Siliceous and Delicious

Spent the morning on the FIB with the engineer in charge, Nicholas Antoniou. He was very helpful and trained me in how to stigmate, focus, and align apertures on the various ion probes in the FIB. It’s a monumentally finicky process, but I think I get the basic picture of what needs to be done now. However, once we’d gone through this exercise, it became clear in a) a couple of trial milling operations we carried out using the now perfectly-tuned ion beams, and b) the conversation we had throughout, that using the higher-current beams (which would reduce the milling time from many—as many as dozens of—hours per specimen to a slightly more reasonable hour or so) is simply not an option. No matter how well adjusted the beams, the 23 and 45 nA probes produce an ion beam that is simply not sharp enough to cut open the frustule cleanly—the result is an unevenly cut frustule caked in ‘curtaining’, step-shaped deposits, and other milling artifacts.

This, needless to say, was a heavy blow. The probe sizes which gave halfway decent results (delivering 6.5 and 13 nA of current) will take countless hours (on the order of a day or two) to cut through the larger frustules I’m currently working with (the Cretaceous sample I’m looking at at the moment, for example, has frustules in the 100 µm length range). It’s simply beyond the size range of what the FIB is designed to efficiently handle (it’s really set up to work with nm- to few-µm size features on electronic chip circuits).

There was some moderately good news, though it was hard for me to really get excited about it. Nicholas talked at the end about the possibility of setting up automated milling operations overnight. This approach would involve milling registration marks (“X marks the spot”!) next to frustules to be milled, which are then sequentially found using pattern recognition in the course of a night’s automated milling. The FIB locates the desired feature on the stub and aligns it using the registration marks, then carries out a pre-set milling operation, then moves on to the next registration mark, and so on. Theoretically this would make very long milling operations (at least on the order of several hours) more feasible.

The amount of time this implies, though, is still staggering. Assuming I can mill 2 or 3 specimens overnight, this will still require on the order of 100 nights of milling to generate measurements for the 50% taxonomic coverage case I suggested in my thesis proposal. This is patently ridiculous, in spite of being a probably wildly over-optimistic assessment of how much time it’s actually going to take to make these measurements.

I’m distraught and demotivated. Nothing seems to want to work at the moment. For better or worse, we have an early lab meeting this afternoon—a dull paper about some early Proterozoic pyrite concretions that are suggested to be biological, maybe even multicellular and (heresy!) eukaryotic in origin. In any case, daft and dreary as it may be, I’m going to have to spend the remainder of the day reading the assigned papers, so I don’t look like any more of a fool in front of my advisor. Yech.