Fe-bearing minerals beyond density functional theory

Scientific problem

Traditional DFT techniques often fail in reproducing Iron-bearing minerals, for which we seek alternative methods. For this purpose we have successfully employed so called hybrid-functional DFT to scan the phase space of the mineral wüstite (FeO).

At zero pressure and temperature FeO crystallise in a rhombohedrally distorted B1-type of structure , as shown in the phase diagram of FeO (Fig. 1). Despite its simple crystal structure DFT wrongly predicts FeO being a metal. The aim of this work is to use a method, which correctly predicts the r-B1 structure of FeO being an insulator, and which allows us study all polymorphs in the phase diagram of FeO using one single technique

Such an investigation with the hybrid-functionals requires many independent calculations that are ideal to perform on the Condor-pools. With the SRB tool we can easily store and access the data for later analysis.

The structure and magenetic properties of troilite (FeS) have been investigated using density functional theory. Comparison of the unit cell and bulk modulus obtained with and without spin interpolation showed that inclusion of antiferromagnetic ordering of the Fe atoms is required to accurately simulate this system.

Scientific results

For the first time we have successfully determined bulk properties and phase transitions in wüstite applying the same technique to all investigated polymorhps independent on their magnetic structure.
In agreement with experiment we predict the rhombohedrally distorted B1-structure (r-B1) to be the energetically most stable structure at zero pressure and temperature. This structure has an anti-ferromagnetic (AFM) structure, which implies that the Fe²⁺ ion show a high-spin configuration. If we apply a homogenous pressure the r-B1 structure transforms into an i-B8(AFM) structure at pressures of ca. 80 GPa. Continuing to apply higher pressures we predict a second phase transition at around 145 GPa. At this transition our calcualtions suggest that the i-B8 strucuture transforms into a B8-structure, associated with a spin collaps on the Fe²⁺ ions. The high pressure structure is a non-magnetic structure.

We have shown that we have a technique which allow us to study Fe-bearing minerals, where traditional DFT techniques fail, which can be extended to study more complex Iron-bearing minerals.

In this work we make use of the following tools:

• Condor-cluster/Mini-Grid: To scan the phase diagram and to calculate a wide range of properites for FeO, employing more than 10 Hamiltonians, we make use of the computer clusters available within the mini-grid, which allow us to perform these calculations in months instead of years.
• SRB: To store the increasing amount of analysises generated from our calculations.
• PIG: To communicate and solve problems faster.
• Globus: Allow us to analyse and store all data on one computer, as well as and controlling and starting the calculations from one and the same machine, independt if the job is running at UCL, Cambridge etc.

Credits

This work was carried out by Maria Alfredsson, John Brodholt and David Price (UCL).