Measured Host Optical Constants Shift Deep-Blue OLED Outcoupling by up to 14% (Relative) While Leaving the Stack Design Intact

Outcoupling designs for organic light-emitting diodes rest on the complex refractive index of every layer, yet host films are often given a constant index without absorption. We show what measured constants change, and what they do not. Variable-angle spectroscopic ellipsometry of three deep-blue phosphine-oxide host films (S1, S3 and S4), refitted from the raw data with several dispersion models, pins the index of the S1 and S4 films to a model-andspot spread of about ±0.002 (about ±0.004 at 95% once substrate-oxide, incidence-angle and roughness systematics are propagated) and reveals a weak absorption tail in S4 that follows its triphenylamine-centered lowest singlet transition; the S3 film is bracketed. Propagated through a dipole model of a deep-blue stack in which the neat host film is the emitter, the measured constants leave the design itself (best emitter position, optimum electron-transport-layer thickness) unchanged within a few nanometers but shift the predicted outcoupling by up to 14% relative (15.1 against 17.2% at the optimum, about 2 percentage points), with the blue dispersion of the index and, for S4, the absorption tail in the emission band contributing comparably (up to about 7% each). Emitter orientation and the emitter–cathode distance act far more strongly, by factors of 2 to 5. Both films are close to isotropic, yet the calculated transition dipoles of S3 and S4 lie in the molecular plane, so growing these hosts with flat-lying molecules is the clearest route to higher outcoupling, worth about a factor of two in the model.