Adaptive Optical Proximity Correction (OPC) Dissection
Publication Date: 2002-Nov-11
The IP.com Prior Art Database
The quality of optical proximity correction results, as well as all proximity corrections, for is heavily dependent on the dissection, or segmentation, of the layout. A method for adaptively dissecting layouts is proposed in which the OPC program, and/or a subsequent silicon-versus-layout verification tool, checks between evaluation points for large variances, aka ripples, and adjusts the segmentation, e.g. increases it, for those areas with large variances.
1. Descriptive Title
Adaptive OPC Dissection
(1)The quality of OPC results greatly depends on the dissection. The fine dissection can produce better lithography result but the data volume and simulation time may be prohibitive. On the other hand, coarse dissection may result in poor OPC quality.
(2)The choice of OPC dissection parameters can be very difficult because of the tradeoffs between performance, data volume and accuracy. Sometimes, there is simply no one set of dissection parameters that are suitable for the whole layout.
(3)As shown in Fig 1, OPC only does simulation on one evaluation point (stars) for each dissection segment. It is possible that simulation results are all good (within the tolerance) at those locations, but the contour still exceeds the tolerance in between those evaluation points. Currently, we rely on silicon-vs-layout checking to detect the problem.
(1)It is best if OPC can detect and correct this kind of problem by itself before generating the results. This patent proposes a new adaptive dissection method which involves detecting possible errors by simulating at additional locations and then correcting them with newer and finer dissection segments.
(2)After OPC engine is finished, it can check between those evaluation points. In the case of Fig. 1, it will find a violation between the 3rd and the 4th stars. OPC
then may adaptively insert one dissection segment as in Fig. 2 and redo OPC around that region to fi...