The basis for the study of materials science and engineering is the discovery of structure-property relationships. The observation of structure at the atomic scale is critical for the field of materials on the nanoscale, and one instrument we use is the atom probe microscope. This technique routinely produces observations that are in the category of ‘big data’, with 100 million data points per experiment.
The challenge with big data is quantitative interpretation. Instead of impossibly trawling through each point manually, we must write software to interrogate the data points and their relationships with other data points. Automation has the advantage of consistency and reliability, but also has drawbacks of generating spurious analysis points.
Once we have a quantitative interpretation of the data, it is up to the materials researcher to engage our critical thinking skills to deduce the hidden structure–property relationships within the material. This requires careful application of existing knowledge considering the new data and analysis. It may also require additional computations such as density functional theory or Monte Carlo techniques, to enable us to determine these structure–property relationships.
We have recently recognised that owing to the highly detailed and technical nature of the analyses, our current quantitative interpretations need extra steps to correlate with the physical reality of the situation. Next in this project, we will reimagine our analysis tools so they are more consistent with the physical theories that have been established in materials science and engineering.