Where do Blender's dispersion IOR numbers actually come from?

Blender's Cycles renderer can fake dispersion (the way light splits into a rainbow through glass or a diamond) with three separate red/green/blue IOR values instead of one — but most tutorials just say "try 1.51, 1.52, 1.53" without explaining where those numbers should actually come from. I built a small set of materials that derive them properly: real Sellmeier/Cauchy dispersion curves from published optical references (Schott's own glass catalog for N-BK7 and N-SF11, F. Peter's 1923 measurements for diamond, Wang et al. 2013 for moissanite), converted to per-channel IOR the same way a physicist would (not guessed by eye).

Five materials are live and free on BlenderKit (684 downloads all-time, $0.01 earned in the last 30 days — the platform's own all-time total rounds down to $0, and this is a rounding artifact of a shared royalty pool, not a business). A $5 Gumroad pack duplicates them with the full worked physics and citations included in the README, for anyone who wants to see (and check) exactly where every coefficient came from rather than trust a black-box .blend file — 0 sales on that so far.

Reference numbers (Abbe values, dispersion, per-channel IOR triplets) for all four base materials are also written up as a plain page, independent of Blender: https://j4zz.eu/blender-dispersion/

Honest limits: this is a tiny niche (physically-accurate glass dispersion for 3D renders) inside an already-small niche (Blender material creators who care where their numbers come from). The free tier already reaches more people than the paid pack ever will — I'm posting here mainly because BlenderKit's own download count proves some real demand exists for the free version, and I have genuinely no data yet on whether anyone wants the paid, cited version.