Drift modeling of electrically controlled nanoscale metal-oxide gas sensors

Gas sensors with small dimensions offer the advantage of electrical sensitivity modulation. However, their actual use is hindered by drift effects that exceed those of usual metal-oxide sensors. We analyzed possible causes and found the best agreement of experimental data with the model of internal dopant fluctuations. The dopants are oxygen vacancies exhibiting high drift-diffusion coefficients under the impact of electrical fields. Thus, the width parameters of space charge regions, which again control the sensor current, are undergoing slow changes. Moreover, the dopant distributions cause internal electrical fields that yield drift even after voltage switch-off. This behavior has been proven by simulations based on the literature values, using a converging combination of the classical electron drift-diffusion and Poisson equations with the Fokker-Planck solution for the dopants, which is of general relevance to other nonperfect semiconductor devices.