Spectroscopic Investigation of Complex Redox Equilibria in Cupric Chloride Subtractive Etching for Cu High Density Interconnects

Abstract

The continuously increasing demand for innovation in the miniaturization of microelectronics has driven the need for ever more precise fabrication strategies for device packaging, especially for printed circuit boards (PCBs). Subtractive copper etching is a fundamental step in this processes, requiring very precise control of etch rate and etch profile. Cu etching baths are typically monitored with several parameters including oxidation-reduction potential, conductivity, and specific gravity. However, the etch rate and etch profile can be difficult to control even under strict engineering controls of those monitoring parameters. The basic understanding of the Cu etching is shown in the figure below [Fig 1], whereas the reality is that the mechanism of acidic cupric chloride etching, regeneration and recovery is complex [Fig 1], and the current monitoring strategies can have difficulty controlling the complex interlocking chemical equilibria. We report that thin-film UV-Vis spectroscopy has the capability to effectively monitor the electrochemical processes and its complex redox changes to the etch bath from beginning to end. We report that the complex equilibria has direct control on the etch rate and its regeneration process. UV-Vis spectroscopy also reveals various underlying mechanism reasons for etch bath behavior and illuminates the roles of H+ and Cl- to the etch bath while also providing a means to monitor the Cl-. By implementing thin film UV-Vis, we are able to probe and monitor the equilibria in real time to maintain a consistent etch rate throughout while also achieving proper regeneration and recovery for repeated use. Additionally, we’ve shown that this technique is sufficient in supplementing oxidation-reduction potential, conductivity, and specific gravity methods. This addition be can utilized to improve current monitoring strategies as it can identify and predict etching behavior that the current standard methodologies may have difficulty predicting. Figure 1