Material transport in the deep Earth's mantle
The models considering fast mantle upwelling without an addition of volatile-bearing components are failed due to very high melting temperatures of mantle silicates. Without melting the material transport will be hardly possible. Thus, considering mantle plumes one should add fusible component to the system. In most geodynamic models fusible component is involved indirectly, for example, just changing the viscosity. Almost nobody try to consider the real process of melt movement through presumably non-porous mantle matrix or try to understand the nature of melt, which assist material transport in the very deep mantle. Some mechanisms, such as melt percolation or hydraulic magma fracturing, are applicable for lithospheric depths and cannot be considered as a reliable mechanism for deeper mantle with high plasticity and low porocity. The most likely mechanism, which can operate in the deep mantle to assist plume or diapiric ascent is the stress-driven dissolution-precipitation, however, the possible fusible component of plume melt in the deep mantle is a matter of debates. Here, I discuss possible compositions of melt in the upwelling mantle, which can drive material transport under superplumes and hot spots originated from the transition zone of from the core-mantle boundary.
The most likely candidates for fusible chemicals in the mantle plumes are alkali-bearing species, C-O-H volatiles, and carbonates. An important requirement for plume motion would be stress-induced melting and dissolution-precipitation of the fusible component at the front and rear of the plume, respectively. For this process one would have a volatile-bearing melt with low solubility of silicates (ca. 5-10%, but not zero) at the temperature of mantle geotherm (or slightly higher). The possible candidates are alkali-bearing silicate melt, hydrous silicate melt, carbonatite melt, and hydrocarbon-bearing melt. Alkaline silicate melt and hydrous silicate melt cannot be considered, since a huge amount of silicate can be dissolved in these melts and the process of plume ascent will be easily terminated by progressive reactions with the surrounding silicate matrix. Carbonated or carbonatite melt is a likely candidate, especially for transition zone. Phase relations in the alkaline carbonatite systems indicate that major melting of subducted carbonates should occur at the transition zone depths. Taking into account the amount of subducted carbonated (1-2 wt.% CO2) in the top 500 m of model slab we proposed a model for mobile carbonatite melt diapirs, generating from the slab in the transition zone, migrating upwards, modifying and oxidizing possibly reduced mantle section, precipitating diamonds, creating enriched source regions, and initiating volcanism at the surface. Dehydration of stagnant subducted slabs in the transition zone may accompany carbonatite diapir formation, however significant involvement of hydrous species into melting in the transition zone is difficult due to very high solubility of water in ringwoodite and wadsleyite, the major minerals in the transition zone.
It should be noticed that carbonate or carbonatite melt may not survive through the lower mantle due to reduction to diamond or other carbon-bearing species (carbide) if we assume redox state of the lower mantle close to the iron-wustite (IW) buffer. Thus, hydrocarbon-bearing or hydrous hydrocarbon-bearing melt might be the best candidate for the liquid portion of a mantle plume arising from the core-mantle boundary. There is limited amount of information about hydrocarbon phase relations and reactions with silicates in the lower mantle due to an extremely difficult experimental setup. The data for melting of volatile-bearing peridotite in the system buffered by the IW buffer at 1-3 GPa indicated negligible solubility of silicates in coexisting CH4-H2O fluid. However, recent melting experiments on peridotite and eclogite systems with reduced C-O-H fluid at 3-16 GPa indicated significant solubility of silicates in the coexisting C-O-H fluid. The diamond or graphite traps contained abundant microinclusions of silicates after experiments. The composition of fluid was not measured in the experiments, whereas theoretical estimates indicate a mixture of H2O with methane and possibly heavier hydrocarbons. Similar fluid/melt containing H2O and hydrocarbons with a relatively low solubility of silicate components along the mantle geotherm can exist through the lower mantle and can be considered as the most reliable candidate for the fusible component of mantle plumes from CMB. In support of hydrocarbon-bearing melt/fluid in the deep mantle, we demonstrated recently that formation of a hydrocarbon mixture is highly probable under reducing conditions at the core-mantle boundary.