Thesis for a Candidate Degree in Technical Sciences: Specialty 05.27.01 – Solid State Electronics. – National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", Ukraine, Kyiv, 2020.
The dissertation is devoted to the development of physical bases of insulator choosing for MIS structures with a certain semiconductor substrate and to the study of cerium dioxide films as an alternative to silicon dioxide film.
The influence of the dielectric on the electrophysical parameters of the MOS transistor with induced channel is analyzed. It is shown that almost all parameters and characteristics of the transistor depend on the properties of the dielectric. The use of a dielectric with high dielectric constant (high-k dielectric), instead of the classic silicon dioxide, avoids tunneling current through the dielectric by increasing the film thickness, also increase the breakdown of the MIS structure. It is also shown that the existing insulator choosing criterion are designed exclusively for the silicon substrate, do not take into account the influence of the insulator-semiconductor interface and do not give unambiguous results.
Not only the dielectric constant affects the characteristics of the MIS device , but also the density of the effective charge in the insulator. The nature of this charge components (the charge of mobile ions, the trapped charge in insulator, the fixed charge in the insulator and the surface trapped charge) is studied. The first two lead to instability of the device characteristics, but can be eliminated by improving the technological process. The last two are due to the presence of broken connections in the insulator and at its boundary with the semiconductor. For a particular semiconductor substrate they depend only on the insulator parameters. That is why the minimization of these charges density was chosen in the development of the choosing criterion of insulator for MIS structures.
A general criterion for choosing a dielectric has been developed. It is based on the minimum difference in the density of broken bonds at the dielectric-vacuum interface and at the semiconductor-vacuum interface. There was a need to build a model of the crystal-vacuum separation boundary to determine the number of broken bonds at this boundary. This turned out to be impossible within the classical (nodal) method of describing the crystal lattice. this is impossible within the classical method of describing the crystal lattice. Therefore, an alternative - interstitial method of describing the crystal lattice was developed, which physically and mathematically does not contradict the classical aspect. Based on the phenomenology of the internodal aspect, in the first approximation, the density of broken bonds at the crystal-vacuum interface is inversely to the square of the primitive loop of the flat grid at this interface. In the general case, this area is difficult to determine due to the large number of possible orientations of the substrate, or almost impossible, due to the amorphousness of the dielectric film. Therefore, to generalize the proposed criterion, a universal parameter is introduced - the average bond length, which characterizes the average distance between the components of the crystalline substance. Under these conditions, the area of the loop of the flat grid on the surface of the crystal is calculated as the square of the average bond length. To confirm the acceptability of the use of this parameter, a relationship has been established between it and the electron work function of the crystal. The difference between the density of broken bonds at the dielectric-vacuum and semiconductor-vacuum boundaries is calculated. This calculation showed that for a silicon substrate, from a number of dielectric applicants according to the developed criterion, the most suitable are Dy2O3, CeO2 and La2O3. Cerium dioxide was selected for the experiment. Analysis of methods for producing cerium dioxide films on a silicon substrate showed that, regardless of the method, they have the structure of CeO2 nanocrystals in the amorphous Ce2O3 phase. To confirm the effectiveness of the use of cerium dioxide in MIS structures, aluminum-cerium dioxide-silicon structures were fabricated by two methods - the metal mirror oxidation method and the flash method. The quality of the cerium-silicon dioxide interface was evaluated by the charge density at its boundary. The paper proposes a simplified method for determining the charge density at the dielectric-semiconductor interface according to the experimental volt-farad characteristics of MIS structures. It is shown that the effective charge density at the cerium dioxide-silicon interface is not more than at the silicon dioxide-silicon interface. Thus, the replacement of silicon dioxide by cerium dioxide, allows to reduce the tunnel currents through the dielectric and does not impair the quality of the dielectric-silicon interface.
