An ideal collaboration

I just had about the best possible experience with a collaboration that one can expect to have. Indeed, I worry that I've now used up my "collaboration karma". Here's how these things are supposed to work....

Back at the APS March Meeting in 2006, my student Zach was presenting his masters work on electronic conduction through atomic-scale Ni junctions. Specifically, he had been doing some experiments to try and examine whether atomic-scale contacts between ferromagnetic metals had unusually large changes in electrical resistance when placed in a changing magnetic field, as had been reported in the literature. (We found that the answer is "No", but the magnetoresistance does depend in detail on the precise atomic configuration of the device. This work was independently confirmed simultaneously by Dan Ralph's group at Cornell.) Anyway, at the end of the session, I met Carlos Untiedt, who was just getting going as a faculty member at the University of Alicante in Spain. I'd read some of Carlos' earlier work on metal junctions made using mechanical means, and he'd read our work, too. He mentioned to me that the Spanish government has a program that allows Spanish graduate students to spend time abroad working in other labs, and suggested that we try this at some point. I said that this sounded like a good idea, and we should do it.

Fast forward to the beginning of 2008, when Carlos and I got back in touch. He had a very good student eager and interested to come and visit, and, even better, they had some exciting data that they'd been taking in mechanically-controlled (STM-style, middle of this page) atomic-scale metal junctions. The main advantage of mechanical junctions is that you can break and re-form them many times, giving you serious statistical information about junction properties. Now, in my lab we often use an alternative technique for making atomic-scale junctions that doesn't involve mechanical motion. While our method (electromigration) is more time-consuming and therefore not well suited to really large statistical samples, it has one main advantage: the junctions we make have enough geometric stability that we can look at a single junction over many temperatures. This can't really be done in STM-style junctions. This was a relatively rare situation: there was an ideal point of scientific collaboration, and we had the person and the resources to make things happen.

So, we did it. Carlos' student, M. Reyes Calvo, came and spent a little under four months working in my lab with my group. She was able to make junctions with our approach that were analogous to the ones that she'd been studying in Spain, and measured them as a function of temperature in our system. The results were very nicely consistent with her data from Spain, and the whole scientific story hung together well. After her visit and a number of fun conversations with theorist colleagues at Alicante, a paper was written that came out today in Nature. It just doesn't work any better than that. I'll write about the science in a separate post....