The Artificial Neural Networks Applied for Microelectronics Intergranular Relations Determination

Abstract

This paper is based on fundamental research to develop the interface structure around the grains and to control the layers between two grains, as a prospective media for high-level electronic parameters integrations. We performed the experiments based on nanoBaTiO3 powders with Y additives. All results on dielectric parameters on submicron level are the part of global values the same measured characteristics at the bulk samples. The original idea is to develop the new computing ways to network electronic parameters in thin layers between the grains on the way to get and to compare the values on the samples. Artificial neural networks are computing tools that map input-output data and could be applied on ceramic electronic parameters. These are developed in the manner signals are processed in biological neural networks. The signals are processed by using elements which represent artificial neurons, which have a simple function to process input signal, as well as adjustable parameter which has an influence to change output signal. The total network output presents the sum of a large number neurons outputs. This important research idea is to connect analysis results and neural networks. There is a great interest to connect all of these microcapacitances by neural network with the goal to compare the results in the standard bulk samples measurements frame and microelectronics parameters. The final result of the study was functional relation definition between consolidation parameters, voltage (U), and relative capacitance change, from the level of the bulk sample down to the grains boundaries.