Full text loading...
Schematic representation of the TD-SHG experiment on aluminum oxide passivated silicon with a corona wire. An external electric field can be applied between the surface of the metal-oxide-semiconductor structure and the corona wire. Due to this electric field, an internal field in the space-charge region in silicon is induced.
SHG response upon applying a positive corona field to an Al2O3 passivated Si wafer. The initial SHG response corresponds to the uncharged sample. This situation is schematically represented in situation (a). The strength of the electric field in the SCR of Si is represented by the magnitude of the arrow. At 0 s, a positive corona field of 2.7 kV is applied, resulting in a drop in SHG intensity. The negative doping charges are attracted to the interface, shown in situation (b). After several hundreds of seconds, the corona field was turned off, resulting in a recovery of the SHG signal to its original intensity.
SHG response upon applying a negative corona field to an Al2O3 passivated Si wafer. Similar curves were obtained for another thickness of the Al2O3 layer. The initial SHG response corresponds to the uncharged sample. This situation is schematically represented in situation (a). The strength of the electric field in the SCR of Si is represented by the magnitude of the arrow. Upon applying a −3 kV corona field, a large increase in SHG intensity is observed. This correlates to the increase in electric field in the space-charge region of Si, shown in situation (b). From a certain electric field over the oxide layer, tunneling of holes from Si to Al2O3 occurs, which reduces the effective electric field in the silicon space-charge region. Hence, the SHG intensity drops exponentially, as predicted by theory. Schematically, the situation is described by situation (c). Upon switching the corona charge off, the SHG drops logarithmically due to the removal of charges in the space-charge region in Si. However, the SHG signal does not recover to the initial intensity due to remaining positive charges in the oxide layer, enabling a larger electric field in the silicon space-charge region. In inset (d), the fitting on the time-shifted tunneling to an exponential decay of situation (c) is shown. Similar relaxation times (313 s vs. 320 s) are obtained for both the 10 nm and 30 nm thick oxide layer. In inset (e), the logarithmic decay upon turning the corona field off is shown.
Capacitance-voltage characteristics before and after negative corona poling. For the positive corona charging, there is no significant difference in the CV response ((a) vs (b)), confirming that the structure was not electrically altered during the corona poling. For the negative corona, there is a clear difference in CV response ((c) vs (d)), indicating that there is an electrical difference between the samples before and after poling. Due to the shift in flat-band potential and the increase in hysteresis in (b), it is clear that the amount of charges in the oxide layer increases and change sign to positive charges.
Article metrics loading...