Mentor Graphics Corporation

Siemens Industry Software Gmbh

Shumay Dou Shang

John L Sturtevant

Azat M Latypov

Yuri Granik

Konstantinos G Adam

Germain Louis Fenger

Le Hong

Gurdaman Khaira

Fan Jiang

Lianghong Yin

Hyejin Jin

Lisa Swallow

Various aspects of the present disclosed technology relate to techniques for classifying layout patterns. First, a set of density feature vectors for a set of layout regions in the layout design are extracted using a set of rings. Each component of a density feature vector in the set of density feature vectors corresponds to a ring in the set of rings. The set of rings do not overlap with each other and cover a whole area of a circle when being placed together. Next, a machine learning-based clustering process is performed to separate layout features in the set of layout regions into clusters of layout features based on the set of density feature vectors. Each of the clusters of layout features may be further divided into subclusters based on one or more properties.

Machine learning-based clustering for curvilinear layout designs

Systems and methods for calculating a printed area metric indicative of stochastic variations of the lithographic process are disclosed. Lithography is a process that uses light to transfer a geometric pattern from a photomask, based on a layout design, to a resist on a substrate. The lithographic process is subject to random stochastic phenomena, with the resulting stochastic randomness potentially becoming a major challenge. To characterize the stochastic phenomena, a printed area metric may be generated analytically (rather than via simulations) and comprise one or more defined moments for a printed area distribution associated with the printed area that are indicative of one or more aspects associated with printing. For example, the printed area metric may be indicative of the likelihood of printing within the printed area or the variance of printing within the printed area due to stochastic randomness in one or both of exposure or resist process.

Method and system for calculating printed area metric indicative of stochastic variations of the lithographic process

Aspects of the disclosed technology relate to techniques for using hotspot simulation to make wafer rework decisions. Metrology data of photoresist patterns created based on a layout design for a circuit design by a photolithographic processing step are received during a lithographic process. Hotspots of interest are selected based on comparing the metrology data with simulated metrology data associated with hotspots. The simulated metrology data and information of the hotspots are generated by performing lithographic simulation on the layout design before the lithographic process and stored in a library of potential hotspots. Lithography simulation is performed on the selected hotspots of interest using process conditions of the photolithographic processing step to generate simulated hotspot data. The simulated hotspot data are analyzed to determine whether rework of the one or more wafers or a wafer lot to which the one or more wafers belong is needed.

Simulation-assisted wafer rework determination

A system for producing mask layout data retrieves target layout data defining a pattern of features, or portion thereof and an optimized mask layout pattern that includes a number of printing and non-printing features. Mask layout data for one or more subresolution assist features (SRAFs) is then defined to approximate one or more non-printing features of the optimized mask layout pattern.

Model-based SRAF insertion

Growing community of inventors

Saratoga, CA, United States of America

Shumay Dou Shang