Integrated circuit interconnection reliability analysis method for modeling based on neural network parameters

Abstract

The invention discloses an integrated circuit interconnection reliability analysis method for modeling based on neural network parameters. The method comprises the steps of acquiring training data by ANSYS simulation; building a neural network model structure based on EM reliability; and training a reliability neural network model and acquiring a model result. The advantages that the neural network has a strong non-linear mapping function and a rapid learning ability are utilized to model an EM reliability input and output relation via the neural network and train the model, results with respect to reliability under various work conditions are acquired after the model is built, reliability data under different conditions can be predicted in very short time, the defect that simulation time is long due to the fact that the reliability results under all input conditions need to be simulated when traditional ANSYS finite element software analyzes circuit reliability in different conditions, and the method provided by the invention is significant in research on interconnection reliability of the integrated circuit.

Description

technical field [0001] The invention relates to the field of reliability modeling of integrated circuit interconnection, in particular to a neural network parameter modeling method applied in reliability analysis of integrated circuit interconnection. Background technique [0002] With the development of integrated circuits, the integration of circuits is getting higher and higher, the structure is more and more complex, and the size of devices is getting smaller and smaller, which brings about the problem of circuit reliability. The reliability of an integrated circuit mainly depends on the reliability of its transistors and interconnection lines. Among them, due to the reduction of metal line width and thickness, as well as the complex and changeable interconnection structure, the current density in the interconnection line gradually increases, and the high current density promotes the migration of metal atoms in the interconnection line, forming an interconnection structu...

AI Technical Summary

Problems solved by technology

This method can quickly and accurately predict reliability data under different conditions, and overcomes the shortcomings of long simulation time caused by the need to simulate reliability results under all input conditions when the traditional ANSYS finite element software analyzes circuit reliability under different conditions

Images