Molybdenum-silicon thin films are used in extreme environments where thermal radiation matters — coatings, emitters, high-temperature sensors. The natural assumption is that composition controls emissivity: more molybdenum, more metallic, higher infrared emission. Density functional perturbation theory calculations across multiple MoₓSiᵧ stoichiometries show this assumption is wrong.
Note (2026-11-10): I checked the abstract of arXiv 2603.17801. The essay says the real finding is about defects. In the paper the defect result comes from test calculations on defected MoSi2, which it says demonstrate that the infrared emissivity of MoSi2 thin films can be substantially increased by introducing defects; the main results are the calculated emissivities of the stoichiometric phases, the lack of a simple correlation with molybdenum content, and a predicted maximal emissivity for hot metallic films of about 900 K at 5 to 10 nm thickness, in good qualitative agreement with experiments on the lower emissivity of hexagonal compared to tetragonal MoSi2. The essay's statements that a perfect crystal minimizes emissivity and that defects create states that couple to infrared radiation are the essay's inference.
Emissivity doesn't correlate with molybdenum content. Hexagonal MoSi₂ has much lower emissivity than tetragonal MoSi₂ — same composition, different crystal structure, different optical response. The difference traces to the electronic structure: a small band gap in the hexagonal phase versus a low density of states at the Fermi level in the tetragonal one. The composition is identical. The property is not.
But the real finding is about defects. Introducing structural defects into MoSi₂ substantially increases infrared emissivity. The imperfection isn't degradation — it's enhancement. Defects create additional electronic states that couple to infrared radiation, turning a mediocre emitter into a strong one.
This inverts the engineering intuition. A perfect crystal minimizes emissivity. Deliberate damage maximizes it. The optimization target for thermal emission isn't purity or stoichiometric precision — it's controlled imperfection. The defect isn't the problem to be eliminated. It's the feature to be designed.