European Commission logo
English English
CORDIS - EU research results
Content archived on 2024-06-18

Deep partial melting of subducted carbon and the formation of sub-lithospheric diamonds and their mineral inclusions

Final Report Summary - DEEPCARBON (Deep partial melting of subducted carbon and the formation of sub-lithospheric diamonds and their mineral inclusions)

This experimental project aims to investigate the behaviour of deeply subducted carbonate, and its relationship to melting, metasomatism, diamond formation and other processes in the lowermost upper mantle (LUM), the mantle transition zone (MTZ) and the uppermost lower mantle (ULM). As a result of intense debate about climate change, great effort has gone into the study of the Earth’s exogene (land-ocean-atmosphere) global carbon cycle. However, the Earth has another, much deeper carbon cycle, which is much less understood, even though it exerts a profound influence on greenhouse gas concentrations in the atmosphere over a variety of time-scales from thousands to billions of years. It also profoundly influences many other interrelated deep-Earth processes including magma genesis, metasomatism and diamond formation and destruction.

The key aspect of the deep carbon cycle is how deeply exogene carbon is recycled as part of altered, carbonated, oceanic crust, into the deeper earth via subduction. This process is the major return cycle for carbon back into the deep earth. In this project, I aim to use ultra high-pressure experimental petrology to investigate the behaviour of subducted carbonate at pressures corresponding to the MTZ and the ULM. Earlier experimental studies1-5 have shown that at least some residual crystalline carbonate in oceanic crust can remain stable without decarbonation or melting, and may be subducted to very deep levels in the mantle. It may form carbonate eclogite in the upper mantle and carbonate garnetite (majorite garnet) in the transition zone and uppermost lower mantle. How far subducted carbon can survive this journey to extreme depths depends on the relationship between the pressure-temperature (P-T) path followed by deeply subducting carbonated oceanic crust and its melting relations and solidus temperatures. In conjunction with my proposed scientific host (Prof Michael J Walter), I will conduct multi-anvil experiments at the University of Bristol to precisely determine partial melting relations and melt compositions in the deep upper mantle and mantle transition zone (13 GPa) of carbonate-bearing mafic compositions, similar to altered mid-ocean ridge basalt (MORB). We will explore the influences of pressure (P), temperature (T), and the key bulk compositional variable CO2/(CO2+H2O) on very deep subduction of carbonate in mafic oceanic crust and on the volumes, compositions, and fates of carbonated partial melts or fluids. We will use this new data to constrain the roles of carbon in various deep mantle processes.

Fundamental research outcomes will include understanding of:
• The role of bulk composition in determining melting temperatures and partial melt compositions of very deeply subducted, carbonate-bearing oceanic crust, and hence how deep carbonate melting can occur;
• How these carbonate-rich melts interact with surrounding ambient peridotite mantle and what sort of geochemical sources, hybrid lithologies and deeply derived magmas/fluids could be so generated (kimberlites, carbonatites, diamonds, CH4-fluids etc);
• The formation of sub-lithospheric diamonds and their exotic mineral inclusions.

The processes occurring at these depths can be simulated using solid state high pressure apparatus such as the multi-anvil device. Samples of synthetic rocks with carefully chosen chemical compositions are held at high pressures and temperatures (corresponding to conditions in the deep mantle) until the stable mineral ± melt ± fluid assemblage crystallises and chemically equilibrates. The experiment is then quenched to room temperature, preserving its high pressure, high temperature condition, the sample recovered and polished, and mineral phases and other run products quantitatively analysed using electron probe microanalysis and imaged using scanning electron microscopy.

Progress on DeepCarbon to date has included the following:

A series of appropriate bulk compositions (based on the average, mafic, altered carbonate-bearing, oceanic crustal composition [ATCM1] of Thomson et al. 2016) were prepared in the lab using synthetic, high purity oxides, carbonates and hydroxides. ATCM1 is essentially mid-ocean ridge basalt (MORB) + 2.5 wt% CO2. We have prepared a series of 6 related mixes, to which H2O as well as CO2 was added, producing a series of ATCM1 compositions with 5wt% total volatiles and CO2/(CO2+H2O) varying from 0 to 1. These were used in all subsequent experiments.

The experimental design was finalised and a decision to use tiny Au capsules to enclose the sample in each experiment and LaCrO3 heaters in a Bristol-stye 14/8 multi-anvil assembly was made.

The initial 6 or so experiments were unsuccessful because of heater failure. It transpired after a few weeks of attempts, that a faulty batch of LaCrO3 was in circulation amongst worldwide multi-anvil labs and was causing the problem. The decision was then made to change the assembly design to use graphite heaters and these proved to have a higher success rate. Other significant delays occurred in relation to delays in ordering and delivery of new batches of some consumable components to the lab, in particular Au tubing and W/Re thermocouple wires.

A series of experiments at 13 GPa (390 km depth = deep upper mantle, near the top of the MTZ) using ATCM1+2.5 wt% CO2 + 2.5 wt% H2O were conducted over a range of temperatures (1100-1300°C), expected to encompass the sub-solidus and part of the super-solidus regimes. This was based on the earlier work at Bristol University (Thomson et al 2016) in which the ATCM1 + 2.5 wt% CO2 was established at 13 GPa and 1350°C.

These experiments were examined using a JEOL 8530F field emission gun electron probe microanalyser. All experiment contained garnet with a slight super-silicic composition (majorite component, average Si cations = 3.03 per 12 oxygens), jadeite + diopside-rich clinopyroxene and stishovite. In no experiments, was evidence of carbonate or water-bearing crystalline phases observed. Previous experiments in this system (but without H2O) produced similar silicate mineral assemblages, and quenched carbonate melt or sub-solidus, crystalline carbonate phases (dolomite, magnesite, etc.). However, the new H2O+CO2-bearing runs are quite porous in texture, with spaces between crystal boundaries and at triple grain boundary junctions, indicating the likely presence of a fluid phase, likely a super-critical fluid (see attached backscattered electron image of a run product). This remains to be verified, and the subsequent experiments were aimed at this.

A series of additional experiments have been performed at 13 GPa and 1200°C, using a range of volatile compositions with ATCM1, such that XCO2 [XCO2=CO2/(CO2+H2O)] varies from 0 to 1. These experiments are expected to demonstrate the presence of a supercritical fluid in those runs in which the volatile compositions are mixed, although a H2O-rich fluid and a carbonate-rich melt may form in the runs with XCO2 = 0 and 1 respectively. These experiments are not yet analysed by electronprobe microanalysis as their fine-grained and porous nature requires examination using a field emission gun electron probe. This will be undertaken in coming weeks at the Australian National University (the former Marie-Curie Fellow’s home institute), where such an instrument has recently been installed.

If this is the correct interpretation of these results, it will be the first such finding of the likely existence of supercritical fluids (rather than melts or fluids) under these deep mantle conditions. Such fluids may be important agents of mass transfer in the deep mantle, may metasomatise regions of the mantle producing over time sources with radiogenic isotope and trace element signatures of a range of interesting magma types such as kimberlites (the main source of diamonds) and carbonatites (carbonate melts which are the main sources of Rare Earth Element ores and other critical metals) and may also be important in the formation of sub-lithospheric diamonds. This outcomes of this project are expected to be highly important in understanding of the earth’s deep carbon cycle. Publication is expected in coming months in international, peer reviewed journals, including possible inclusion of some outcomes from these experiments in a chapter to be written by the Marie-Curie Fellow on carbonate melting in the earth, for a book to be published in 2018 by Cambridge University Press and the Deep Carbon Observatory of the Carnegie Institute of Washington. This publication will be a comprehensive treatise on all aspects of the earth’s deep carbon cycle.