The origin of cratonic peridotites: A major-element approach

Cratons formed in tectonic events that occulted at levels extending to as much as several hundred kilometers within the upper mantle, clearly deeper than the thicknesses of Phanerozoic oceanic plates. These events were primarily Archean, but Re depletion ages for the peridotites indicate that some extended into the Proterozoic. The peridotites that comprise craton roots are spinel facies to depths of similar to 100 km. The section between 100 and 200 km is coarse garnet peridotite with equilibration temperatures ranging upward to similar to 1100 degrees C. The peridotites of deepest origin commonly are deformed, metasomatically enriched in Ti and Fe, and record temperatures up to 15000C. The peridotites that form the principal mass of the craton root are buoyant because their Mg/(Mg+Fe) is greater than that of circumcratonic peridotites, and because they contain less garnet-forming Al. Most also are characterized by being orthopyroxene-rich relative to oceanic peridotites. The low-temperature peridotites have provided our best estimates of the bulk composition of the craton mantle, although they have not escaped late-stage metasomatism. Balancing bulk analyses for Fe and Ca against the probe analyses and modes for these elements shows that, in some cases, there has been significant introduction from the kimberlite during eruption. Cratonic peridotites are depleted in most magmaphile elements, but their compositional variations are not well characterized by depletion models, even with allowance for metasomatism during eruption. Plots of Fe vs. Si: for example, have a negative trend-opposite from that expected if these elements were jointly concentrated in melt rather than in the residue. Negative Fe/Si could be the product of a mixing or unmixing process involving olivine and orthopyroxene, or it could be the product of a metasomatic reaction of olivine with melt to form orthopyroxene. Kelemen and Hart (1996) have observed that there will be enrichment of Ni in orthopyroxene in the event that it is generated by melt-rock reaction, but their interpretation is not unique. Similar enrichments can be generated by cooling assemblages of orthopyroxene and olivine that originated with variable modal proportions (Herzberg, 1998). An excellent positive correlation of Ni in opx with modal opx is found for Premier peridotites, which is the result either of cooling or melt-rock reaction. Those from Kimberley, however, show a flat trend. The opx-rich and opx-poor peridotites at Kimberley appear to have equilibrated following cooling to ambient conditions. The origins of cratonic peridotites remain an imperfectly understood, first-order geologic problem.