I do. I’m not sure why, but I get an odd amount of enjoyment of plotting data out and seeing different relationships displayed graphically. I know, I know, there’s no hope for me. I’m reminded of the Simpson’s episode where Homer has suddenly become smart (I believe from having a crayon removed from his brain) and is finding himself rather sad. Lisa mentions that happiness and intelligence are often inversely correlated and says, “sigh... look, I made a graph of it.”

Anyhow, here’s one of my latest little ones :

During our experiments it’s often necessary to keep track of the temperature accurately. However, it’s often quite difficult to get the conventional temperature measurement device, the ubiquitous thermocouple, to function in our system (usually it’s just impossible to use due to my heavy reliance on RF induction heating). However, we still need an accurate way to determine the temperature both for our results and also so that we don’t melt the crystals.

One of our standard tricks is to us the common piece of knowledge that (most) things expand when heated. Now, we can’t put calipers inside our chamber or do anything like actually measuring the size difference of these little crystals in the normal sense. However, because we’re so often doing x-ray scattering experiments, we are able to clearly determine the average distance between the individual atoms in the crystals. As the temperature increases we are capable of seeing the distance between the atoms change by a very tiny amount. The relation between that expansion and temperature is well known.

So we just measure those distances periodically and convert them to temperature. Now, it takes a little time to do that and time is a precious quantity at a light source. So we do it only when needed/reasonably possible and then interpolate for places in between. Above is that interpolation. The blue points are the measured vales along with the accompanying experimental uncertainty in each point. A red line is then fit to those data points (I excluded the room temperature-0 power data point as it looks like a non-linear jump in temperature from the heater just being turned on at its lowest power setting). So now you can get the temperature for any power seeing to within a very reasonable accuracy. We have to do that each time we change samples or even whenever there’s a large change in the sample position relative to the heating coils, but that’s the price we pay. The accuracy is much, much higher than the other temperature measurement device easily available, a pyrometer (which determines the temperature from the light being emitted by the glowing crystal).