The goal of CALCE LED research group is to show potentialities and preceding results of PoF-based and data driven-based prognostics and health management (PHM) into useful life dynamics on degradation failures of LEDs and LED lightings to correlate degradation of LED light output with other multi-factor failure modes such as degradations of light output, color parameters, and electrical degradations.

Development of prognostics and health management (PHM) for LEDs is critical for fast and economic LED technology insertion in lighting applications. LEDs have been considered as new light sources for the enhanced lighting services because LEDs offer design flexibility; easy adjustment of line-scale, area-scale, and color dimming; high energy efficiency resulting low power consumption; wide range of color; and no low-temperature startup problems. Although technological advances in recent years have improved LED performance, reliability still remains an issue for LEDs. The structure of LEDs is somewhat similar to microelectronics; however there are unique design/reliability requirements and materials in LEDs.

As the LED products in the lighting industries gradually replace the conventional light sources such as the fluorescent lamp and the white bulb, information about LED reliability under a variety of operating conditions is required by the users. This data allows product designers to deliver the best combination of purchase price, performance, and cost of ownership for the life of the product. A number of techniques to predict remaining useful life of LEDs have been proposed, but they turn out to be of limited use since the knowledge of failure modes and mechanisms is incomplete and the factors of the degradation are numerous. End customers would like to have the LED industry guarantee lifetime of LEDs at usage conditions.

The prediction of LED lifetime can be varied with how interpreting the results of the accelerated life tests. The conventional accelerated factor mainly reflects the temperature difference between the junction temperature and the outer temperature. This method for predicting remaining useful life has ambiguous assumptions. An advanced method of LED lifetime estimation needs to reflect total consideration of factors including LED package geometries, operating conditions, materials.

The reliability is dependent on the temperature of the p-n junction of the LED die. As the junction temperature is increased, the lifetime is degraded. Possible other stress factors are needed to estimate the lifetime of LEDs more specifically at the user environment in terms of failure mechanisms of LED packages. The definition of the LED lifetime is based on the time to reach certain level of light output degradation such as 50% and 70%. However, failures of LEDs can manifest in many modes including yellowing of (epoxy) encapsulant, degradation of lens/ encapsulant material, degradation of die-attach epoxy, discoloration of metal reflector, phosphor thermal quenching, delamination.