A novel multiple-stress-based predictive model of LEDs for rapid lifetime estimation. (November 2017)

• Record Type: Journal Article
• Authors: Huang, Su-Dan; Zhou, Lin; Cao, Guang-Zhong; Liu, Huai-Yuan; Hu, Yi-Min; Jing, Gang; Cao, Ming-Gao; Xiao, Wen-Peng; Liu, Yan
• Is Part Of: Microelectronics and reliability. Volume 78(2017)
• Page Start: 46
• Page End: 52
• Publication Date: 2017-11
• Subjects: Adaptive genetic algorithm (AGA) -- Multiple stresses -- Light-emitting diodes (LEDs) -- Lifetime estimation -- Predictive model; Electronic apparatus and appliances -- Reliability -- Periodicals; Miniature electronic equipment -- Periodicals; Appareils électroniques -- Fiabilité -- Périodiques; Équipement électronique miniaturisé -- Périodiques; Electronic apparatus and appliances -- Reliability; Miniature electronic equipment; Periodicals; 621.3815
• DOI: 10.1016/j.microrel.2017.07.094
• Languages: English
• ISSNs: 0026-2714

Abstract

Abstract: This paper proposes a novel multiple-stress-based predictive model (MSBPM) to rapidly assess the lifetime of light-emitting diodes (LEDs). The MSBPM addresses the lifetime estimation of LEDs with respect to temperature, humidity, and current; these three stresses are rarely considered simultaneously in the assessment of reliability. Using several degradation data sets from accelerated life tests (ALTs) without using extrapolation method, a designed adaptive genetic algorithm is employed to identify five unknown parameters of the MSBPM. A simulation of the proposed MSBPM is presented as validation. By applying the degradation data from the ALTs under high stresses, an MSBPM for the LEDs is established. Under the nominal conditions of 25 °C/22.5% RH and a current of 0.35 A, the lifetime of the LED is estimated using the established MSBPM. The effectiveness of the proposed MSBPM is further verified through the estimated results. Highlights: A multiple-stress-based predictive model is proposed to rapidly estimate LED lifetime. The relationship between lifetime and primary stresses is characterized for LEDs using the proposed model. The proposed model is capable of estimating LED lifetime accurately using as much of the sample data sets as possible.