Research On Computational Lithography For Nanometer-scale Circuits

With the constantly increasing integration level of integrated circuits (IC) and the smaller and smaller critical dimension (CD), IC industry evolves into nanometer scale. These will inevitably lead to the higher demands for manufacturing techniques of IC. Due to the slow development of illumination system, the advanced IC technology nodes still use the 193nm light source. Since the 90nm technology node and beyond, using illumination of 193nm lithography in IC production will lead to severe optical proximity effect (OPE). Many resolution enhancement technologies (RETs) have been proposed to compensate the presence of the OPE, such as off-axis illumination (OAI), optical proximity correction (OPC), phase shift mask (PSM), double patterning (DP) and etc. However, as the technology node has entered 45nm and beyond, the traditional RETs have been facing great challenges. Computational lithography (CL) as a new resolution enhancement technology, has been becoming one of the solutions beyond 22nm technology node. The computational lithography technologies can be divided into two classes, namely, gradient-based (GB) approach and level-set-based (LSB) approach. As the level-set-based approach has currently become the hot study topic in industrial and academic area, the research work in this thesis mainly focuses on the LSB approach, listed as in following.Regularized LSB computational lithography algorithm for IC mask synthesis. Computational lithography as a kind of pixel-based OPC treating the mask synthesis as an inverse problem, is able to get a properly optimized layout for its more flexible ability to modify the topology of the layout. However, the topology of the optimized mask is usually complicated, which has brought great challenges for mask production due to the complicated features. We put forward a new algorithm called regularized LSB computational lithography algorithm, which has the advantage of reducing mask complexity. By adding the TV term and Laplacian term, the algorithm, which successfully suppresses the isolated irregular holes and protrusions in the optimization process, reduces the complexity of mask by 40% and improves the manufacturability of the optimized mask.A new LSB computational lithography algorithm for process robustness improvement. In the manufacturing process, the mask needs to show high pattern fidelity under the condition of process variations. To satisfy above requirement, a new LSB computational lithography algorithm for process robustness improvement has been proposed in this thesis. In order to account for the process variations in the optimization, we adopt a new form of the cost function by adding the objective function of process variation band to the nominal cost. In the non-standard process conditions, such as the offset of the focus plane and the variation of the exposure energy, the optimized masks have better pattern fidelity. Compared with the LSB-ILT algorithm without PV band consideration, the new algorithm can reduce the process manufacturability index (PMI) by 41.37%.Hybrid conjugate-gradient-based LSB computational lithography algorithm. Most computational lithography algorithms are time-consuming, with slow convergence in the optimization. In order to improve the convergence speed of the algorithm, we put forward a hybrid conjugate-gradient-based LSB computational lithography algorithm, which overcomes the shortcomings of the FR and PRP methods in application. The algorithm has faster convergence speed and reduces the simulation time nearly 40%, compared with the traditional method. The optimized masks have better pattern fidelity and process robustness.