Multilayer optical proximity correction (OPC) model for OPC correction

An apparatus comprising a machine-readable medium having instructions which when executed by a machine cause the machine to perform operations comprising:

• creating a semi-physical model of a mask for a current layer in an IC design layout using physical parameters of a lithography process used to create the mask, the semi-physical model specifying contours of the plurality of features of the mask;
• determining from design information whether the current layer is deformed by the one or more reference layers that overlap the current layer near the contours;
• responsive to determining that the current layer is deformed by the one or more reference layers, inputting the semi-physical model and the design information of the one or more reference layers into a trained machine learning algorithm to generate a contour shift prediction for the current layer, the contour shift prediction estimating a residual error of the semi-physical model;
• and
• using the contour shift prediction for multilayer OPC correction of the current layer.

A system comprising:

• a mask created via a lithography process;
• a physical silicon wafer fabricated using the masks, the physical silicon wafer having a plurality of features embodied therein as defined by the mask;
• a memory of a computer system, the memory to store:
• a semi-physical model of the mask created using physical parameters of the lithography process used to create the mask, the semi-physical model specifying contours of the plurality of features of the mask;
• and
• images of the plurality of features embodied within the physical silicon wafer;
• a multilayer model predictor executing on one or more processors configured to:
• quantify differences between (a) the contours of the plurality of features of the mask as specified by the semi-physical model and (b) the plurality of features embodied within the physical silicon wafer as captured by the SEM images to define target variables corresponding to the quantified differences;
• calculate density and optical parameters are for one or more overlapping reference layers in a proximity to the contours;
• use the target variables and the calculated density parameters of the reference layers to train a machine learning algorithm to characterize the difference between a measured contour and a prediction of the residual errors in a multilayer OPC model;
• responsive to determining that a current layer of a new mask is deformed by the one or more overlapping reference layers of the new mask, input a semi-physical model of a plurality of contours of the current layer and design information of the reference layers into the trained machine learning algorithm to generate a contour shift prediction for the current layer, the contour shift prediction estimating a residual error of the semi-physical model;
• and
• apply the contour shift prediction to the semi-physical model to generate a final multilayer model for the current layer that is used to create a corrected physical mask for the current layer.