Synthesis and Characterization of Nanostructured Materials And Tribological Studies Of CMP Retaining Rings

Abstract

The first project contained within this script refers to the controllable synthesis and characterization of novel thermally-stable silver-based particles. The experimental approach involves the design of thermally-stable nanostructures by the deposition of a thick interfacial layer of active titania between the primary substrate (SiO2 particles) and the metal nanoparticles (Ag NPs), as well as the doping of Ag nanoparticles with an organic molecule (Congo Red, CR). The nanostructured particles were composed of a 330-nm silica core capped with a granular titania layer (10 to 13 nm in thickness), along with monodisperse (5 to 30 nm) CR-Ag NPs deposited on top. The titania-coated support (SiO2/TiO2 particles) was shown to be chemically and thermally-stable, and promoted the nucleation and anchoring processes of CR-Ag NPs, which prevented the sintering of CR-Ag NPs when the structure was exposed to high temperatures. Different characterization techniques were proposed, and utlized, to examine and determine the thermal stability of the silver composites, among which, electron microscopy (i.e SEM, TEM and EDS) played an important role due to its high resolution capabilities. Based on the information obtained via high resolution imaging and elemental mapping, it was possible to resolve changes on the surface topography, crystalline structure, and chemical composition at different temperatures on the nanometer scale. The second project presented in this dissertation concerns the tribological characterization of retaining rings in Chemical Mechanical Planarization (CMP). This project was divided into several studies, all of which contributed valuable information which made the determination of tribological mechanisms possible. These eventually contributed to minimizing the environmental impact, and reducing cost of ownership (COO) of the CMP module. The first study was conducted to obtain the tribological mechanisms in the ring-slurry interface by analyzing the Stribeck and Stribeck+ curves. Both methods gave consistent results with the lubrication mechanisms starting at “boundary lubrication,” and transitioning into “mixed lubrication” as pseudo-Sommerfeld numbers increased. Rings made of different materials were analyzed (i.e. PPS and PEEK). Stribeck curves were used to infer wear rate information about the ring, which in conjunction with Stribeck+ curves could help choose process parameters that balanced wafer removal with ring wear. In the second work, tribological studies were conducted on two differently designed rings made of polyether ether ketone (PEEK), one with sharp slots and the other with rounded slots. Lubrication mechanism, coefficient of friction (COF) values, and a new parameter, referred to as directivity, were analyzed to understand the tribology in the ring-slurry interface. Raw data from the Stribeck+ curves was used to construct a new “kinetic” plot to help infer the relative wear rate of the ring, thus allowing one to choose process parameters that balanced it against wafer removal rate. Validation studies involving the extended wear of eight differently formulated retaining rings showed a correlation between the actual wear rate and the “inferred” wear rate, which renders credibility to our approach regarding the above-mentioned “kinetic” plots. The final study concerned correlating previously reported slurry bow wave widths with the tribological attributes of retaining rings having different slot designs that were subjected to the same polishing conditions as those in the bow wave width studies. Retaining rings with different slot geometries were shown to affect the slurry distribution. Tribological parameters such as COF and directivity captured the effect of the slurry distribution on the pad-ring interface. This study also demonstrated the utility of directivity as a key metric for determining polishing performance since the fluctuations in the applied forces were captured, which would have otherwise been ignored when COF was utilized.