Garnet and spinel in the Upper Mantle: Results from thermodynamic modeling in fertile and depleted compositions

Spinel–garnet relations in the upper mantle have long been investigated from different perspectives, by means of petrological studies on natural samples (orogenic massifs, xenoliths, diamonds), high-P-T experiments, thermodynamic calculations, and geophysical observations. Here we report a refined thermodynamic model that allows one to predict phase relations and mineral compositions in a wide range of realistic mantle compositions. The generated phase diagrams show that the garnet+spinel stability field is always broad at low temperatures and progressively narrows with increasing T. In lithospheric sections with hot geotherms, garnet coexists with spinel across an interval of 10–15 km, at ca. 50–70 km depths. In colder, cratonic, lithospheric sections, the width of the garnet–spinel transition strongly depends on bulk composition: in fertile mantle, spinel can coexist with garnet to about 120 km depth, while in a strongly depleted harzburgitic mantle spinel is stable to over 180 km depth. These results are in agreement with the observed extension of the Hales gradient zone (a seismic impedance increase in the mantle that is usually attributed to the spinel-to-garnet transition) in various geodynamic settings. The model predicts that formation of chromian spinel inclusions in diamonds is restricted to pressures between 4.0 and 6.0 GPa. The calculated modes of spinel decrease rapidly to less than 1 vol% when garnet enters the equilibrium assemblage, hence spinel grains can be easily overlooked during the petrographical characterization of small mantle xenoliths. The very Cr-rich nature of many spinels from xenoliths and diamonds from cratonic settings may simply be a consequence of their low modes in high-P assemblages and does not require ultra-depleted compositions. The model also shows that large Ca and Cr variations in lherzolitic garnets in equilibrium with spinel can be explained by variations of pressure and temperature along a continental geotherm and do not necessarily imply variations of bulk composition.