Post-CMP cleaning effectiveness is best assessed by studying the cleanliness of polished wafers obtained directly from the CMP process. In this study, a Taguchi DOE was designed as an L27 orthogonal array, and 3 levels of ten CMP and ultrasonic parameters were incorporated in the experimental matrix. The CMP parameters studied were: CMP velocity, pressure, slurry concentration, pH and temperature. The sonication parameters investigated were: frequency, amplitude, pH of the cleaning solution, surfactant type and concentration. The output properties measured were erodability and cleanability. High Level Analysis (HLA) of the DOE data indicate that under conditions of CMP that can embed particles with greater force, lower-frequency (more cavitational) conditions are more effective in particle removal. In general, ultrasonic frequencies in excess of 132 kHz are ineffective in removing strongly-adhered or embedded particles. Ionic surfactants are, in general, more conducive to CMP residue removal compared to non-ionic surfactants (natural, or man-made).
Introduction to CMP and post-CMP CleaningWhile chemical mechanical planarization (CMP) is in widespread use for forming device interconnect structures, post-CMP cleaning remains a critical aspect of achieving high reliability and process yield. Zantye et al.(1) indicate that CMP is the planarization method of choice in the microelectronic industry today, and also dwell on post-CMP contamination and cleaning issues that remain unresolved. Shallow-trench isolation (STI) using CMP is an attractive isolation alternative to local oxidation of silicon (LOCOS) for deep-sub-micron processes, but CMP leaves the surface contaminated with slurry particles and metallic impurities which may damage VLSI patterns and introduce crystal defects in Si wafers; these must therefore be removed by post-CMP cleaning. Wang et al. (2,3) have described a multi-chemical spray cleaning process for this application. Tan et al. (4) have compared various cleaning processes for low-k dielectric trench structures, with N 2 /H 2 plasma treatment followed by wet clean yielding the best results for contaminant removal. Zhang et al. (5) have optimized post-CMP cleaning via statistically-designed experiments involving double-sided brush scrubbing and spin-rinse drying. Yeh et al. (6) have described a post-CMP cleaning procedure for copper using buffered HF solution and ozone water. This combination minimizes organic as well as inorganic residue.Ein-Eli and Starosvetsky (7) have reviewed post-CMP cleaning solutions for copper. They emphasize the fact that particulate contamination of the polished surface depends strongly on CMP conditions, slurry type and hardness of the slurry layer, and point out that particle adhesion forces increase as a function of time; hence, effective just-in-time
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.