A Simplified Testing Approach for Performance Analysis of Integrated Circuits through Reliability Testing

During the COVID (coronavirus disease) period, the semiconductor market was hit by the shortage of electronic chips affecting many industries. To cope with the issue in Moxa, two major steps were suggested, first is to use C-grade ICs (Integrated Circuit) instead of A-grade ICs and second, is to use the ICs of alternate vendor instead of the regular vendor. ICs are commonly classified into different grades based on their operating temperature specifications. For example, A-grade ICs are rated for temperatures between -40°C and 85°C, while C-grade ICs are rated for 0°C to 70°C. Although both grades of ICs share the same design and are manufactured using identical processes, the vendor employs a sorting method to categorize these ICs according to their respective operating temperature ranges. To verify that such steps will not impact the reliability of the system, accelerated reliability tests namely temperature-humidity and temperature cycle are conducted for the A-grade and C-grade ICs. Same tests are conducted for the alternate vendor and regular vendor ICs to compare their performance. However, testing complex ICs having multiple pins is not a straightforward way as it requires complex test circuits. This work proposed a simple method of testing such ICs by using just the I-V characterization pins of the ICs. I-V characterization is the process of measuring the current-voltage relationship of electronic devices. It involves sweeping the voltage from a specified minimum to a maximum value and recording the corresponding current. The proposed method is based on the fact that the failure in bond wires and bond pads are the common causes of ICs failures in the operational environment. This article discusses temperature-humidity and temperature cycling testing, detailing their results, and computing reliability under normal environmental conditions. It starts by using I-V data to identify the resistance change associated with the pins of IC and define failure criteria. Through a series of failure analysis, we have identified corrosion as the root cause of IC failure. Consequently, leveraging this understanding of failure physics, we ascertain that its reliability model conforms to the PoF (Physics of Failure) models. Furthermore, analysis of test data reveals a statistical model consistent with log-normal distribution. By integrating information on product usage environment, we derive an acceleration model and subsequently estimate the 5-year reliability of the IC under normal operating conditions. The 5-year reliability results show that both A-grade and C-grade ICs as well as the alternate and regular vendors ICs are resilient to temperature-humidity but vulnerable to temperature variations. Interestingly, the performance of A-grade and C-grade remains consistent over 5 years. However, the 5-year performance of DUTs (device under test) from alternate vendors surpasses that of those from the regular vendors.