Maximum IOR?

[Ihno]

Hi,
I'm wondering for quite some time now what range for IOR is physically plausible?
I think it has to be higher than 1 but what is the max?
2, 5 10?

F.e. I'm creating a guitar paint that has a clear coat on top. The clear code looks pretty close with an IOR of 2.5 and a reflection amount of 1.
I could, on the other hand get a pretty similar result with an IOR of 3 or more and a darker reflection amount.
Sure, the transition from 0° to 90° and the value at 90° would be different but it's sometimes hard to tell because of unknown light sources and unknown general lighting in the reference images. I think knowing the IOR would unlikely be more than 3 f.e. would help to get a range to play with.

After some measures, I've seen this strange valley emerge. Starting at an IOR of 4.
Is that where it stops to be realistic?

IOR: 2


IOR: 4


IOR: 10

[^Lele^]

I think it has to be higher than 1 but what is the max?

For refraction, there is no theoretical maximum IoR, as one can -theoretically- slow light down infinitely, before stopping it completely.
In essence, as n = c/v (n is the IoR, c the speed of light, v the phase velocity of light in the medium), so if v becomes very small, n will become very big (in math wording, for the limit of v going towards 0, n becomes plus infinity.).
In practice, periodic table elements go to about 4.0 (so yes, you're right. Although what you see *may* be an issue with the maths model used. I wouldn't know for sure.), while metamaterials can go to about 40 and exotic states of matter in the lab can slow light to 15 km/h (down from 300.000 km/s!), making the IoR some insanely high number (~72,000,000).
There are also cases of IoR < 1 and even of negative IoRs, albeit not so often in the visible spectrum, and these happen because the speed at which the phase (ie. the position of crests in lightwaves) changes isn't bound to that of light.

For reflection, complex IoR in metals (with the n/k coefficient pair, and a slightly different equation than the one above) will be between 0 and 1, and will quite strongly vary with the wavelength they're subjected to.
Which is why places like refractiveIndex.info show triplets, instead of single values.

As such, you're generally better off looking up a table of some sort with the values of IoR you're looking for.

p.s.: i ripped the links and some parts of the reply off this or the other website.

[Ihno]

Thanks for the answer! (I was hoping the graphs would attract you)
Reading that science is able to slow light down to a speed where one could overtake it with a bicycle was really mindblowing.

Refractive IOR isn't a big problem usually. Since the Internet is full of tables for those and in the daily work all you need is: water, glass, acrylic glass, oil and maybe diamond.

After reading your post I went to google and searched for a while for reflective IOR tables.
Again, I seem to do this a few times a year and I always endup not trusting any of those.
It increasingly gets interrupted by pbr evangelists saying: IOR for none metals shouldn't be touched at all, stay at 1.6. Since "For a smooth dielectric surface, F0 will reflect between 2-5%". Completely ignoring the fact that one could use a higher IOR and lower the overall reflection amount.
Also, there are some guys saying "compound materials like wood, stone, concrete etc should have 3-6" (and a low reflection amount)

So to sum it up: there seem to be 3 types of people(for opaque non-metal's):
The PBR "fuck it all I don't care about IOR nor the reflection amount" guy.
The "refl amount should always be white and you only edit the IOR" guy.
And the "edit both: what ever works, works. If 6 and a reflection amount of 0,12 looks right, go for it" guy.

I tend to agree with the last one but its all opinion and very unscientific.

I'm more than happy that its solved now for metals with the metalness Parameter and the table vlado posted at the Chaos Group Labs blog.

I think I'll measure the refractiveIndex.info data. At least I'll get an idea of the curve for some plastics.

[^Lele^]

Yeah, that's the thing with perception: many things will "work" for indvidual "A", and not for individual "B".
On the other hand, it's a true issue to get data off of nature: the n/k pairs WILL vary a lot depending on how they are measured, how the sample is (f.e. a thin sheet, versus a thick block), and so on and so forth.
You may try (for the fun) to measure some of the most recent (stress on most recent!) VRScan metals, ideally without the procedural clearcoat, and see what you get.
Then do the same with a wholly principled metal of the same type, and see what the differences may be.
Perhaps they aren't as big as made to be, perhaps they are. ^^