Analysis of Large Particle Count in Fumed Silica Slurries and Its Correlation with Scratch Defects Generated by CMP

Edward E. Remsen,Sriram Anjur,David Boldridge,Mungai Kamiti,Shoutian Li,Timothy Johns,Charles Dowell,Jaishankar Kasthurirangan,Paul Feeney
DOI: https://doi.org/10.1149/1.2184036
IF: 3.9
2006-01-01
Journal of The Electrochemical Society
Abstract:The large particle count (LPC) of fumed silica slurries was evaluated and correlated with scratch counts created on SiO2 films by table-top, chemical-mechanical planarization (CMP). Particle sizing results obtained by static light scattering, capillary hydrodynamic fractionation, and dual-sensor single particle optical sensing (SPOS) pointed to the latter as the superior method for quantitative analyses of the LPC. Dual-sensor SPOS is a new technique that determines the LPC on a silica sphere-equivalent, light-scattering diameter scale for particles as small as 0.469μm . LPC measurements used in combination with dark-field optical microscopy for scratch metrology afforded linear correlations between scratch counts and the LPC. Particles producing scratches had silica sphere-equivalent, light-scattering diameters exceeding 0.68μm . Inclusion of these particles in the LPC produced a two-fold increase in the number of scratch-forming particles in the correlation relative to correlations generated via single-sensor SPOS measurements of LPC. Experimental uncertainty in scratch counts limited the correlation as a scratch predictor. Slurries differing in LPC by a minimum of 1.8×105particles∕gslurry had statistically different predicted scratch counts at the 95% confidence level. Additional method development is needed to extend LPC-based scratch prediction to other CMP processes producing scratch defects.
electrochemistry,materials science, coatings & films
What problem does this paper attempt to address?