Demand for better pattern fidelity and the adoption of ILT are increasing pressure to shift to curvilinear mask technology.
Curvilinear masks are rapidly moving into high-volume production. This transition is driven by the need for better pattern fidelity, larger wafer process windows, and more effective use of inverse lithography technology (ILT) and curvilinear optical proximity correction (OPC). However, curvilinear masks also create a new challenge for mask data preparation (MDP): when curvilinear MULTIGON patterns are converted into piecewise linear representation (particularly in high fidelity applications) the data can grow significantly, increasing turnaround time and potentially degrading the original curve intent.
This is why the industry is adopting native curve data formats such as MULTIGON, the curvilinear extension to the SEMI P49 OASIS specification. MULTIGON enables mask data to represent real curves, including Bézier and B-spline curves, directly rather than approximating them through thousands or millions of short line segments. For manufacturers preparing for high-volume curvilinear mask production, native curve handling is becoming essential to preserve accuracy while managing file size, runtime, and downstream mask-writer requirements.
Curvilinear mask adoption is also enabled by the continued deployment of multi-beam mask writers, which can write complex curved shapes more efficiently than traditional variable-shaped beam (VSB) approaches. As a result, mask flows can now benefit from MDP solutions that maintain the benefits of curvilinear patterning across the full data path—from OPC or ILT output through fracture, mask rule checking, and final mask-writer preparation. Key benefits of curvilinear MULTIGON masks include:
The fracture engine is a critical component of the mask data preparation flow. It converts complex post-OPC or post-ILT layout data into mask-writer-ready data while preserving the intended geometry as accurately and efficiently as possible. In traditional rectilinear flows, this process is well understood. In curvilinear flows, however, the challenge is fundamentally different: the fracture engine must process smooth curves, comply with mask-writer constraints, and avoid unnecessary data growth or loss of geometric fidelity.
Simply digitizing curves into piecewise linear polygons is not an ideal solution. Piecewise linear polygonization introduces an approximation step, which can create edge placement error on silicon, increase vertex count, and make downstream verification and fracture operations more expensive. For curvature-sensitive operations, smaller digitization tolerances do not always guarantee better results because the computed geometry can become highly dependent on how the curve was sampled. Native curve processing avoids this class of problem by operating directly on the curve representation.

SmartFracture addresses this challenge by bringing native MULTIGON handling into the fracture flow supporting Boolean and fracture operations. SmartFracture is built on the same SmartEngine foundation used for SmartMRC, combining proven curvilinear and Manhattan data handling with the production fracture expertise of Synopsys CATS. This creates a unified platform for fracture, MRC, and mask correction workflows, helping users deliver curvilinear mask production without forcing a tradeoff between accuracy and operational efficiency.
Although SmartFracture was initially developed to address the emerging requirements of curvilinear mask flows and multi-beam mask writer formats, its value extends across the full range of mask manufacturing needs. SmartFracture supports both advanced MULTIGON-based curvilinear flows and established mask writer formats used in Manhattan-based production. This makes SmartFracture the solution of choice for users seeking best-in-class turnaround time and high-quality mask output, whether they are running mature rectilinear flows today or preparing for the most advanced curvilinear production flows.
The key advantage for advanced flows is native curve preservation. SmartFracture processes MULTIGON and Bézier-based data directly, avoiding the accuracy loss that can occur when curves are first converted into polygons. This helps preserve the design intent of curvilinear OPC and ILT while reducing the need for excessive polygonal decomposition. For users, this provides a production-oriented path to curvilinear mask manufacturing: accurate curve handling, controlled output data volume, and a platform aligned with modern multi-beam mask writer requirements.

SmartFracture also includes optimization for MULTIGON monotonicity requirements. In some cases, mask-writer or format constraints require a curved shape to be split into monotonic sections. Unnecessary cut-lines can increase point count and expand output file size. SmartFracture can choose an optimized cutting direction, helping reduce unnecessary cut-lines and produce output files that are smaller with faster turnaround time. As shown in the graph below, SmartFracture has significant advantages over competing tools in both turnaround time (TaT) and file size. The figure on the right shows the processing results for a circular hole pattern; SmartFracture clearly produces more periodic fracture segments, which greatly contributes to the reduction in file size.


As curvilinear masks become more important for advanced manufacturing, users need MDP technology that can preserve native curve accuracy while meeting production requirements for file size, runtime, and mask-writer compatibility. SmartFracture provides that path. By combining native MULTIGON fracture/curvilinear Boolean operations, optimized curve handling, and the SmartEngine platform foundation shared with SmartMRC, Synopsys enables users to deploy curvilinear mask production with confidence. Users can engage with Synopsys now to bring SmartFracture-enabled MDP flows into production quickly, improving readiness for today’s Manhattan-based mask requirements while accelerating the transition to advanced curvilinear MULTIGON-based manufacturing flows.
References
“A Study that Enhances the Accuracy of Multigon-based Curvilinear Mask Rule Check,”Jae Kwang Kim, et al., Photomark Japan 2025, Proc. of SPIE Vol. 13655.
“Comparing curvilinear layouts for the verification of mask data preparation,” Masakazu Hamaji, et al., Proc. of SPIE Vol. 13655.
“CATS SmartFracture – A New Fracture Engine for Curvilinear and Multigon Mask Data Input,” Kokoro Kato, et al., Proc. of SPIE Vol. 13216.
Leave a Reply