Method for setting up semiconductor simulated model

Abstract

The invention discloses a semiconductor simulation module establishing method, comprising the steps of: firstly by dependence of length and width, establishing device model of Global region, namely G model; then devices on the boundary between Bin region and Global region use G model as their own single device models; then singly extracting models of each device in Bin region not bordering Global region; then each Bin block synthesizes its own block model by device models at its four corners; and finally integrating G model and several Bin models into a whole model. And the invention has advantages of both global model and blocked model and besides, overcomes their disadvantages, obtaining higher model accuracy and improving model maintaining performance, and thus raising IC design efficiency and accuracy, shortening product design cycle and reducing cost.

Description

technical field [0001] The invention relates to a method for establishing a simulation model, in particular to a method for establishing a simulation model of a semiconductor device. Background technique [0002] The device model plays a very important role in the design of integrated circuits. It can greatly shorten the design and production cycle of products, improve the yield of products, save costs and so on. At present, there are generally two methods for establishing device models, one is the global model method (Global Model), and the other is the block model method (Binning Model). [0003] The global model simulates all sizes of transistor devices with one model. According to the physical theory of semiconductor devices, the performance of metal oxide field effect transistors (MOSFET), such as threshold voltage Vt, channel current Ids, etc., mainly depends on its short-channel and narrow-channel effects, and there is a very large relationship with its gate length a...

AI Technical Summary

Problems solved by technology

The block model has the advantages of good fitting accuracy because it makes a specific model in a specific size, but the simulation speed of this model is slow, and most of the parameters have no specific physical meaning. Because the model is composed of many sub-models, the model maintenance becomes easier. It is more difficult to obtain, and it is easier to have continuity problems between blocks, causing non-convergence

Images