Interface defect engineering for high-performance MOSFETs with novel carrier mobility model: Theory and experimental verification

As the conventional hydrogen-termination method has a limited ability to improve the interface quality between SiO2 and its Si substrate, an alternative termination method to reduce the influence of interface states is necessary. Interface engineering using first-principles calculations to suppress the influence of interface states is proposed based on the findings that silicon with dangling bonds is their primary origin. First-principles calculations indicate that the interface states can be terminated with oxygen when incorporated into the SiO2/Si interface without additional oxidation, which generates other interface states from an appropriate oxygen-anneal process. It is experimentally shown that such an oxygen termination can be realized in slow and low-temperature annealing, and the oxygen-termination method is a promising alternative for hydrogen termination. The stronger Si–O bond introduced from the oxygen termination compared with the Si–H bonds from hydrogen termination ensures a better interface quality. As one oxygen atom terminates two silicon atoms, the oxygen-termination method can efficiently suppress the number of interface defects compared with hydrogen and fluorine termination. The mobility degradation due to the interface states was improved more from oxygen termination than from hydrogen termination because the strength of Coulomb scattering due to Si–O dipoles is reduced from the heavier oxygen mass. Theoretical predictions were verified using experiments, indicating that the oxygen-termination method under appropriately optimized annealing conditions (speed and temperature) is a promising candidate to improve the interface quality by reducing the influence of interface states.