Research on Multifault Testing Method for MIV Based on Grid Search and Random Forest

Abstract

With the fast development trend of highly integrated electronic products, as the key technology of 3-D interconnect circuits, the research on monolithic intertier via (MIV) testing technology is particularly critical. Due to the high integration density, the MIV itself is very prone to manufacturing faults, moreover, the uncertainty of fault location, the complexity of fault type, that will have a greater impact on the signal transmission performance of the MIV. Therefore, in order to effectively isolate the faults and reduce the loss of assembly yield, a new nondestructive MIV fault testing method is proposed. The method adopts random forest (RF) classification model optimized based on grid search (GS) optimization algorithm. This test method can effectively solve the problem that traditional MIV test methods are difficult to accurately test for MIV faults. As well as the study of multifault in response to the fact that the existing methods are less studied for the presence of multifault in MIV. The simulation results show that the GS-RF-based MIV fault testing method can avoid damage to the MIV under testing during the MIV fault testing process, and the accuracy of single-fault testing reaches 96.11%, and the accuracy of multifault testing reaches 91.38%.