We have interests in a variety of characterisation methods and are developing a series of toolboxes to help materials scientists get the most from their data.
Current interests and expertise include:
Toolboxes for characterisation
In order to understand the properties of materials there are three distinct areas that need consideration: (1) What is the state of the material (it's structure, defects, grains size etc), (2) what properties does it have (it's mechanical, electrical properties which result from its state), and (3) can we predict its properties?
Of these, defining the state of a material by Materials Characterisation methods is integral to understanding and producing better materials. Because, without good characterisation we are not able to apply the results of mechanical (or other properties) tests to other materials, and we cannot evaluate the effectiveness of material models.
The problem though is that the property of a material is dependent on a number of different microstructural features. Hence, in order to understand the material requires measurements using different test techniques that can give information about different features and also the same features at different length scales. For example, we may want to be able to characterise the dislocations (defects that control plastic deformation) in a sample. We can view these dislocations by using Transmission Electron Microscopes (TEM). The problem with this technique is we only see a tiny proportion of a sample, but dislocations can change significantly within the same grain and more so within other grains in the sample. Alternatively, we could measure multiple regions using TEM, or use indirect methods that give information about the dislocations in multiple grains, e.g dppa, hardness measurements, ebsd.
Hence, characterising a material and how this effects its macro properties, such as the dislocations present, is non-trivial. We can either view the microstructural details directly but have the problem of not knowing how representative this is of the sample, or we can get information of the bulk sample but have the problem of what assumptions to make so that we are accurately representing the sample. The best approach would seem to be a combination of both types of technique. But...
The human mind is a better lawyer than it is a scientist, so effectively we look for trends to justify our beliefs. And this is problematic when interpreting materials science data. Whatever method we employ to characterise a sample, multiple direct methods, one indirect method or a combination of both, we will tend to interpret the results to meet our expectations. And the problem is in most cases we don't know we are doing this. There is no simple way of getting out of this paradigm, but we can try our best by being aware of its existence, using multiple techniques and combining this with modelling. But most of all being humble and trying to discuss the limitations of our results and understanding is really the best we can do.
Thomas H. Simm - "My goal is to enhance characterisation techniques, explore what works and what doesn't and help others get the most from their data"
@ All Rights Reserved