Optical proximity correction: A cross road of data flows

Abstract

This paper reviews the various data flows that occur in the generation and verification of optical proximity correction (OPC) for an optical lithography photomask. First we review the models that are or can be used in the OPC model-calibration flow, with some emphasis on those models that are not yet standard practice. Through an efficient selection of calibration structures, the data amounts needed for model calibrations can be usually kept reasonably small. This is much less the case in the OPC verification step (computationally first and on-wafer afterwards), where data volumes can be very large. Especially the inspection of printed test wafer, where the printability of large number of structures needs to be assessed throughout the intended process window, currently presents important challenges as “hotspot” detection and printability quantification needs to be combined with a short turn-around time. We will discuss some of the approaches that are being employed to deal with these conflicting requirements.

Optimized test pattern selection with machine learning method

“…Since the CD of the technological nodes shrink below 90nm, the resolution enhancement technology represented by the optical proximity correction 1,2 , source optimization [3][4][5] , inverse lithography techniques [6][7][8] et al becomes more and more critical to guarantee the process window of specific gauge points spreading over the optical lithographic layer. Provided a fixed illumination condition and film stack configuration, a OPC process model should be built based on the CDSEM metrology data obtained by the optical lithographic experiment.…”

VLSI Mask Optimization: From Shallow To Deep Learning

“…VLSI mask optimization is one of the most critical stages in manufacturability aware design, which is costly due to the complicated mask optimization and lithography simulation. Recent studies have shown prominent advantages of machine learning techniques dealing with complicated and big data problems, which bring the potential of dedicated machine learning solution for DFM problems and facilitate the VLSI design cycle [1,2].…”

Optical proximity correction (OPC) in near-field lithography with pixel-based field sectioning time modulation

“…The proximity effect correction method has been introduced to adjust the unintended optical interaction between field structures, and enhance pattern quality by utilizing dose correction combined with assist features for the original target pattern image, for good pattern fidelity and fabrication efficiency [29][30][31][32][33][34][35][36]. In conventional optical projection lithography used in the semiconductor industry, optical proximity correction (OPC) is extensively used to design and adjust the mask patterns affected by the imaging distortion induced by optical interference between nearby features.…”

“…In conventional optical projection lithography used in the semiconductor industry, optical proximity correction (OPC) is extensively used to design and adjust the mask patterns affected by the imaging distortion induced by optical interference between nearby features. The simplest optical model is defined according to the image intensity on the basis of convolutions of mask functions with kernels, and the calculated intensity is converted into the expected final pattern via resist and etch models [27,29,36]. Because maskless lithographic operation is intrinsically based on the serial process, the whole pattern shape is identified with a trajectory and field accumulation of single beam spot exposure in photoreactive material with no optical interference.…”