The other thing that struck me about bait-and-switch was this

I gravitated toward a scientific life with fantasies of sci-fi movies running through my head, with large machines emitting lightning at the flip of a huge Frankenstein-type switch, or several people poring over softly-glowing computer screens as an experiment produces fantastic data in real-time, and great discoveries are made. I thought this kind of thing actually happened even as I started grad school (even if I had never seen it in my various research summers…)

It doesn’t happen often, but it does happen (depending on your definition of “great”). Back in my first postdoc, at TRIUMF, we trapped radioactive potassium atoms for nuclear-decay tests of the standard model. Or, more precisely, we planned to do this, since the research had progressed only to the point where stable potassium had been trapped when I started working there. Not too long after my arrival we were scheduled for a few stretches of beam time, with an appropriate target to produce the radioactive isotopes we were trying to trap.

Since these were radioactive isotopes, the exact frequency for trapping them was unknown, though the presence of stable isotopes meant (in principle) that the isotope shift could be calculated to some degree of accuracy and narrow down the range of frequencies for the trapping and repump interactions. Since the linewidth of the transition is somewhere around 5 MHz, and you should be able to see a trap with a laser detuning of somewhere between a half a linewidth and several linewidths to the red of resonance, we set up to scan in discrete steps of several MHz, pausing at each step to look for fluorescence at the center of the trapping region — literally looking: we integrated the output from a CCD camera and displayed it on a computer screen, along with a graph of the total fluorescence.

The scan was done by stepping an acousto-optic modulator (AOM) arranged in a double-pass. AOMs shift the light as it passes through, but by reflecting the beam and sending it in the opposite direction, you retrace the path and un-do the physical shift of the beam: the light will take (to first order) the same path no matter what your frequency setting is. The one additional trick is a quarter-wave plate after the first pass, which becomes a half-wave rotation upon retroreflection, so the polarization changes 90º and is transmitted through the polarizing beamsplitter that had originally reflected light into that path.

We could scan around 80 MHz in this fashion, and then re-tune the other AOMs as necessary to expand our scanning range. The first attempt was unsuccessful; it turns out that two people had done the isotope shift calculation and made the same math error, so we were searching in the wrong area in frequency-land, and only found the shrieking eels. The calculations were re-done and we set sail again. After a couple days, we still had bupkus to show for our efforts, and we were tired. Beamtime at an accelerator is 24/7, with occasional scheduled shutdowns (and some unscheduled ones as well), and at that time there were only two of us in the group with laser/optics experience — me, because I was an atomic physics guy, and the other postdoc, because he had to learn all that stuff. The students were nuclear folks of various flavors, and we all pulled 12-hour shifts during the experiment.

It was the last day of running, and the last shift. The other postdoc refused to go home at this point, because of the pressure we were under to trap the atoms. Finally on one scan, there was a blip on the graph and a circular blob on the monitor, distinct from the background, though my colleague was so exhausted he didn’t see it. When the scan was done, he started to get ready to set up for another frequency range, but before I could ask him why, he announce he was going to go get something to drink and try and wake up. So I re-did the scan and stopped it when the signal reappeared. My colleague returned to the glow of a ball of trapped K-37 on the monitor. We spent the last few hours of the shift doing more precise scans and quick code update to allow for a screen capture so we could show off our bundle of joy to other people in the research group.

But it was real-time discovery, with a softly-glowing computer monitor. We even had a Frankenstein knife-switch in the lab, though we never threw it — it was used as a DC bus bar. So even though it’s not the norm, that scene does play out occasionally.

(my first reaction to the bait-and-switch post)