查看文章

The metal-oxide-semiconductor (MOS) structure plays a central role in integrated circuits including as the active channel in field effect transistors. Intense research in the 1960s demonstrated that oxidizing Si into SiO2 creates a high-quality oxide-semiconductor interface; despite interest in novel materials to replace it, SiO2 remains ubiquitous as a dielectric oxide, with popular estimates suggesting that more than 1022 SiO2-based transistors have been manufactured to date. An important parameter for the performance of these capacitors is the threshold voltage Vth (the voltage at which the device changes from depletion to inversion). Imperfections such as charge traps in the bulk of the insulator or at the interfaces can create unwanted modifications to Vth and other critical device parameters. A textbook example of such a trap is that naturally occurring at the interface between the crystalline Si and amorphous SiO2, where the misfit nature of the interface gives rise to unpassivated bonds [1]. From Poisson's equation, the presence of such charge traps directly modifies the electric potential of the MOS capacitor, modifying its switching characteristics including Vth. If, further, such charges are mobile (eg Na contaminants), long term stability issues arise.

Given the importance of understanding MOS performance, a wide array of characterization techniques has been used to probe charge in these structures since their initial development [2]. The majority of these are “global” probes such as internal photoemission [3] which, while quantitative, are unable to address the location of charges within the MOS structure. Such information is crucial in enabling the …