In the modern electron microscope, for some materials (typically metals), we can form atomic resolution images relatively easily and and with sufficient clarity to understand atomic positions, bonding and defects to in how the atoms are packed with a crystalline ordering. These atomic scale ordering defects can lead to large scale changes to the properties of the material (such as charge-transfer mechanisms, magnetic structures or strength) - so understanding the atomic structure and ordering of materials is key to understanding global scale properties of materials.
However, many societally important materials (such as photovoltaics - used in solar cells, or battery materials, or pharmacuetical materials) cannot be imaged in this way - the electron beam used to form the images instead damages the sample (damage mechanisms can include heating, localised charging, knocking atoms out of the sample, amongst others). These processes prevent clear images being obtained by using standard electron microscopy imaging protocols. Lacking clear images of these materials hampers our understanding of how their structure and properties are correlated. My overall research objective is to lift this bottleneck, and enable high-resolution, clear images of all materials science samples.
The societal value of this fundamental research comes through in the applications it enables - clearer images of prototypical photovoltaics will enable a more rapid development of efficient sola panels, clearer images of new battery materials will accelerate our understanding of why different options fail to cycle successfully - and so on - materials science progress is built upon progress of accurate materials characterisation.
In this fellowship, I studied a family of algorithms referred to as "ptychography", investigating the potential of these algorithms to enable us to form higher resolution images with a lower electron dose. Ptychographic algorithms (as applied in the scanning transmission electron microscope) make use of every scattered electron that lands on the detector in a diffraction pattern - analysing the position it is scattered to, and how these positions vary with the position of the incoming focussed electron beam can tell us more about the sample, at a higher resolution, than can be interpreted from the conventional bright-field, or dark-field imaging datasets.
Early in this research, it was realised that we lacked verification that these methods can be applied to thick samples, as these scatter the electron beam more than once as it passes through the sample, in a manner not accounted for in the mathematics underlying the algorithm - verifying that these samples could be imaged by using electron ptychography, and verifying the limits of this, became the main focus of this project as the first hurdle to overcome towards the greater goal.
