Carbon systematics of the Icelandic crust and mantle

In recent decades there has been an increased interest in the carbon content of Earth’s geochemical reservoirs due to the impact of atmospheric carbon on the habitability of our planet. Earth’s interior likely hosts a greater mass of carbon than that of the oceans, atmosphere and crust combined, which has buffered the carbon content of the atmosphere over geological time. Yet only a few direct measurements of carbon from the upper mantle, and none from the lower mantle, have been made. Undegassed basalts erupted at mid-ocean ridges have previously been used to estimate the carbon content of the upper mantle. However, due to the low solubility of carbon within silicate melt, these undegassed basalt suites are rare. The majority of basalts have lost their mantle carbon information en route to eruption through the crust. Various crustal processes act to modify the geochemistry of melts before eruption, therefore it is important to be able to characterise the effect of these processes to better interpret the volatile signals preserved in erupted products. Pressure, and therefore depth, is a key parameter controlling volatile solubility and can be estimated using a variety of igneous barometers. This thesis presents results from crys- tallisation experiments conducted on basaltic glass from the Miðfell eruption, Iceland. The experiments provide new data that has been used to test a variety of barometers and crystalli- sation models used by igneous petrologists, and could aid future barometer recalibration. A key part of this work was the development of an experimental method for stabilising 5 kbar conditions in a piston cylinder apparatus. The experiments have shown that clinopyroxene- liquid barometry is more reliable than multi-reaction barometry. However, knowledge of equilibrium clinopyroxene compositions is crucial for accurately determining pressure using the clinopyroxene-liquid barometer. More experiments conducted at mid-crustal pressures are required for a full recalibration of these barometers. The results of testing igneous barometers and crystallisation models have been applied to two suites of olivine-hosted melt inclusions from the Kistufell and Miðfell eruptions to help determine the melt evolution history of these basalts. These eruptions were targeted due to previously measured noble gas isotopic ratios that suggest a primordial mantle component present in their melting regions, and therefore evoking the possibility that they could hold information about deep mantle carbon. Barometry suggests that Miðfell phases equilibrated, and therefore crystallised, at mid-crustal pressures (5–7 kbar), which could allow for the entrapment of undegassed melt inclusions within olivine. The two melt inclusion suites were found to differ in trace element variability, with the observation that the more trace element enriched eruption, Kistufell, had lower relative trace element variability than the more depleted eruption, Miðfell. Several processes, both in the crust and the mantle, are likely responsible for the level of trace element enrichment and variability, including extent of mantle melting, source heterogeneity, and melt transport. The depleted nature of the Miðfell melt inclusions has allowed them to preserve some of the highest CO$_2$/Ba and CO$_2$/Nb ratios ever recorded in basaltic glass, with ratios over five times greater than undegassed mid-ocean ridge basalt values. This carbon enrichment is not due to any crustal melt modification process, but rather pertaining to lower mantle carbon-rich lithologies that have been tapped by the Icelandic mantle plume. The carbon reservoir beneath Miðfell is estimated to contain 744 $\pm$ 188 ppm carbon, 15 times greater than the depleted upper mantle. This value matches estimates of bulk mantle carbon from planetary mass balance calculations and provides evidence for carbon-rich domains within the Earth.