Studying the chemical composition of some of the planet’s oldest rocks has revolutionised our understanding of how our continents formed.
Sometimes even the tiniest of rock fragments can hide big secrets. Our recent chemical analysis of African rocks has revealed that the continents we know today may have started to form more than a billion years earlier than was previously thought.
Earth formed about 4.6 billion years ago, from material in a giant molecular cloud called the solar nebula. Gravity caused this material to assemble into a sphere – Earth, with the densest forming the core, and the least dense forming the mantle. The crust and the upper part of the mantle – which together comprise the lithosphere – formed rigid plates, which move horizontally on top of the more malleable lower part of the mantle – the asthenosphere (figure 1).
The organisation of these plates has changed dramatically over time (figure 2). About 2.5 to 4 billion years ago – during what is known as the Archaean eon – the lithosphere was partitioned into plates much smaller than the continents we know today. Later, during the Proterozoic eon, the plates joined together, forming one large supercontinent called Rodinia. Traditionally, this is believed to have been the situation 1 billion years ago. Subsequently, the continents started to drift away from these masses, progressively forming the globe we now recognise. The break-up of Rodinia is referred to as modern-style plate tectonics and is traditionally thought to have started around 900 million years ago.
As this process occurs, plates collide. When one plate moves under the other and sinks into the mantle, it is called subduction (figure 1). Subduction is a slow process that happens at high pressure (about 10 kilobar) and a temperature of less than 500 °C, and with a thermal gradient of less than 15 °C per kilometre.
We didn’t set out to investigate plate tectonics, however. Instead, the purpose of our study was to use a new technique to learn more about the formation of metamorphic rocks about 2 billion years ago. We had not expected our work to have any implications for plate tectonics, which was generally thought to have started nearly 1 billion years later.
For the first stage of our study we visited several hundred geological sites around Africa (figure 3) and collected samples of greenstones. These rocks are known to have undergone metamorphosis – a change from one rock type to another – about 2 billion years ago. Based on previous knowledge about the metamorphism of rocks during this period, it was thought that they must have formed under conditions of low pressure (no more than 5 kbar) and temperatures ranging from 200 to 700 °C.
Next, we investigated the composition of minerals in these rock samples using microprobe analysis. This is a range of techniques that includes microscopy and back-scattered electron imaging, which distinguishes heavy elements, which scatter electrons well, from light elements that do not. We also performed chemical mapping, which shows where particular minerals are found in the samples.
Furthermore, we carried out experiments at the European Synchrotron Radiation Facility (ESRF; see box) to decipher the very fine chemical structure of some of our samples. Synchrotron X-ray beams are billions of times brighter than the beams produced by a hospital X-ray machine, allowing them to resolve the structure of matter at a level of detail impossible to reach with standard X-rays.
Using very thin slices of rock, we were able to map their chemical composition. We found that they contained quartz, garnet, phengite, chlorite and iron oxides (figures 4 and 5). But what did this tell us about how the rocks formed and under what conditions?
To interpret our results, we used computer calculations based on different chemical parameters that we measured. For example, we analysed the ratio of H2O to CO2 in the fluids trapped within the quartz, and measured the ratio of Fe3+ to Fe2+ present in the rocks (figure 5). There are many different chlorites (e.g. magnesium chlorite, iron chlorite) and several different forms of phengite (which may contain, for example, magnesium or iron). The precise chlorites and phengites that we observe in the metamorphic rocks depend on the conditions at the time of rock formation. These are the H2O:CO2 and Fe3+:Fe2+ ratios as well as on the pressure and temperature. Measuring the ratios of these different chemicals in our rock samples therefore allows us to work backwards to calculate exactly the temperature and pressure conditions under which the rocks formed.
Using these calculations, we demonstrated that the chlorite and phengite composition in the rocks of western Africa was obtained under high pressure (about 10 kbar) and a low temperature of less than 500 °C. This was surprising, because these pressure and temperature conditions are found only in subduction zones. Since the rocks we studied date back more than 2 billion years, our results imply that modern-style plate tectonics existed 2 billion years ago, far earlier than the 900 million years ago that scientists had previously thought.
Our discovery has changed the scientific understanding of the geodynamics of Earth. So when, then, did modern-style plate tectonics actually begin? And how widespread were these gigantic land movements? To address these questions, our next step will be to study other rocks of the same age and older. In particular, we plan to visit Yilgarn Craton in Australia and the Barberton area in South Africa, to examine their chlorite- and phengite-containing metamorphic rocks.
The European Synchrotron Radiation Facility (ESRFw1) is one of the most intense sources of X-rays in the world. Thousands of scientists come every year to ESRF to carry out experiments in materials science, biology, medicine, physics, chemistry, palaeontology and cultural heritage. ESRF is a member of EIROforumw2, the publisher of Science in School.
The authors would like to acknowledge the help of Dominique Cornuéjols, from ESRF’s communications department, in preparing and translating material for this article.