Combined Methodology for Accurate Evaluation of Distance and Direction of Chromaticity Shifts in LED Reliability Tests

Light Emitting Diodes are an integral part of modern illumination systems. While their long-term stability in terms of lumen maintenance used to be in the focus leading to drastically increased nominal lifetimes over the past decade, the color shift or respectively chromaticity shift is lately attracting more and more attention especially in lighting applications that demand for a high chromatic stability. Common ways of displaying and reporting color shifts in LEDs and LED products provide only limited possibilities to accurately classify changes in color according to the current understanding of the underlying degradation processes. This paper therefore presents a methodology to overcome the observed discrepancies in analyzing time-dependent chromaticity data in terms of their shift direction and distance within the CIE 1976 (
$u',v'$
) chromaticity diagram. The use of an extrapolation function extending the shift towards the spectral locus enables the determination of the corresponding nearest spectral locus intersection wavelength. Thereby, it is possible to classify changes in color according to different shift regions, which basically allows for performing an automated comprehensive color shift mode analysis. As part of this paper five different algorithmic approaches for determining the intersection wavelength and shift region are discussed and compared in terms of their accuracy and ease in implementation.