The macroscopic optical appearance of materials -- transparent, colored, black, or lustrous -- cannot be fully classified by the band gap alone. Both metals and graphite possess a vanishing band gap, yet the former exhibits metallic luster while the latter appears black with cleavage-plane gloss. We propose that existing physical quantities -- the inverse participation ratio (IPR) and the Wannier function spread -- can serve as a unified descriptor (delta) for optical classification when combined with an effective conduction dimensionality Dₑff. Using five carbon allotropes (diamond, C60, SWCNT, graphene, graphite) and hexagonal boron nitride (h-BN) as a control, we demonstrate that a decision tree based on Eg and Dₑff correctly classifies 6 of 7 materials (85. 7%), with the sole failure -- single-layer graphene -- requiring the layer count N as a third variable. Tight-binding simulations with Kubo-formula optical conductivity and Fresnel reflectivity confirm the classification chain. Anderson disorder simulations show that systems with high delta x Dₑff maintain their optical response under disorder while low-delta x Dₑff systems rapidly decohere. DFT + Wannier90 calculations for diamond and graphite independently confirm the delocalization ordering. The recently proven optical sum rule connecting Wannier spread to Imepsilon (omega) /omega (Cardenas-Castillo et al. 2024) provides a rigorous mathematical foundation for the delta-optics connection.
Kazuyoshi Miyauchi (Sun,) studied this question.