Yield Estimation of Molecular QCA Memory Structures with Geometric Analysis

Nanoelectronic devices and circuits are likely to contain multiple defects and exhibit faulty behavior, especially as larger circuits are built. In this work, we will present a new method for estimating the yield of memory structures constructed from molecular QCA devices. This method takes patterns of missing cells that were found, via full physical simulation, to cause wire failures and compares them to the geometry of a defective memory-like circuit to determine if the memory-like circuit operates properly. The memory-like structure used here is ´snake´ shaped and is a series of straight and bent wire segments. We vary the number of cells between the corners (the aspect ratio), the cell width of the wire segments, the total length of the memory structures, and the percentage of cells that are removed. Our results show that this new method is viable for yield estimation. We also show that computational speedups of several orders of magnitude are achieved by using this method.