The weather got bad, and I couldn't do much due to clouds, so I decided to do some dark image experiments instead. I parked the scope so that it pointed towards the northern horizon and put the lens cap on, but kept the cooling set to -20C on. I wanted to understand how many darks that are actually needed for a good dark image calibration of my camera. I have a camera based on the popular Sony  IMX571 CMOS sensor. Since I normally use 5 minute subs I set it to do 60 such images without any light, thus a total of 5 hours of just dark nothing. At least, I thought it would be just dark nothing.

I viewed the sequence as a movie in the PixInsight tool 'blink'. There was a lot of things going on in this movie. The random noise dominated the individual frames and behind that was a stationary pattern which I interpret as the Dark Signal Non Uniformity (DSNU) of the sensor, basically the fixed pattern noise that I want to subtract from my image subs.  There were also blinking pixels, which I interpret as Random Telegraph Noise (RTN). Clearly they will be present in my image subs as well and I have to think about what they can do to an image (hopefully solved by dithering). I could also see a faint amp glow in the amp glow free sensor. 

But, I was fascinated by the many transient features. I read about similar tests done by other people and found some CCD references. One of them was an Astrobin reference from May 2020, written by Bruce Rohrlach. He gives a long and interesting explanation. You can look it up if you are interested. 

It turns out that many of the phenomena are actually secondary particles from cosmic high energy particles, mostly protons, that are colliding with atoms in the atmosphere. The cascade of particles that comes out of such collisions contains muons, a kind of heavy electrons (about 200x more massive). They are the most likely particles that will survive all the way to the surface of the Earth, and apparently to my camera. Muons are really strange particles, they have a very short life time of about 2 micro seconds after which they decay into electrons and neutrinos. They only reach the ground due to relativistic effects.

When muons enter an active image sensor such as a CMOS or CCD sensor, they generate charges much like light would do, and since muons are very penetrating they will potentially move through many pixels before they continue on the other side of the sensor. They are not bending much do to fields or other obstacles in the sensor and  make straight lines that are long or short depending on entrance angle.

In my camera, looking at the horizon, a muon coming from zenith would create the longest tracks, since the sensor is standing up. A muon coming from the horizon would give a small or no  track. I found plenty of events in my experiment. I tried to count them and found about 10 events per hour. In this image I show one of the largest. I also show the frame before and after to prove that they are muon free.

So, my cloudy night turned out to be a super interesting lab in particle physics. I wonder where this cosmic ray came from? Maybe a distant supernova?  For imaging of astronomical objects they are of course not good news, so look out and do some muon cleaning of your darks before use!