All coils blazing
The main Tesla coils pages are very focused on theory/construction and don't include much lightning pictures. So here's a page that does.
My very first encounter with voltages higher than mains happened just after our NST were delivered, a 9kV 25mA and a 6kV 25mA model. The bigger one was used with the Zeus coil, and the other one eventually found use in the Hyperion coil.
Maël Flament and I then tried all sorts of manipulations with our new 6kV transformer.
We put a paper sheet between the electrodes of our tiny spark gap at some point, to see if the arc would be arc to pierce it. During the very early stages of Zeus' construction we briefly though about using paper as the dielectric for the primary caps. Seemingly a bad idea ; low-density polyethylene (LDPE) is a far better choice for many reasons.
We also tried to make a Jacob ladder out of a pair of aluminium electrodes. The results were somewhat disappointing. We then added salt, hoping the heat would separated the Na+ and the Cl- and make the whole system more conducting. It worked ; the arc could reach the top of the ladder and remain alive for several seconds. It also got a vivid yellow colour (famous sodium doublet ?).
Now we're entering the heart of the matter. Let's start with Zeus. These four awesome pictures were taken by Jean-Louis Colot with a Canon EOS 5D. And you really should see them in high resolution.
To catch Tesla coil sparks, one should work with exposure times of about 1s. Longer times (about 3s) can be used to give a hair-like appearance to the sparks, while shorter ones (down to 1/50s) should be able to catch single strikes. The f-ratio should be as small as possible to have as much light as possible. ISO sensitivity is more problematic as spark pictures will often involve a lot of dark zones ; one should proceed by trial-and-error to find the best balance between sensitivity and noise. The use of a tripod is mandatory. I also recommend using a self-timer (about 10s) to allow vibrations to dissipate if the photograph is taken manually.
One can also use some external lightning, like in the photographs below. For instance, in the right picture, the coil is lit by a neon tube. Actually, it is interesting to note that the tube is glowing because of the electrical field gradient generated by the coil, a case of (tiny) wireless energy transfer. The left picture was taken with a flash.
The Zeus coil, photographed by Jean-Louis Colot. Left: f/2.8, 800 ISO, 1s. Right: same with 1600 ISO.
I don't know where do the weird dark bands come from, probably camera firmware.
The following pictures were taken by me with a Canon EOS 1000D, a much cheaper model, but still quite good. The following pictures were taken looking at Hyperion, with notably longer exposure times. Second and third picture are head-on shots, looking straight at the breaker. Last picture is a close-up on the filaments.
All pictures taken at f/5.6.
Upper left: 400 ISO, 3.2s. Upper right: 200 ISO, 3.2s.
Lower left: 400 ISO, 3.2s. Lower right: 400 ISO, 2.5s.
The light from the sparks we see come from the de-excitation of molecules present in the air (mainly nitrogen and oxygen), which have been excited in the first place by the very high voltage generated at the coil. It is a well-known quantum-mechanical fact that the energy levels an electron bond to an atom can occupy are quantified. Therefore we expect the light coming from these sparks to show a discrete spectrum.
This is indeed what we see when we split light into its components using a diffraction grating. The seemingly continuous background on which the brightest lines seems to lie are actually a due to a multitude of very closely-spaced levels. Trying to identify the main lines would prove very difficult in this crude setup.