Reporting on the Second

The Atomic Physics course I'm taking has a serious lab component (35%). It consists of four 3-hour lab sessions with one set of experiments involving the Cs-133 hyperfine splitting, which are used to define the S.I. unit of second: 9192631770 periods of the radiation between the two hyperfine lines.

The experiment, as intended by the lecturer, was to be less undergraduate lab style and more real research style (we're using, according to the lecturer, research-grade lasers). There is no lab manual, no complete instructions, no fixed way of setting up the apparatus. Hell, we don't even get the exact specifications of the apparatus and have to make our own approximations (such as estimating the intensity at which the detector saturates and thus rendering it useless). However, it was quite fun, since it is not those find-this-line or record-this-number kind of experiments, and there's lots of fooling around exploring that we made. There was plenty of thrill and frustration (trying to align the laser, for example) involved.

So with the lab done and data collected, the first part is done. Which, however, was not so much for the data but for the experience of research, because everyone's data was made available for everyone else to use. The reason for this is that some groups had inferior apparatus and some were given flawed information, so it's only fair for everyone to have the better data. After all, this is the first time this course is ran this way, so some kinks has yet to be ironed out.

However, if grading is not so much on the data, then it has to be on the report. What's interesting is that the lecturer wanted the report to look more like a journal article than a standard undergraduate lab report. We are allowed to give it a spin of perspective, such as pretending that we're writing a paper on making cheap atomic clocks. Sounds fun, but it sure isn't easy. Right now, I've been throwing in crazy ideas such as linking gravitational wave detection to the hyperfine splitting lines of Cs-133, or encrypting messages using this methods of measurement. The former is pretty down the drain now, but the latter seem to hold quite some promise.

I'm open to any ideas if anyone has some. Somehow I've got this murky idea of dark matter hovering at the back of my head, but I think it'd be wise to abandon that idea. At the moment, I'll go with the encryption idea.

The report's due at the end of the semester (which, now looking at it, is a mere one month away), though the lecturer is pretty flexible with things and is pretty fine with a submission beyond that. Still, it's a wise move to start early, and, definitely, if the theme can't be fixed, we can start working on the data analysis itself first.

Only problem is, data analysis is one of the most dreaded part of any lab for me. Just imagine those wretched error bars... URGH!