Proceedings Article | 31 March 2014
Proc. SPIE. 9052, Optical Microlithography XXVII
KEYWORDS: Semiconducting wafers, Global system for mobile communications, Optical lithography, Photomasks, Finite-difference time-domain method, Computer simulations, Silicon, Lithography, Wafer-level optics, Optical proximity correction
Reflection by wafer topography and underlying layers during optical lithography can cause unwanted
overexposure in the resist [1]. In most cases, the use of bottom anti reflective coating limits this effect. However, this
solution is not always suitable because of process complexity, cost and cycle time penalty, as for ionic implantation
lithography process in 28nm bulk technology. As a consequence, computational lithography solutions are currently under
development to simulate and correct wafer topographical effects [2], [3]. For ionic implantation source drain (SD)
photolithography step, wafer topography influences resulting in implant pattern variation are various: active silicon
areas, Poly patterns, Shallow Trench Isolation (STI) and topographical transitions between these areas. In 28nm bulk SD
process step, the large number of wafer stack variations involved in implant pattern modulation implies a complex
modeling of optical proximity effects. Furthermore, those topography effects are expected to increase with wafer stack
complexity through technology node downscaling evolution. In this context, rigorous simulation can bring significant
value for wafer topography modeling evolution in R and D process development environment. Unfortunately, classical
rigorous simulation engines are rapidly run time and memory limited with pattern complexity for multiple under layer
wafer topography simulation.
A presentation of a fast rigorous Maxwell’s equation solving algorithm integrated into a photolithography
proximity effects simulation flow is detailed in this paper. Accuracy, run time and memory consumption of this fast
rigorous modeling engine is presented through the simulation of wafer topography effects during ionic implantation SD
lithography step in 28nm bulk technology. Also, run time and memory consumption comparison is shown between
presented fast rigorous modeling and classical rigorous RCWA method through simulation of design of interest. Finally,
integration opportunity of such fast rigorous modeling method into OPC flow is discussed in this paper.