Research Themes

Nucleation of the Earth’s Solid Inner Core

Figuring out when the Earth’s solid inner core began to freeze is easy right? The core is cooling down, so just find the time when the temperature of the core drops below the melting temperature. The trouble is that all liquids actually have to be supercooled to below the melting temperature before they begin to freeze. Did you know that water can be supercooled to 30 C below its melting temperature before it forms hail in the atmosphere, provided there is no dust to help the nucleation along. Even with dust and supercooling, there is still a waiting time for liquids to freeze. This is because the solid only starts to grow once the random vibrations of atoms arrange to form a solid of a critical size. When we estimate how much supercooling is needed to kick start inner core growth, we get between 700 and 1000 K, huge! The problem is that if we estimate how much supercooling is possible in the core, based on not freezing too much of the core, we get 200 to 400 K. This is the inner core nucleation paradox and it shows that mineral physics implies that the inner core either hasn’t seen enough supercooling to begin growing yet, or should have frozen most of the core, neither of which are true! I have developed methods to study the nucleation process using molecular dynamic simulations at conditions of Earth’s core and shown that whilst pure iron, FeSi and FeS are hopelessly incapable of resolving this paradox, FeO gets us a bit closer and FeC might actually offer a resolution to the paradox! Watch this space for a review we have coming (hopefully) soon which shows how and why this paradox matters for the thermal history of Earth.

I am also interested in what happens when this barrier to nucleation is overcome and supercooled liquids at the centre of the Earth begin to freeze. The initial growth of the inner core might have been extremely rapid but to understand how rapid, we need to learn more…

Cellular automaton of FeO freezing in Earth’s core (0.01 mm² 0.02 seconds). Simulations like this, based of thermodynamic properties from experiments and first principles, teach us about the early stages of inner core growth.

Precipitation of Light Elements from the Earth’s Outer Core

Ever wonder why we and our ancestors in the primordial soup aren’t being cooked by solar radiation? It’s because the Earth has had a strong dipolar magnetic field for at least the past 3.5 million years of course! But during this time the dynamics of the Earth’s liquid core (which produces this field) have changed significantly, so how has it been so consistent? Well it turns out that this is quite tricky to do, the problem even has it’s own paradox. The geodynamo producing the field was originally driven by the gradual cooling of the core, until the inner core began to grow and the associated processes from this continue to provide plentiful power to this day. The bit between these two is the hard part, by the time the inner core began to grow, there was likely very little heat left for powering the dynamo and models suggest that the field should have been extinguished, which is not what we see in the palaeomagnetic record, which preserves the signal of the field. So, we need to find other ways of powering the field. One way is radiogenic heating, but we find that this just isn’t sufficient. Another route is through light element precipitation. Iron is actually a pretty good solvent at high temperatures and pressures (as conveniently found in Earth’s core) and can incorporate elements like silicon and magnesium. As the core cooled, it became a less efficient solvent and had to precipitate some fraction of these elements to maintain chemical equilibrium. This would happen at the coldest point, which is the top of the core, and leave behind an iron rich liquid, which is dense. This dense liquid would then sink, helping to drive the geodynamo through the energy contributed to convection. It turns out that this process is pretty efficient, and the main question is how much precipitate can be released and how fast. We find that silicon precipitation can provide enough power to not let the magnetic field collapse, and magnesium can too, but with some caveats.

Thermodynamic Properties of Silicate Liquids

The Earth used to be hot. This because the collisions of small planets and asteroids which assembled it all imparted heat, the sum of which was sufficient to melt the majority of the planet. But what happened as the planet cooled? One of the biggest questions in the evolution of the deep Earth is when and how our solid silicate mantle froze from the magma ocean that proceeded it. This is such a conundrum because we just don’t know them melting temperature of silicates at these extreme conditions that well. The reason we don’t is that the experiments are very difficult, and the equivalent calculations are very difficult in a different way. Specifically, to calculate the melting temperature of a material precisely using computers we need to know the “free energy” of both the solid material and the liquid equivalent. When we find the conditions where these energies are the same, we have found the melting temperature. This is all hard enough, but the really tricky part is finding the free energy of the liquid because one part of this energy is particularly elusive. The entropy of a material is the measure of its “disorder”, or how its energy is distributed across all possible microstates. If all of the atoms in a substance are neatly arranged, it is not disorder and the energy is distributed across a narrow range of states, its entropy is low. If, as is the case for liquids, the atoms are arranged in many different ways and freely move between these arrangements, then many energy microstates are occupied and the system is disorder, high entropy. For solids there are defined ways of calculating entropy, because things (atoms) tend to play by the rules and stay where they’re meant to. For gases, the opposite is true, atoms fly around wildly but in a predictable way because they don’t bump into one another very often, so we can calculate (or estimate pretty well) the entropy of these systems too. Liquids are sort of a combination of the two. At short distances, a predictable structure is present, like the solid, and at long range things are random, like the gas. My PhD involved testing and developing methods for calculating entropy in liquids from molecular dynamic simulations. It went quite well and to prove it, we calculated the melting curve of one of the most abundant phases in Earth’s mantle (previously magma ocean), CaSiO₃. </details>

Stability of Thermochemical Piles in the Lowermost Mantle

Seismology tells us that there are two large features in the Earth’s mantle sitting on top of the core, we call them LLSVPs (large low shear velocity provinces) because of their seismological expression. They look something like piles, leading people to question whether they have been sitting down there a long time, being swept around by mantle convection. If this is the case, how old can they be? If they are being swept around are they not being eroded as well? In this CIDER 2018 summer project, we used 2D mantle convection models to assess just what combination of thermal and chemical properties (both defining relative bouyancy compared to the bulk mantle) these piles would need in order to not be completely eroded. My role in this project was assessing what the outputs from these models would look to seismologists, and define which of our models might be representative of what we see down there. Take a look at what we did!

Sources of Seismicity in the Martial Crust

Mars is often compared to Earth. Its rocky, close-ish to the sun, is somewhat similar in size (if your really squint) and looks like it might have some water tucked away somewhere. There are of course many differences to Earth. It doesn’t have a magnetic field, so its lost most of its water and atmosphere, and plate tectonics have come to a halt. To understand these differences we really need to understand what the interior structure of the planet is, we have an idea but not a clear picture. In 2018 NASAs insight mission placed a seismometer on Mars offering the first chance to resolve some key details of Mars’s interior. The issue was that to make use of the markquake data that would be detected (and was!) the source of seismicity would have to be known. Prior to the mission I worked with Dr Peter Grindrod to understand whether a possible source of marsquakes was active. I studies boulder populations along the valley floors in the Cerberus Fossae Region to determine weather the faulting in the area is active. As it so happens, the boulders in these valleys are clustered in specific areas, rather than evenly distributed, which implies that marksquakes originating from the faulting which created these valleys shook these boulders loose rather than the valley walls crumbling down.