Diamond formation in the deep lower mantle: a high-pressure reaction of MgCO3 and SiO2

Abstract

Diamond is an evidence for carbon existing in the deep Earth. Some diamonds are considered to have originated at various depth ranges from the mantle transition zone to the lower mantle. These diamonds are expected to carry significant information about the deep Earth. Here, we determined the phase relations in the MgCO3-SiO2 system up to 152 GPa and 3,100 K using a double sided laser-heated diamond anvil cell combined with in situ synchrotron X-ray diffraction. MgCO3 transforms from magnesite to the high-pressure polymorph of MgCO3, phase II, above 80 GPa. A reaction between MgCO3 phase II and SiO2 (CaCl2-type SiO2 or seifertite) to form diamond and MgSiO3 (bridgmanite or post-perovsktite) was identified in the deep lower mantle conditions. These observations suggested that the reaction of the MgCO3 phase II with SiO2 causes formation of super-deep diamond in cold slabs descending into the deep lower mantle.

Stability and Reactions of CaCO3 Polymorphs in the Earth's Deep Mantle

“…Schematic reaction line of () based on the recent experimental constraints and simulated data in this study. The phase diagram of the MgCO 3 ‐SiO 2 system, with measured reaction line of () (in green) and phase boundaries of relevant phases (St: Stishovite; CS: CaCl 2 ‐type SiO 2 ; se: Seifertite; mgs: Magnesite; MgsII: Magnesite‐phase II; Brd: Bridgmanite; PPv: MgSiO 3 ‐post‐perovskite), follows that proposed by Maeda et al (). Phase boundaries of CaCO 3 polymorphs predicted in this study are included with light‐gray solid lines.…”

“…The MgCO 3 ‐phase III and MgCO 3 ‐ P 2 1 2 1 2 1 are found to be closely equivalent in energy and are stable above ~144 GPa. Therefore, in the inset plot of Figure , only three polymorphs are involved and the phase finds its stability in a limited regime, which explains the recent experiments that magnesite transforms directly to MgCO 3 ‐phase II at high temperatures (Maeda et al, ).…”

“…As shown in Figure with the bold solid line, the reaction line of has already been tightly constrained up to over 50 GPa in our recent measurements with synchrotron X‐ray diffraction in laser‐heated diamond anvil cells (Li et al, ). To extend this line to even higher pressures without the need of modeling CO 2 and O 2 , which are still inadequate and challenging (Litasov et al, , ), we find the following reaction examined by Maeda et al ()…”

“…The reaction line of ((R2)) (ΔG R2 = 0) is shown in red. Geotherms of subducted slabs are shown after Maeda et al (2017).…”

“…It should be noted that this indirect way of extending the reaction line of (R3) may involve noticeable uncertainty, mostly from that of the reaction line of (R4). Although Maeda et al (2017) constrains the decarbonation reaction boundaries of MgCO 3 much more precisely with increased number of high T-P measurements as compared with its previous studies (e.g., Seto et al, 2008), the temperature and pressure uncertainties for each data point are still remarkable (up to 10 GPa and 370 K, see Figure 2 and Table S1 of Maeda et al, 2017). The steep gradient of decomposition boundary of CO 2 as shown in Figure 5, in particular, should be highly uncertain and is apparently inconsistent with the previous estimations .…”