Ugh… I’m just going to cut this one immediately, so anyone who doesn’t want to hear me whining about wavelength calibration can go on about their business!
The biggest pain in any kind of science is that arduous bit in the middle. Planning stuff out is fun, writing the proposal to figure out what you might find, being excited about possibly getting to use cool equipment to do interesting science with interesting stars (or molecules, or laser beams, or whatever it is you’re using this time). Likewise, having a set of nice clean data to analyse and pick apart is so much fun. Combined with the added exhiliaration of the fact that you did this. These data with all their mysteries and unidentified features? These are your data. There are many like them, but these ones are yours.
Ok, right. But the bit in the middle? That’s just a headache. Sometimes literally. Taking raw data and putting it into nice plots is troublesome. Calibration sometimes moreso. But anyone who’s ever heard me grumbling about this (as I do, from time to time) will know that my least favourite part of reducing astronomical data is wavelength calibration. Because it’s a pain. Chemists, take note. The way you can put a sample into a spectrometer and get a set of wavelengths out? Well, that’s because the spectrometer’s nicely calibrated, and you don’t normally need to mess with that. With astronomical data, for reasons I don’t care to go into right now, you have to calibrate the wavelengths every time. At least you do with optical data, I don’t know so much about other wavelengths.
So how does that work? Basically, you take a spectrum of an arc lamp each time you take a spectrum of an astronomical object, using the same spectrometer settings. Then you can check the reference spectrum (always something with a well known spectrum, like Argon for instance) and use it to calibrate your astronomical spectrum. Which, honestly, sounds like a really trivial thing to do. And it can be. But it can also be… troublesome.
For instance, these data I’m tinkering with right now have used a Thorium-Argon arc lamp as a reference. From the instrument manual, the spectrum of the Th-Ar lamp used looks like this:
That region I’ve outlined in red is the region I’m looking at. So that’s not hugely helpful. Thankfully, there are other spectra out there. The same observatory has nice convenient spectra (in fits format) of their arc lamps available to download. So the specific patch of the spectrum I’m looking at looks like this:
Looks pretty messy, huh? Yeah, both Thorium and Argon contain plenty of electrons. That’s a lot of little quanta of energy bouncing about in those atoms, creating all kinds of lines. Well, this still shouldn’t be so bad. So, I thought, let’s get an independently recorded spectrum. Just to see which are the lines I should be paying attention to. So I had a look, and I found this:
Ok, that looks good. So if I ignore the mess near the baseline of the spectrum, the main peaks are easily picked out. Cool. Ok. Well in that case, it should be easy, right? Let’s take a look at my arc spectrum and assign some lines! Now where did I put my arc spectrum… Ah, here it is:
(And that big peak in the middle isn’t even from the arc lamp. It’s a cosmic ray hit which, by a freakish coincidence, just happened to be in the wrong place.)
It’s done now, in any case. I just felt I needed to rant a bit. Sorry about that. I hope it was educational, at least. Ok, I’m going to go and eat chocolate now…