks during intrusion or by metamorphic or hydrothermal, activity and mixing calculations show that enhancement of the radiogenic Os during intrusion could only have been small. The evidence points to some extremely localized process which could, howeve r, have acted in two ways: either a preexisting high concentration of rheni um and PGE (a xenolith) was incorporated into the magma chamber, remained e ssentially unmixed with the majority of the material in the chamber, and af ter extrusion of the gabbro, PGE nuggets were weathered out into laterites, or alternatively concentration processes within the magma chamber produced a localized concentration of rhenium--possibly in copper sulfides or late magmatic iron oxides in acid intrusions-which, when chemically dissolved in the supergene environment, precipitated PGE nuggets with extremely radioge nic Os-187/Os-188 signatures. The data presented in this paper in conjuncti on with work elsewhere have been used to demonstrate that the processes pro ducing highly radiogenic Os-187/Os-188 signatures cannot have resulted from large-scale processes and have been used to dismiss a number of previously proposed models. However, a resolution between the two alternative models proposed is not possible using the present data.