Comparative study of tunneling currents through silicon dioxide and high-κ dielectric hafnium oxide partly embedded with nanocrystals and nanotubes in metal oxide semiconductor structures

Abstract

We have investigated the influence of silicon nanoparticles and carbon nanotubes as the charge storage element embedded in a narrow layer in the thin gate dielectric of metal oxide semiconductor (MOS) structures. The gate dielectric is assumed to consist of two layers. The thin pure oxide near the silicon substrate is the first layer and on top of it is the composite oxide layer embedded with silicon nanocrystallites or carbon nanotubes forming the floating gate of the MOS structures. We have studied the tunneling currents through the barrier formed at the interface of the silicon substrate and the embedded gate dielectric. Mainly, two types of tunneling phenomena have been investigated. One is the direct tunneling and the other is the Fowler–Nordheim (FN) tunneling at low and high applied gate voltages, respectively. Three parameters of the composite gate dielectric, such as the barrier height, the effective mass, and the dielectric constant have been modified in comparison with the pure oxide gate dielectric to calculate the tunneling currents. In our MOS structure model, we have used the silicon nanocrystallites and single-walled semiconducting carbon nanotubes as the embedded inclusion particles in the gate oxide of either in silicon dioxide (SiO2) or in high-κ hafnium oxide (HfO2). The direct tunneling currents in both the cases have been observed at the negative and at very low positive gate biases whereas the FN tunneling current has been observed at high applied gate bias. In terms of the tunneling currents (both direct and FN) the carbon nanotubes showed a superior electrical performance and are more effective as a charge storage node compared to silicon nanocrystallites, such as the lowest direct tunneling leakage current and the lowest FN tunneling onset voltage irrespective of the gate dielectric in which they are embedded.