4D quantification of micro-scale feedbacks in dehydrating, deforming rocks

This research project uses a novel methodological approach to determine where mineral dehydration reactions can trigger failure in deforming rocks. This link between dehydration and failure is important at convergent plate boundaries. Where plates collide, the shallow portions of the Earth's crust are affected by so-called thin-skinned tectonics. There, dehydration reactions enable the emplacement of tectonic nappes, which shape mountain belts such as the Swiss Jura, or the Appalachians in the US. Plate collision also leads to the subduction of tectonic plates, where dehydration reactions are suspected to trigger seismic events at depths of several tens of kilometers. In both tectonic settings hydrous minerals in rocks become unstable as temperature increases. They start to transform into denser minerals by releasing water in dehydration reactions. The density increase produces pores, which are filled by the water. The pores, the fluid pressure in them, and the newly grown minerals weaken the reacting rock mechanically. It may become unable to support tectonic stresses and fail.

The processes that control large-scale tectonics start at the grain scale. These grain scale processes entail a series of complicated, intertwined developments that involve the chemistry, hydraulics and mechanics of a dehydrating rock. Coupled chemical, hydraulic and mechanical processes may facilitate the self-organization of the dehydrating rock into a state where it ultimately fails. Unfortunately, neither classical laboratory experiments nor field-based studies allow a spatial and temporal (4D) characterization of these coupled processes on the micro-scale. Models to explain failure in dehydrating rocks therefore lack a robust observational basis.

We will use a unique combination of new methods to overcome this severe limitation. Our interdisciplinary team of experienced researchers will establish a technique to directly observe dehydration reactions in deforming rocks. We will employ the most powerful x-ray sources in the UK and Switzerland to observe dehydration reactions in a new generation of experimental pressure vessels. These vessels are transparent to x-rays and allow us to reproduce conditions at the base of tectonic nappes and at intermediate depths in subduction zones. They are designed and built in Edinburgh. Combining these vessels with time-resolved (4D) x-ray microtomography will enable us to document mineral dehydration at a wide range of conditions. The resulting 4D microtomography data sets will have a volume of several tens of TB. New analysis techniques based on machine learning will allow us to extract the relevant information from these vast quantities of data. Our analyses will determine conditions where dehydration causes rocks to become unable to support tectonic stresses. Using these analyses, we will test and advance theoretical concepts used to link dehydration and deformation in numerical simulations.

The first direct observation of the complex grain-scale developments during dehydration reactions will significantly advance our understanding of some key processes in tectonics. Because our data are time-resolved and dynamic, they will support the interpretation of field data that otherwise capture a static, fossilized picture of dehydration reactions. Our data will allow testing and refining existing mathematical models that provide a foundation for robust simulations of large-scale tectonic processes. Ultimately, our findings will support the assessment of risks associated with plate collision. Our project will also make a new experimental imaging method available for research on geothermal energy, CO2 sequestration and nuclear waste storage. The method combines time-resolved x-ray microtomography in our new experimental vessels with advanced data mining and image analysis and computational simulation.

Our technology-led proposal will directly benefit industry and governmental stakeholders in their R&D. We will establish a new experimental technique at the Diamond Light Source that is uniquely suited to observe and document fluid-rock interaction at the grain scale. Such processes are at the core of geothermal heat extraction, CO2 sequestration and nuclear waste storage. The chemical-hydraulic-mechanical interactions involved are further relevant for applied concrete research and the processing of industrial minerals. Time-resolved (4D) in-situ microtomography experiments provide unprecedented insights into grain-scale developments that impact on these operations on larger scales. The legacy cell will be transferred to Diamond at the end of our project, with staff being trained in its use during our project-specific work at Diamond.

Industry will also benefit from the publications of the technical drawings of our cell. As we have done with other rigs previously [Refs. 4,7], we will do this to encourage reproduction and improvement by third parties once the final design is tried and tested. In this way, our x-ray transparent fluid-rock interaction cell "Sleipnir" was copied and is now used by an oil & gas research centre in Pau (France) and by an applied research group at Stanford University (USA), and available to general users at the Advanced Photon Source (USA).
Data analysis is generally the bottle neck in 4D x-ray imaging, mostly due to the significant volumes of data involved and the variety of codes available for different tasks. At Diamond, we will also establish a new data acquisition protocol that reduces the amount of waste data and thereby facilitates data processing and analysis. We will also devise a new data processing and analysis framework that combines and optimizes a range of available codes and integrates novel data mining algorithms. All of these codes are open source. Towards the end of our project, once tried and tested, this framework will be freely shared on Github. Both actions will reduce the time from experiment to result and support the uptake of 4D x-ray microtomography by a wider R&D community.

The general public will benefit from our outreach activities during the project. In our interdisciplinary research we will utilize one of the UK's flagship research facilities to conduct pioneering and challenging experimental work that will push limits both on the geosciences side, the synchrotron imaging side and the data analysis end. During the project, ongoing work will be disseminated through social media and a project-specific blog on our departmental website. We will further embed 4th undergraduate students on our "Geoscience Outreach" course in our synchrotron campaigns, where they will a) experience cutting edge science in action first hand and b) communicate their experience through their outreach projects. This can involve organising sessions with school children to highlight UK research and encourage careers in science, entertaining blogs and webpages and feeding into Facebook and Twitter.
We will show our science output at Diamond's open days. These attract several thousand visitors every year. The principal ideas and achievements and their relevance will be shown in
displays in Diamond House, where one or two of our researchers will be present to answer questions before and after visitors go on tours through the synchrotron.
As a source of information that lasts beyond the duration of the project, we will establish two displays for the Cockburn Museum at the Grant Institute of Geology (Edinburgh). These will 1) detail the interdisciplinary methodological approach of our research project, which spans geosciences, informatics and engineering, and 2) illustrate the most important research data and results for highlight the significance of micro-scale processes for plate tectonics. During "Doors Open" days, these displays will be staffed and presented to the public.