1：Experimental Design and Method Selection:

The study compared refractive index (RI) and densitometry measurements for monitoring slurry composition changes. Three silica-based slurries were diluted with ultrapurified water (UPW) to simulate concentration variations. Linear correlations between density and RI were analyzed, and limits of detection were estimated based on instrument resolution and slope of calibration curves.

2：Sample Selection and Data Sources:

Three industrially-relevant slurries were used: Fujimi PL-7106 (copper CMP), Klebosol 1501-50 (ILD CMP), and CMC W7801 (tungsten CMP). Initial solutions were prepared according to manufacturer specifications, and UPW was added in increments from 0% to nearly 10% by volume.

3：List of Experimental Equipment and Materials:

Equipment included a Mettler Toledo Densito 30PX densitometer, a K-Patents Semicon Process Refractometer PR-33-S, a custom-built flow loop with a magnetically levitated pump (Levitronix), a Malvern Zetasizer Nano ZS for dynamic light scattering, and ultrapurified water. Slurries were from Fujimi, Merck (Klebosol), and Cabot Microelectronics.

4：Experimental Procedures and Operational Workflow:

For each slurry, initial suspensions were prepared and circulated in the flow loop at 6.5 L/min. RI was measured inline and logged every 0.33 Hz, with averages taken over 60 seconds after 20 minutes of circulation. Samples were withdrawn for offline densitometry, with five measurements per sample. Temperature corrections were applied using specified formulas. Dynamic light scattering confirmed no particle size changes.

5：5 L/min. RI was measured inline and logged every 33 Hz, with averages taken over 60 seconds after 20 minutes of circulation. Samples were withdrawn for offline densitometry, with five measurements per sample. Temperature corrections were applied using specified formulas. Dynamic light scattering confirmed no particle size changes. Data Analysis Methods:

5. Data Analysis Methods: Data were analyzed for linear trends using regression analysis. Limits of detection were calculated as the ratio of instrument resolution to the slope of the calibration curve. Standard deviations and R2 values were computed to assess repeatability and correlation.