CMP Retaining ring

The progressing miniaturisation and complex nature of semiconductors are challenging and driving the advancement of new semiconductor fabrication equipment. A major concern for all processes is cmp wafer defects that can cause device failures and significantly affect yield and cost of ownership. Among different influencing factors, chemical mechanical polishing (CMP) induced defects are a critical yield concern, directly impacting device performance and yield. While the industry research and development focus has been mostly on consumables such as pads, slurry, and conditioners, the cmp retaining ring is now increasingly under investigation, as it also has a major impact on the yield and overall CMP performance. 

Ensinger, one of the leading suppliers of high performance plastic materials for CMP retaining rings, has cooperated with an outside independent laboratory, to evaluate the impact of different cmp ring materials on CMP performance. A technical article was published in the Electrochemical Society Journal of Solid State Science and Technology titled “Tribological, Thermal, Kinetic, and Pad Micro-Textural Studies Using Polyphenylene Sulfide CMP Retaining Rings in Interlayer Dielectric Chemical Mechanical Planarization.” (See study, the product is called TECATRON SE in the article, which is the former name of TECATRON CMP). This case study is an overview of that article with an emphasis on the key findings. 

CMP testing was done using 300 mm cmp retaining rings fabricated from Ensinger’s unfilled PPS tube TECATRON CMP and an industry standard PPS, both were manufactured using the extrusion manufacturing process. DuPont IC1010^® polishing pad with a K-groove design was used during testing with Fujimi PL-4217 slurry that has abrasive silica particle content of 10 percent by weight was used at a constant flow rate of 250 ml/min. This particular slurry is primarily intended for use in inter-layer dielectric (ILD) applications. A 3M A165 conditioning disc was utilised with a rate of 95 RPM while sweeping 10 times per minute across the pad surface with a downforce of 10 lbf. The platen rotation rate was held constant at 50 RPM.

A key finding has been that different PPS retaining ring materials can modify the micro-texture of the pad, cause different levels of pad fragments and shavings, and also result in different shapes of pad asperities. TECATRON CMP resulted in lower levels of pad fragment generation during wearing, less obscured pores and entrained pad debris and shavings, and sharper asperity tips compared to industry standard PPS. With sharper asperity tips and lower entrainment of fragments and other types of debris, the resulting pad micro-texture of TECATRON CMP potentially reduces the chances of wafer-level defects. Additionally, the micro-texture of the pad created by the TECATRON CMP potentially can improve material removal rates. Both retaining rings caused the pad to wear at comparable rates. However, regarding the ring wear rate, TECATRON CMP seemed to wear faster than industry standard PPS. This was deemed inconsequential in high volume manufacturing because the pads are changed out regularly, and the CMP rings are changed after a certain number of pad changes depending on the application. Hence, small differences in the ring wear rate do not impact the life of the ring.