Developing a system for the measurements of solid and melt’s density under mantle conditions using neutron tomography

Accurate density measurements of melts and solids under pressure are crucial for the correct interpretation of seismic tomography data allowing a better understanding of Earth interior. Melt density controls magma migration and viscosity which is an essential prerequisite to understanding Earth evolution and dynamics. Also, a knowledge of density as a function of pressure and temperature is the minimum information necessary to conduct atomistic simulations, such as molecular dynamics and Monte Carlo, to understand melt atomic structure. Containing melts and measuring density at elevated pressure and temperature is difficult since melts are very reactive, and suitable containment capsules, such as platinum, are very absorbing of X-rays. Using a large volume cell and neutron radiography or tomography seems a possible way forward. The fundamental parameters necessary for a successful system were explored using the DINGO neutron tomography instrument at the Australian Center for Neutron Scattering, Australian Nuclear Science and Technology Organisation at Lucas Heights, as a test bed. Four system design aspects were considered: design of sample containment assembly, optimised image processing methodology, optimised instrument configuration and design of pressure cell. A combination of finite element analysis and experimental measurements were used to explore the design parameter space. An available Rotational Paris-Edinburgh cell (roPEC) was used for the experimental component. Sample containment assembly: To use the roPEC at the tomography beamline to explore the usefulness of this approach, a sample assembly that can be used with conical-shaped anvils that is largely transparent to neutrons and will achieve the pressures and temperatures of the upper mantle needed to be developed. I developed two assemblies: the first can achieve 3.2 GPa and 750 °C and contain four samples with a volume of 4 mm3 each, the second can achieve 4 GPa and 900 °C and has a single sample volume of 11 mm3. Optimised image processing: I determined the relationship between grey values in radiographs and tomography and density of the sample and assembly components, allowing an estimation of the minimum density change measurable with this system. Optimised instrument configuration: A complete characterisation of the tomography beamline was necessary to ensure optimal configuration during density analysis measurement and to allow an assessment of the expected image resolution, which will allow optimisation of sample size. I concluded that the best setup for the tomography studies using a high-pressure cell is using the Iris 15 CCD camera with a 100 mm lens, high-intensity beam mode flux, the 6LiF/ZnS:Ag scintillator screen, and an exposure time between 4 and 8 seconds. Design of pressure cell: The preceding work demonstrated that the traditional four post Paris-Edinburgh cell design will not work for tomographic studies since too few radiographic images can be collected due to the post obscuring the neutron beam. My solution is to replace the four posts with a cylinder, thus distributing the material evenly around the sample and allowing sufficient neutron transmission. I used finite element analysis to calculate the von Mises stress distribution within a Cylindrical Paris-Edinburgh cell frame structure. The results show that for a cylinder with a 3 cm thick wall of 819 WA steel, the material easily sustains maximum stresses of 474 MPa, equivalent to 220 tonnes. If the cylinder is exposed to these conditions for 12 hours, its safety factor is 3.16. The overall result of this work is a design for a system that will allow melt density measurements at pressure and temperature using neutron tomography with a density resolution of 0.03 g/cc, which equates to a pressure resolution when using a liquid indium internal standard of 0.1 GPa.