Detection of an Iron Spin Transition in Ferro-periclase in the Lower Mantle

Seismic detection of an iron electron spin pairing transition in (Mg,Fe)O ferropericlase (Fp) would yield an important constraint on lower mantle composition because the presence of Fp requires that (Mg+Fe)/(Si-Ca)>1 (e.g., as required for pyrolite or harzburgite). The absence of a clear signal of this transition in global seismic profiles could be taken to mean that (1) a substantial portion of the lower mantle is Si-enriched (Fp-poor) relative to pyrolite, (2) seismological resolution is insufficient to see it, and/or (3) the predictions of mineral physics are inaccurate. W e also have a variety of evidence that subducted lithospheric slabs containing abundant Fp penetrate well into the lower mantle, so we focused our efforts on interrogating the characteristics of fast velocity anomalies in the relevant depth range. We found that S wave tomography models resolve fast material continuously with depth but P wave models show a drop-off in detection of strong fast anomalies at depths expected for the mixed spin state . This is because P wave velocity loses sensitivity to temperature in the mixed-spin region, such that cold (or hot) anomalies do not manifest a strong velocity anomaly, unlike S wave anomalies. Our findings are broadly consistent with the proposal that large highly viscous regions called bridgmanite- enriched ancient mantle structures (BEAMS) stably reside in the mid-lower mantle and organize the pattern of deep mantle convection over billions of years.