Amine-Based Chemistries for the Megasonic Removal of Alumina Nanoparticles from Oxidized Silicon Carbide

Abstract

Increasing demand for improved Integrated Circuit (IC) technology has propelled the advancement from silicon as a semiconductor material to Wide Band Gap (WBG) materials (i.e., Silicon carbide (SiC), Gallium nitride (GaN)) due to their intrinsic properties (i.e., increased thermal stability, wear resistance, etc.). While these traits make WBGs ideal semiconductor materials, the Chemical Mechanical Planarization (CMP) process becomes difficult as these surfaces are more chemically inactive. To overcome this, the CMP process has shifted to incorporate more aggressive chemical and mechanical conditions (i.e., highly oxidative chemistries, increased shear force, etc.) to reach desired material removal rates (MRR). Traditional post-CMP (pCMP) cleaning processes combine high shear forces via a Polyvinyl Alcohol (PVA) brush with oxidative cleaning chemistries to remove CMP residue. This, however, results in secondary defect generation (i.e., scratches, pitting, brush debris) on the substrate surface. A low-shear force, non-contact cleaning process via megasonic energy is one way to combat secondary defect generation. The megasonic energy serves a dual function of reducing the substrate’s surface boundary layer, which allows for enhanced surface interaction with various cleaning solutions while also generating reactive oxygen species (ROS) that activate the chemistry. This work focuses on amine-based cleaning chemistries for oxidized SiC substrates. The introduction of the amine leads to a reduction of the oxidized surface, allowing for a weakened residue-substrate bond for an improved particle removal efficiency (PRE). Initial results have shown that amine-based standard clean 1 (SC-1) produced comparable results to previously utilized cleaning chemistries (i.e., Supramolecular chemistries), with a PRE of 11.4%. The PRE of SC-1 increases significantly in proportion to sonication power, with the highest PREs ranging between 90% to 96%. This work will provide insight into oxidized SiC surface behavior by incorporating other amine-based cleaning chemistries.