Research on Acceleration Test Model and Parameter Determination Method of Integrated Circuit

With the development of microelectronics manufacturing technology, the reliability level of electronic components has also increased rapidly. The failure rate of common commercial semiconductor chips has been as low as 10-9. It is difficult to obtain quantitative reliability data of devices by traditional reliability tests. Therefore, the accelerated life test with the advantages of higher efficiency and lower test cost has been rapidly developed in the field of reliability evaluation of electronic components. The difficulty of accelerated life test is to choose the appropriate acceleration model and determine the parameters of the model. This paper studies the characteristics and application scenarios of three acceleration models commonly used in accelerated life tests of integrated circuits— the Arrhenius model, the Coffin Manson model, and the Hallberg-Peck model. By studying the test conditions of life test, temperature cycle test and humidity test in common test standards of integrated circuits, the potential laws behind the standard conditions are excavated. Introduce the test conditions specified in the commonly used standards into the model parameter determination method: 1) For the life test, the steady-state life and aging test conditions in GJB548C-2021 are introduced into the Arrhenius model parameter determination method; 2) For the temperature cycle test, combined with JESD47G The specified test conditions are introduced into the Coffin Manson model parameter determination method; 3) The test conditions of GB/T 4937.4-2012 on the HAST test are incorporated into the Hallberg Peck model parameter determination method. The research results show that the life test conditions conform to the Arrhenius model, the temperature cycle test conditions conform to the Coffin-Manson model, the temperature and humidity comprehensive test conditions such as humidity resistance and HAST tests conform to the Hallberg-Peck model, and the corresponding model parameters are obtained through fitting calculations. It can be seen that there is a linear correlation greater than 0.95 between the two, which provides a basis and ideas for formulating an integrated circuit accelerated life evaluation plan, selecting key technical parameters such as accelerated model, accelerated stress conditions, and accelerated time.