The Research Of GaN LED Degradation Model

The reliability of GaN LED is an important factor in GaN LED application.Lifetime and the failure rate of GaN LED should be guaranteed before the effectiveand reasonable application. With the development of process and technology, thereliability of GaN LED is better than before. GaN LED devices can work effectivelymore than decades, which lengthen the evaluation time of GaN LED device reliability.In the accelerated test, it takes at least thousands of hours to evaluate the devicereliability by the electronic device reliability evaluation criteria of U.S. militarystandard, GJB and GB, etc. What’s more, in the process of evaluation, the details ofthe degradation are neglected, simultaneously. The information before failure isutilized inadequate and the failure mechanism can’t be distinguished in the wholedegradation process. Therefore, the evaluation of device reliability is imperfect in theaccelerated test.The reliability of GaN LED can’t be evaluated by the accelerated test rapidly andaccurately. Based on the degradation process of GaN LED in the accelerated test, weanalyze the details of degradation process. GaN LED degradation model is establishedaccording to the degradation process. Firstly, Effectiveness of the degradation modelis guaranteed via the consistency of failure mechanism. Then, the degradation modelis verified dependability with the degradation process corresponds to the failuremechanism. Finally, the experimental error is discussed which caused by thermalshock in the off-line measurement accelerated test. We avoid the test errors via on-linemeasurement which ensure the accuracy of the degradation model. The researchcontents are as follows:1. GaN LED degradation is established. The degradation process of GaN LEDcan be described by the degradation model. With the help of the chemical kinetics, thedegradation model is established based on the temperature effect on the reaction rateand the change of reaction volume concentration in the physical/chemical reactions.Moreover, the accelerated test is carried out for the GaN LED, figuring out theparameters for the degradation model which including the ratio of differentdegradation processes and the time constant.2. The consistency of failure mechanism is researched. The degradation model isestablished based on the degradation process which is corresponds to the failuremechanism. The failure mechanism is stimulated by accelerated stress in thedegradation may be different from that under normal condition, which would lead toinvalid accelerated test. The altered failure mechanism will cause the invaliddegradation model. To solve the problem, we study the relation between the Arrheniusequation and the distribution function in the degradation process. Two relationships of distribution parameters can identify the consistency of failure mechanism.3. The degradation process corresponds to the failure mechanism. The details inthe degradation process are analyzed with the information before the GaN LED failed.We study the degradation process corresponding to the failure mechanism, andextrapolate the device lifetime, which guarantee the validity of the degradation model.The proposed method may provide a solution to evaluate the reliability ofsemiconductor products, which improving the productivity and assuring the accurateof the results.4. The measurement error in the accelerated test is analyzed. In the acceleratedtest, the offline measurement leads to the thermal and current shock, which affect theaccuracy of the degradation model parameters. In order to eliminate these errors andthe operating error, this paper uses the online measurement. The automated test andmeasurement systems which consist of computer, GPIB and HP-VEE programachieve accurate results under the electrical and constant temperature stress. In theautomated test and measurement systems, computer is used as a control terminal, thedata communications infrastructure realized via the international Common InterfaceIEEE-488bus line, and the experiment is controlled by HP-VEE program. The onlinemeasurement collects real-time experimental degradation data, eliminate theexperimental error, and improve the accuracy of measurement.