A Neural Network-Based Correction Method for Electron Beam Proximity Effect

Abstract

The invention belongs to the field of computational lithography, and discloses a method for correcting the proximity effect of an electron beam based on a neural network. The present invention first corrects the dose of the designed training layout through any kind of proximity effect dose correction method, then adjusts the neural network parameters, and performs neural network training on the training layout according to the prescribed input method to obtain an adaptive neural network. Network, and finally predict the correction dose of any exposure layout according to the specified input method. The invention realizes the correction of the proximity effect of the electron beam based on the neural network, and solves the correction calculation of the proximity effect of the electron beam exposure. On the basis of ensuring high-precision correction, the calculation efficiency of the correction is greatly improved, and it has high calculation accuracy and high calculation efficiency. Efficiency and other characteristics.

Description

technical field [0001] The invention relates to a correction method for proximity effect in accelerated electron beam lithography. Through a neural network training method, on the basis of ensuring high-precision correction, the calculation efficiency of correction is greatly improved, and belongs to the field of computational lithography. Background technique [0002] Electron beam lithography (EBL) is an important technology for micro-nano processing to prepare high-resolution nanoscale lithography layouts, and has broad application prospects. Proximity effects in e-beam lithography degrade pattern quality in small-scale exposures below 10 nanometers, and direct exposure without proximity correction (PEC) has a large impact on resolution, limiting the range of e-beam exposures. resolution. [0003] As the smallest features in integrated circuits are continuously reduced to the nanometer scale to increase circuit density, proximity effects are no longer negligible. The co...

AI Technical Summary

Problems solved by technology

This method can obtain high-precision correction, but its calculation process is complicated, and it is difficult to apply to large-scale layout calculations.

[0005] Although there is a method to accurately calculate the proximity effect correction, the application of large-scale integrated circuits cannot be quickly and effectively solved due to the large number of iterations of the physical method or the calculation steps include complex calculation units, which limits the masking capacity of integrated circuit manufacturing. Membrane precision

Images