Using spatial information to analyze correlations between test structure data

A modeling strategy that captures the dependence of performances on the spatial position of the chips on the wafer is presented. The information from this model can be used to determine whether the variance and correlation of parameters are due to random variation or are a function of wafer position. Simulation or knowledge from process engineers can be used to determine whether the spatial correlation between two parameters is due to a single underlying cause or multiple causes. The spatial correlation can also be used to split the correlation matrix into two parts: the first matrix would contain correlations due to spatially dependent disturbances and the second would contain those due to disturbances not spatially dependent.<
>