Interesting. I feel like this should be straightforward to replicate in a lab, though.