| Popis: |
Inverse Lithography Technology (ILT) has become one of the key computational lithography solutions, which may produce mask output that results in better process latitude and CD control on wafer than the one using conventional OPC. However, the curvilinear inverse lithography recipe parameter tuning is a non-trivial task, which involves optimization for various common objectives such as common DOF, NILS, and pvband (process variability band) [1]. It is known that test patterns used for recipe development play a critical role in achieving optimized ILT masks in terms of mask-friendliness, OPC convergence or multi-structure common focus range. The traditional way of test pattern selection is usually a clip-level manual search by taking into account of design rules, which inevitably may cause lack of critical design representations. In this paper, we introduce Mentor Graphic’s Calibre SONR[2], a Machine Learning (ML) method to implement design layouts clustering and automatic pattern selections for ILT recipe tuning on a full-chip level. Calibre SONR leverages design layouts and process models to create necessary features as input and then combine with a proprietary clustering algorithm to fully analyze and cluster a full layout in feature vector space. The degree of pattern reduction is controlled by an adjustable similarity metric[2]. The paper has four parts: the first part is on the feature generation and data collection based on the design layout, the lithography and etch process models. A wide range of over 100 features are extracted to represent the layout and process etc. The second is on the SONR model building and pattern selection from the full-chip; the third on the ILT recipe creation process with patterns selected from SONR and the traditional method; the fourth on the pattern coverage analysis and verification by evaluating the recipe quality through OPCV checks, and compare the results between SONR-based and traditional pattern selection methods. |
