Time Evolution Degradation Physics in High Power White LEDs Under High Temperature-Humidity Conditions

Abstract

A high temperature-humidity test is commonly employed to evaluate the humidity reliability of electronic devices. For an integrated circuit, the degradation mechanism under the high temperature-humidity test is metal corrosion, and Peck's model is used for extrapolating the test results at accelerated test conditions to the normal operating condition. Such extrapolation is possible as the underlying degradation physics is invariant from the accelerated test conditions to the normal operating condition for integrated circuits. However, this is not true for high power LEDs, as found in this paper. The degradation in the LEDs undergoes time evolution at either 95% or 85% relative humidity (RH) and 85 $^{\circ}\hbox{C}$ . We also found that the degradation physics are completely different among the various RH levels from 95% to 70%. The degradation process begins from bond pad contamination and Kirkendall void formation, galvanic dissolution, phosphor dissolution to encapsulant, and die attach delamination. Such time evolution degradation physics renders the inapplicability of the Peck model and presents a challenge in extrapolation of test results to the normal operating condition for lifetime prediction.