Study of pn junctions using genetic algorithm optimizer

Abstract

An efficient general-purpose optimizer based on "genetic algorithms" has been developed. The efficacy of the optimizer has been well demonstrated by implementing it for parameter extraction of pn junction diodes. With this technique, device parameters like built-in potential and grading coefficient for any diode can be extracted. If the grading coefficient shows the junction to be linear graded, effective diode area and doping gradient have been extracted. If the junction turns out to be an abrupt one, effective area and doping concentrations of each side have been obtained. It is also possible to obtain process parameters for other kinds of junctions with the inclusion of suitable models. This parameter extraction technique has been applied successfully to measured data for real diodes as well as simulated data with different noise levels.