Silicon Carbide power device reliability Chip development and packaging

The narrow concept of quality is to measure whether the device meets the specified standard requirements at present, that is, whether the product performance is consistent with the content described in the specification. In a broad sense, quality is related to reliability, which is a measure of the expected life of a device, that is, how long a device can last to meet the specifications based on reliability results. It is easy to test the compliance of a new device, but difficult to determine whether the physical characteristics of the device change over time and environment.

The previous article introduced the relationship between materials and the reliability of silicon carbide power devices, and then discussed the reliability problem from the aspects of silicon carbide chip development and packaging.


The reliability can be measured from the failure rate of the device. In the typical fault curve bath curve, with the effect of environment, time and electric field, the failure rate of the device can be divided into three periods:

The initial failure zone lasts approximately 3 to 15 months, usually 1 year. The failures occurring in this region are common to most semiconductor components, mainly caused by design and manufacturing reasons, and can be screened; The available period area is generally 10 years, and random failure will occur. Failure in the aging area is generally due to material fatigue and aging.

 

Reliability tests must be carried out before products are put on the market. According to the known failure mechanism of reliability test, the accelerated mode test method is designed to reproduce the failure phenomenon and eliminate hidden dangers, so as to avoid avoidable failure in the process of use.

Each reliability test corresponds to a certain failure mode, and there are different classification of test methods according to different environmental conditions, test items, test purposes and test properties. According to the conventional classification, the test methods can be summarized into five categories: environmental test, life test, screening test, field use test and identification test.