The objective of this paper is to evaluate the delay impact of staggered metal filling (Metal2) on the standard cells and their associated local interconnect (Metal1). A Design Of Experiment (DOE) is used to define a large range of filling pattern shapes and positions. This set of filling patterns is then inserted in a Ring Oscillator (RO). From the filled RO simulations, the RO delay is expressed as a function of the filling pattern features. The maximal timing error between the model and the simulation is 1.3%, validating the model. The filling impact on RO delay magnifies the one introduced by the front-end process variations (PV). Consequently, the filling influence is introduced for the minimal, typical and maximal corners, defined now with Process (P), Voltage (V), Temperature (T) and Filling density (F) characteristics.
In present and future technology nodes, the insertion of metal filling can lead to timing yield losses. In this paper, we present a test structure generation method to study the impact of metal filling on interconnect timings. Such a method is mandatory when dealing with metal filling, because its insertion in real design does not follow standardized rules. The proposed test structure generation method is based on Design Of Experiment (DOE) to minimize the number of structures. The DOE approach links test structures timings to filling pattern characteristics. This timing model is first validated by comparison with the test structures electrical simulations. Then the model is used for statistical analysis about filling impact depending on metal level.
The objective of this paper is to evaluate the delay impact of staggered metal filling on the standard cells and their associated local interconnect on several metal levels. A Design of Experiment (DOE) is used to define a large range of filling pattern shapes and positions. This set of filling patterns is then inserted in a Ring Oscillator (RO). From the filled RO simulations, the RO delay is expressed as a function of the filling pattern features. The maximal timing error between the model and the simulation is 1.3%, validating the model. The filling impact on RO delay magnifies the one introduced by the front-end process variations (PVT). Consequently, the filling influence is introduced for the minimal, typical and maximal corners, defined now with Process (P), Voltage (V), Temperature (T) and Filling density (F) characteristics.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.