MOCVD of High-κ dielectrics

The oxide being currently used in transistors is SiO₂ (a marvelous material as far as interface with Si is concerned). But it is continuously being scaled down to shorten the gate length and suppress the short channel effect. However in doing this, direct tunneling current exponentially increases its power consumption. This is a significant problem that has directed the research focus to many high-κ dielectrics (e.g., Al₂O₃, ZrO₂ and HfO₂). With high-κ dielectrics, the desired Effective Oxide Thickness (EOT) could be achieved concurrently with a reduced current by increasing the physical thickness. The challenge with high-κ dielectrics is to achieve an EOT of at most 10Å and mobility comparable to that of SiO₂. There are related concerns - (a) the low dielectric interface layer which forms between the Si substrate and high-κ dielectric, making it difficult to reduce the EOT to under 10Å; (b) mobility degradation, which directly affects the drain current of the transistors & hence, speed of the circuits. These problems might be remedied with appropriate process conditions or using ultrathin barrier layers.

Calculation of low energy-loss spectra & energy-loss near edge structures (ELNES)
using density functional theory (DFT)

DFT is a well-established approach for the calculation of electronic properties of materials. For calculating the ELNES in case of dielectrics, a supercell with core-hole has to be simulated. Necessity of creation of a supercell makes these computations time-consuming. Low-loss features, on the other hand, arise from interactions of incident electrons with the valence electrons; supercells are not needed for calculating such features. These computations would help in understanding the electronic properties, local bonding, etc. in the high-κ dielectric materials at nanometre resolution.