Comet C/2017 K2 (PanSTARRS) is thought to have travelled in a hyberbolic orbit from the Oort cloud for some 3 million years.  The Pan-STARRS survey in Hawaii first observed the comet between the orbits of Saturn and Uranus, the farthest active inbound comet they’d seen. Now in the inner solar system, closest approach to Earth is on July 14 at 277 million km, and to the sun on December 19. My image was taken a little before closest approach, with the comet at magnitude 9.1. It seems the comet is large and probably as well very active. Estimates of the nucleus size range from 14 to 160 km, but clearly it is a particularly large nucleus, as the majority of comets have a nucleus of 1 to 3 km in diameter. And it developed a huge coma of 130,000 km diameter, a sphere that is about the size of Jupiter!  Early observations detected an incredibly large tail, about 800,000 km long. Given the distance from the sun (especially when first observed) it would not be expected to heat up enough to sublimate huge amounts of ice. This suggests the comet's strong activity may be driven by a mix of ices with substances like nitrogen, carbon dioxide, carbon monoxide and molecular oxygen. 

Regarding the characteric green glow of comets, this is a result of the breakdown of a reactive molecule dicarbon (C2). This molecule absorbs a photon of light to excite it, then a second photon to excite it further into an unstable configuration that  decays, radiating light. In the green there is a strong peak for emission that starts about 490 nm increasing to a peak at 514 nm,  see C2 in Hale-Bopp and C2 in Mitaki and Emission in NEOWISE

I imaged the comet using the L-eNhance (HO) duoband filter whose OIII band is centered at 500 nm with a bandwith (1/2max) of ± 12 nm, thus most of the green emission from C2 is reasonably well captured, though the L-eNhance transmission is falling off at the peak (note there is little green emission above the 514 nm peak, which is very asymmetric). Thus with my horrible light pollution it is very beneficial to use this NB filter. And indeed, the characteristic green glow as very strong in the coma, though the tail was dim--it is know that cometary tails generally have little C2 emission as the lifetime of C2 is short. 

The image was integrated to 2h 53m total exposure, using 20 s subs to enable no image trailing in the comet. This required 3 nights (June 23, 24 and 27th) to get sufficient integration from my location, as the comet is not visible for long each night. Each nights subs were calibrated, registered and median stacked separately in DSS using the comet as reference for the stack (each night needed to be separate as DSS won't comet stack without overlapping stars). Then each night was processed similarly in Startools. Similarly, the stars were stacked in DSS for the 40 of the 20s frames at the end of the first night, then processed in Startools. All three processed comet images were averaged in Photoshop layers, and aligned and added on top of the star layer using a mix of screen and soft light blends. Noise was addressed with NoiseXterminator as a PS plug-in, which I currently have on trial. It did an amazing job with noise-very, very impressive.