Quantifying the Accuracy and Precision of the Transition Dipole Moment Alignment from Realistic Angular Emission Data

Optoelectronic device efficiency depends on the effective orientation of its photoluminescent transition dipole moment(s). This orientation is typically quantified indirectly by fitting the angular emission pattern of a material. However, the accuracy and uncertainty of this procedure are unknown, and thus, the predicted efficiency of a device has the potential to be extremely inaccurate. Here, we quantify the inherent accuracy and precision of finding the orientation of transition dipole moment(s). We created artificial data sets of varying transition dipole moment alignments, refractive indices, and thicknesses and used statistical models to determine the fit accuracy and associated confidence intervals. The inherent confidence intervals are inconsistent across transition dipole moment alignments and samples: uncertainty increases for more horizontally aligned dipoles and for higher refractive indices, meaning that quantum-confined semiconductor films will inherently have a less precise fit. We then showed that accurately fitting the transition dipole moment alignment requires adequate knowledge of the film parameters or computationally expensive fitting methods. Finally, we incorporated realistic nonidealities into our generated data sets that led to extremely inaccurate predictions of the true transition dipole moment alignment, with some cases showing 10-30° difference from the true angle. To address this, we developed a new weighting mask that reduced these inaccuracies to be within a few degrees for most cases. Through this work, we provided a framework to more accurately quantify the transition dipole moment alignment and the uncertainty of the associated fit, enabling better predictions of material properties and future device performance.