The density of glass is normally proportional to the refraction, but you will find a wide variety of refractive materials, the key is normally given not by one material alone, it is the change from one to another. Like lenses work with “air-lenses”, which is nothing else than the space among lenses inside of lens groups e.g., for a camera.
To simulate that we just dial in a specific refraction value, which is based on the idea that the object is surrounded by air, even C4D has no Air in of course. In deep space you might consider the surrounding as pure Vaccum, which will not refract at all. having no practical experience in space ;o) I would assume that the lens refraction is stronger than.
When Light get bought in glass, at least to my knowledge, it has to do with the pureness of the glass, e.g, optical glass vs a cheap glass-water-bottle. The less pure the glass is “cooked” the less everything is melted to a singular “something” and light becomes refracted and reflected. This results then in the effect (i guess) you are after.
In C4D this effect is render intensive, and for animations you might consider twice to do that…
It is the Volume Caustic based on Light objects and Material settings. (You need to set up an material, the light source and the Render Settings>Effects>Caustic to get it.
If you compose later on anyway, you might create an smaller objects, perhaps several, as child and have them just reflective. Place illuminated objects for this reflection close to the light sources, perhaps with a gradient, so the reflections have some tonal qualities. Then render these passes (just reflection and illumination, no lights.) Dial the illumination above 100%. and render of course in 32bit/c floating point as file format.
As most Photographer know, the reflections of light sources work differently than light sources with its inverse square on diffuse surfaces, hence the tip to not use Light source for that pass. more about that in a longer series I’m working on this Summer. (Hint, knowing Photography in and out [not point and shot/automatic stuff] supports the success to be an excellent 3D artists)
If you have now 2 or 3 passes with different textured and fractured/structured surfaces, you can comp these in. Why 32bit/c, because the light inside the glass will be handled as highlights and not as white spots. If you render these in 8 or 16 bit/c integer you end up with clamped values and that is just un-usable (yes there are tricks with the fantastic FrishLuft plug-in in Ae, which might work, but if you ever have tasted HDRI in the pipeline, you don’t want to go back). If you put “smoke” behind, you need these super-bright values, especially for DOF and Motion blur., to separate these two form the smoke of course.
So, two methods, two different render times, heavily different render times, so my suggestion for animation is version two, compose it. Try it with a short test and play with the layers in your comp-app of course.
All the best