Level set mask optimization method for distance regular level set and convolutional neural network

Abstract

A level set mask optimization method for a distance regular level set and a convolutional neural network aims to solve the optimization problem of a mask plate under an advanced process, selects a depth level set mask data set and a corresponding mask label as a sample data set for training a neural network model, selects a group of pictures and corresponding labels in the sample data set, performs random rotation, zooming, translation and cutting processing, and readjusts an image sequence to synthesize two three-dimensional data packets: storing depth level set data in one packet and storing corresponding labels in the other packet, so as to realize a joint segmentation effect of the convolutional neural network and the level set.

Description

technical field [0001] The invention belongs to the field of chip manufacturing. Using the deep learning prediction method to optimize the chip manufacturing mask, the running time of the lithography process is greatly shortened. Background technique [0002] With the continuous reduction of the feature size of the integrated circuit process, the existing manufacturing technology is facing great challenges. Since the pattern size of the circuit is similar to the wavelength of the light source used in the lithography process, interference effects will inevitably occur, resulting in distortion of the lithography pattern, which in turn affects the production capacity of integrated circuits. The application of resolution enhancement technology is a key step to increase production capacity. Optical proximity correction is the core resolution enhancement technology. In the process of optical proximity correction, the pattern on the mask to be processed will be adjusted to compe...

AI Technical Summary

Problems solved by technology

However, the design rules of large-scale design under the advanced technology node increase exponentially, so the quality of the optimized reticle based on the design rule method is limited.

The other is the model-based method, which has a huge solution space, so it will be very time-consuming to get a high-quality result

With the increasing complexity of circuit design, the above design rule-based and model-based methods will face the problem of efficiency or output quality

Images