Abstract:
Described herein is a method to determine a mask pattern for a patterning device. The method includes obtaining a target pattern to be printed on a substrate, (ii) an initial continuous tone image of the patterning device corresponding to the target pattern, (iii) a binarization function (e.g., a sigmoid, an arctan, a step function, etc.) configured to transform the initial continuous tone image, and (iv) a process model configured to predict a pattern on the substrate from an output of the binarization function; and generating, by a hardware computer system, a binarized image having a mask pattern corresponding to the initial continuous tone image by iteratively updating the initial continuous tone image based on a cost function such that the cost function is reduced. The cost function (e.g., EPE) determines a difference between a predicted pattern determined by the process model and the target pattern.
Abstract:
Scanner aberration impact modeling in a semiconductor manufacturing process is described. Scanner aberration impact modeling may facilitate co-optimization of multiple scanners. Scanner aberration impact modeling may include executing a calibrated model and controlling a scanner based on output from the model. The model is configured to receive patterning system aberration data. The model is calibrated with patterning system aberration calibration data and corresponding patterning process impact calibration data. New patterning process impact data may be determined, based on the model, for the received patterning system aberration data. The model comprises a hyperdimensional function configured to correlate the received patterning system aberration data with the new patterning process impact data. The hyperdimensional function is configured to correlate the received patterning system aberration data with the new patterning process impact data in an approximation form in lieu of a full simulation without involving calculation of an aerial image or a representation thereof.
Abstract:
Described herein are different methods of training machine learning models related to a patterning process. Described herein is a method for training a machine learning model configured to predict a mask pattern. The method including obtaining (i) a process model of a patterning process configured to predict a pattern on a substrate, wherein the process model comprises one or more trained machine learning models, and (ii) a target pattern, and training, by a hardware computer system, the machine learning model configured to predict a mask pattern based on the process model and a cost function that determines a difference between the predicted pattern and the target pattern.
Abstract:
Disclosed herein are methods of constructing a process model for simulating a characteristic of a product of lithography from patterns produced under different processing conditions. The methods use a deviation between the variation of the simulated characteristic and the variation of the measured characteristic to adjust the parameters of the process model.
Abstract:
Described herein are methods and systems for determining mask rule check violations (MRC) associated with mask features using a detector having geometric properties corresponding to the MRC. The detector (e.g., elliptical shaped) is configured to include a curved portion to detect a curvature violation, an enclosed area (e.g., a fully enclosed area or a partially enclosed area having an opening), a predefined orientation axis configured to guide relative positioning of the detector with a mask feature, and a length to detect a critical dimension violation. The orientation axis of the detector is aligned with a normal axis at a location on the mask feature. Based on the orientation axis aligned with the normal axis of the mask feature, an MRC violation is determined corresponding to a region of the mask feature that intersects the enclosed area.
Abstract:
Described herein is a method for determining a wavefront of a patterning apparatus of a patterning process. The method includes obtaining a reference performance (e.g., a contour, EPE, CD) of a reference apparatus (e.g., a scanner), a lens model of a patterning apparatus configured to convert a wavefront parameter of a wavefront to actuator movements, and a lens fingerprint of a tuning scanner (e.g., a to-be-matched scanner). Further, the method involves determining the wavefront parameter (e.g., wavefront parameters such as tilt, offset, etc.) based on the lens fingerprint of the tuning scanner, the lens model, and a cost function, wherein the cost function is a difference between the reference performance and a tuning scanner performance.
Abstract:
Described herein is a method to determine a curvilinear pattern of a patterning device that includes obtaining (i) an initial image of the patterning device corresponding to a target pattern to be printed on a substrate subjected to a patterning process, and (ii) a process model configured to predict a pattern on the substrate from the initial image, generating, by a hardware computer system, an enhanced image from the initial image, generating, by the hardware computer system, a level set image using the enhanced image, and iteratively determining, by the hardware computer system, a curvilinear pattern for the patterning device based on the level set image, the process model, and a cost function, where the cost function (e.g., EPE) determines a difference between a predicted pattern and the target pattern, where the difference is iteratively reduced.
Abstract:
Disclosed herein is a computer-implemented method including: determining a first partial image formed from a first radiation portion propagating along a first group of one or more directions, by a lithographic projection apparatus; determining a second partial image formed from a second radiation portion propagating along a second group of one or more directions, by the lithographic projection apparatus; determining an image by incoherently adding the first partial image and the second partial image; wherein the first group of one or more directions and the second group of one or more directions are different.
Abstract:
Disclosed herein is a computer-implemented method to improve a lithographic process for imaging a portion of a design layout onto a substrate using a lithographic projection apparatus comprising an illumination source and projection optics, the method comprising: obtaining a source shape and a mask defocus value; optimizing a dose of the lithographic process; optimizing the portion of the design layout for each of a plurality of slit positions of the illumination source.
Abstract:
Disclosed herein is a computer-implemented method for improving a lithographic process for imaging a portion of a design layout onto a substrate using a lithographic projection apparatus, the method comprising: providing a desired pupil profile; calculating a discrete pupil profile based on the desired pupil profile; selecting a discrete change to the discrete pupil profile; and applying the selected discrete change to the discrete pupil profile. The methods according to various embodiments disclosed herein may reduce the computational cost of discrete optimization from O(a n ) to O(n) wherein a is constant and n is the number of knobs that can generate discrete change in the pupil profile.