Concept of two-dimensional swing curves for critical dimension prediction and optimization of resist/antireflective coating bilayers in topographic situations

In microlithography, the necessity of using bottom antireflective coatings (BARCs) to reduce linewidth variations resulting from reflective notching a nd thin film interference effects has been largely demonstrated. We report a new methodology, based on a 2-D swing curve concept, to predict the litho graphic performance of resist/BARC bilayers in a topographic situation to o ptimize the critical dimension (CD) range. Due to inherent planarization ef fects when coating on topography, both the resist and organic BARC thicknes ses vary, as do the final reflectivity and CDs. As a consequence, it is not easy to determine the optimal BARC thicknesses and to predict the lithogra phic performance, taking into account topography effects over the whole chi p. Usually, lithographic performance (CD swing curves) is measured or calcu lated using modeling over plane wafers. Over the topography of a real chip, however, the resist thickness can cover up to four periods of the swing cu rve, which means that lithographic performance over a real chip cannot be p redicted using swing curves calculated or measured on plane wafers. We ther efore propose a new method of representing the reflectivity and CD swing cu rves in two dimensions to evaluate the lithographic performance of the BARC /resist bilayer. A simulation algorithm is performed that enables the deter mination of both the optimal BARC thicknesses and the lithographic performa nce window over the whole chip. Practical examples are given demonstrating the role of such a simulation. (C) 2000 Society of Photo-Optical Instrument ation Engineers. [S0091-3286(00)02503-4].