The "Ideal" Thesis

I attended my collaborator's defense this past Wednesday, which was an incredible piece of work performed over the short course of three years. While the thesis itself, 200-pages, was indicative of the considerable work done by him over the last few years, I think more impressive was still the nature of the work.

Two of the experiments were performed in collaboration when I visited France and another when he visited Pittsburgh. While some of the other experiments were not performed by him at all, but by former students but all the analysis was new and performed by him solely still in relation to his project. Certainly  the experimental design, execution of the experiment, and data collection takes a large amount of time, but analyzing the data afterwards is just as much of a feat afterwards as well.  In particular, for the nature of our data, orientation maps obtained through electron backscatter diffraction, primarily contain grain size, grain orientation, and grain boundary misorientation information immediately. However many other things can still be analyzed on a local level (especially in our in-situ experiments), such as observing grains nucleating from the areas of highest kernal average misorientation, indicative of concentrated dislocations in an area, or the roughness of the migrating grain boundaries.

But in addition to the experiments, he still managed to perform simulation experiments as well by introducing anisotropy (in particular large variations in anisotropy for a twin boundary) in level-set and phase field methods. After achieving this, this was furthermore implemented into a microstructure featuring twins under grain growth, and observe the evolution of the microstructure and compared to observations made from the experiments performed. Anyone who has worked with computational materials science understands these developments are quite time consuming.

So finally one impressive achievement from his work is that he did a balanced amount of simulation and experimental work that complimented one another. It is often more likely to see thesis that are completely focused on experimental work, with a touch of computational work to support the experimental work performed, or vice versa a computational thesis with some experimental work to back up the results of the simulation.

To achieve both in such a short amount of time (3 years), is an absolutely amazing feat.