Open-Source Differentiable Lithography Imaging Framework

TL;DR

The paper addresses the cost and complexity of optical lithography in semiconductor manufacturing by introducing an open-source differentiable lithography framework that leverages GPU-accelerated differentiable programming. It formulates the forward imaging process as a differentiable chain $f = \eta \circ f_i \circ f_h \circ f_m \circ f_r$ and evaluates Abbe and Hopkins models (including SOCS-based approximations), enabling gradient-based optimization of OPC and SMO. Key contributions include memory-efficient adjoint back-propagation for end-to-end optimization of optical components and masks, rigorous comparison against state-of-the-art lithography models showing superior accuracy on standard benchmarks, and a fully open-source release to accelerate collaborative development. The framework has potential to improve RET optimization and lithography fidelity, supporting continued progress toward smaller semiconductor nodes under Moore's law, with code available at https://github.com/TorchOPC/TorchLitho.

Abstract

The rapid evolution of the electronics industry, driven by Moore's law and the proliferation of integrated circuits, has led to significant advancements in modern society, including the Internet, wireless communication, and artificial intelligence (AI). Central to this progress is optical lithography, a critical technology in semiconductor manufacturing that accounts for approximately 30\% to 40\% of production costs. As semiconductor nodes shrink and transistor numbers increase, optical lithography becomes increasingly vital in current integrated circuit (IC) fabrication technology. This paper introduces an open-source differentiable lithography imaging framework that leverages the principles of differentiable programming and the computational power of GPUs to enhance the precision of lithography modeling and simplify the optimization of resolution enhancement techniques (RETs). The framework models the core components of lithography as differentiable segments, allowing for the implementation of standard scalar imaging models, including the Abbe and Hopkins models, as well as their approximation models. The paper introduces a computational lithography framework that optimizes semiconductor manufacturing processes using advanced computational techniques and differentiable programming. It compares imaging models and provides tools for enhancing resolution, demonstrating improved semiconductor patterning performance. The open-sourced framework represents a significant advancement in lithography technology, facilitating collaboration in the field. The source code is available at https://github.com/TorchOPC/TorchLitho

Figures (5)

• Figure 1: (a) Different representations of illuminator source, conventional source, annular source, dipole source, ans quasar source. (b) Core components of forward lithography process.
• Figure 2: Automatic differentiation flow.
• Figure 3: Comparison of Abbe's and Hopkins' imaging.
• Figure 4: The visualization of the differentiable lithography chain.
• Figure 5: (a) The result comparisons on runtime; (b) The result comparisons on EPE.