SlurrySlurry Handling,Handling, TroubleshootingTroubleshooting andand FiltrationFiltration Levitronix CMPUG Symposium Budge Johl – Phoenix, AZ February 17, 2005 Key CMP Issues & Their Causes Three key CMP issues encountered in field: High defects/microscatches on wafers Change in polishing performance (removal rate, nonuniformity) Short filter lifetime or reduction in lifetime Associated causes of these three key issues: Large particles in slurry Change in slurry particle size distribution High gel content in slurry Terminology of Particles Observed with SEM, ESEM & Light Microscope Particle: single solid sphere or other geometry Aggregate: multiple particles chemically attached to each other (chain), primarily via siloxane bonds “Agglomerate”: particles and aggregates come together into close-packed clumps due to entanglement, drying, and coagulation, denser than the surrounding solution Microgels: aggregates link together and form three dimensional network with water trapped within; same density and refractive index as the surrounding solution (Filter Plugging Problem) Particles Aggregate Agglomerate Microgel Ref.: R. Iler, Chapter One , “The Chemistry of Silica”, John Wiley & Sons, 1979 Potential Defect Causing Agglomerates & Large Particles Gels Membrane with Wet Gel & Particles Same Membrane with Dry Gel & Particles Dry Gels Wet Gels Large Particles Silica slurry was diluted with pH adjusted water, then filtered through a 3 micron Isopore membrane. Possible Sources of Large Particles & Gels: Agglomeration Local drying in shipping container, tank, and fittings Settling pH shock during dilution Temperature fluctuation during shipping and storage Shearing (i.e. certain valves, pumps, piping design) Improper Mixing Large # of Turnovers Lack of PM’s on Distribution System Lack of proper slurry conversion protocols from one slurry type to another, including flushing and system sterilization protocols. Oxide Defects “Large Particles” in Slurry Create Micro-Scratches & Embedded Particles on Wafers Typical defects in oxide CMP Ref.:C. Dennison, KLA-Tencor, Micro, February 1998 W-CMP Defects Nine types of tungsten CMP defects Ref.: C. Dennison, KLA-Tencor, Micro, February 1998 Large Scratch Cu-CMP Defects Large Chatter Marks Cu-CMP Defects Optical image (Compass®) SEM images (SEMVision™) AFM (Veeco) Gouge Cu-CMP Defects Optical image (Compass) SEM images (SEMVision) AFM (Veeco) MicroMicro ScratchScratch Cu-CMP Defects Optical image (Compass) SEM images (SEMVision) AFM (Veeco) Slurry Troubleshooting Takes Detective Work Slurry Trouble Shooting Methods Typical Problem Encountered Associated Cause of Problem Troubleshooting Method Defects including Microscratches Large Particles in Slurry • Particle Counting Eq. (LPD’s on Wafers) • SEM Analysis • EDX (Elemental Signature) Change in Polishing Performance Change in Mean Particle Size • Mean Particle Analyzer Distribution or other parameter • % Solids Measurement • Density • pH • Conductivity 250 • % Assay 200 • Viscosity 150 100 • Zeta Potential 50 0 30N50 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Short or Reduced Filter Lifetime Gel Content of Slurry • Delta p measurement • SEM analysis of used filter •ESEM analysis of slurry Analytical Labs Play Key Role for Slurry Handling, Troubleshooting and Filtration Total % Solids Density / Specific Gravity pH Conductivity Viscosity Mean Particle Size Large Particle Size Filtration Testing & Optimization (Source, Post Mix, Global Loop & POU) Zeta Potential Percent Assay & Assay Development Slurry Characterization, Distribution & Handling Evaluations, Slurry Conversion Protocols Drum Contamination Clean drum surface to remove dust particles before opening When removing lids bung holes should be carefully cleaned before pouring out slurry. Carefully clean the slurry container lids & bung holes before closing left over slurry pails or drums to prevent introducing dry slurry particles. Dip tubes should be clean before introducing into new slurry container. Proper Mixing of Slurry CUS1201A Mixing Time @ 500 RPM Time (minutes) 0 5 10 15 20 25 29 29.5 30 30.5 31 31.5 Some Mixing 32 32.5 Top Bottom Factors PercentTotal Solids 33 Size of Tote Age of Slurry Note: Slurry allowed to remain undisturbed for several weeks for Time 0 sample. Number of Propellers Type of Mixer and Propeller Direction of Mixing RPM Propeller Scraping Side of Drum Minimizing Particle Growth Humidification Need to humidify the air pockets that come in contact with the slurry with >95% humidity. Need to avoid drying of the slurry which can result in formation of agglomerations and aggregates. Non-Humidified Effect of Humidification in CDM’s Effect of CDM’s on Slurry Health Counts Large Particle Population Distribution 500000 450000 400000 Control 350000 CDM-I 300000 250000 CDM-II 200000 150000 CDM-III 100000 50000 0 0.5 1 1.5 2 2.5 Particle Diameter (µm) -> Levitronix Pump Study Magnetic Levitation Technology Klebosol® 1501-50 Dynamic Levitronix Pump Study Storage Tank Turnovers 0 250 500 750 1000 1250 1500 1750 32.5 6 5 30.0 4 Total Percent Solids Viscosity (cP) 27.5 Density (g/mL) 3 2 Total Percent Solids 25.0 1 (g/mL)(cP), Density Viscosity 22.5 0 0 50 100 150 200 250 300 350 Time (Hours) Days 02468101214 0 Klebosol 1501-50 -25 -50 Zeta Potential (mV) Dynamic Levitronix Pump -75 0 50 100 150 200 250 300 350 HOURS Days 02468101214 150 125 Klebosol 1501-50 100 75 50 Mean PSD(nm) 25 Dynamic Levitronix Pump 0 0 50 100 150 200 250 300 350 HOURS Klebosol 1501-50 Dynamic Levitronix Pump Study 6000 5000 Diameter) 4000 T0hrs T24hrs (122 TO) T144hrs (730 TO) T336hrs (1700 TO) 3000 2000 1000 Cumulative Number> (# Part 0 0 5 10 15 20 25 30 35 40 45 50 Particle Diameter (microns) Klebosol 1501-50 Dynamic Levitronix Wafer Removal Rate Response 2.0 STANDARD (0 Turnovers) RR LEVITRONIX (1700 Turnovers) RR 1.5 1.0 Normalized RR Data 0.5 0.0 123 Polisher Head # Klebosol 1501-50 Dynamic Levitronix Pump Defectivity Response 4.0 3.5 STANDARD (0 Turnovers) 3.0 LEVITRONIX (1700 Turnovers) 0.1µm ) 2.5 2.0 1.5 1.0 Normalized> Defect Data ( 0.5 0.0 Head 1 Head 2 Head 3 Polisher Head # Filtration Wonder if this filter will work? Possible Filter Locations Dilution Day Tank Shipping Cleaning Container Chemical Polishing Filtration DI Tool Filtration Water Track Post CMP POU Cleaning Filtration Intake Post Recirculated Filtration Dilution Recirculation DI Filtration Loop Filtration Filtration Filter Retention Curves Filter Ratings Are Not All Created Equal Filter Retention Curves are Very Useful to Determine Effective Micron Size of the Filter at various Percent Retentions Retention Efficiency Curves for Planargard CMP Filters 100 CMP1 CMP3 80 CMP5 60 CMP7 CM11 40 CMP9 Percent Retention CM13 20 CM16 0 0 5 10 15 20 25 30 Particle Size (microns) Defectivity Improvement EPL2362 Defectivity Response 4.0 Dynamic 3 µms Global Loop Filtered m) µ µ µ µ Unfiltered (Static) 0.1 3.5 > Point-of-Use 1.5 µms Filtered (Static) 3.0 2.5 2.0 1.5 1.0 0.5 Normalized Defect Data Respect to Day 0 ( 0.0 0 24 72 144 192 Time (Hours) Large Defectivity Improvement Seen with Global Loop and POU filters. Loop filter : no RR loss Filter : Mykrolis Corp., Planargard CMP3 10” If the filter is optimized for the slurry, there is no RR loss. Foreign Materials Found On Filter Foreign Materials Can Decrease Filter Life Such as Plastic Shavings/Misc. Some Foreign Materials Statically Attracted to PFA Tubing Some Other Possible Sources Unclean Source Container Mixing Blade Improperly Adjusted & Scrapes Side of Drum Dirty Dip Tubes Unclean Environment where Drum Opened Improperly Cleaned Delivery System Filtration Summary CMP slurry filtration may reduce defects and improve process consistency POU slurry filtration allows the use of more retentive filters Higher retention provides better defect reduction Filtration in slurry delivery systems will extend POU filter’s lifetime Filter lifetime is highly slurry and process dependent Lifetime can be monitored by pressure drop and/or flow rate Conclusion Understanding slurry characteristics, handling properties and filtration is key for maintaining distribution systems, troubleshooting and minimizing large defect causing particles. Rohm and Haas Electronic Materials actively seeks to collaborate with distribution system suppliers, filter companies & others to provide the best solutions for the customer. Together as suppliers we can achieve optimized solutions to fit customer needs and to keep them on target. Acknowledgements BOC Edwards and FSI International Publications (John Bare & Budge Johl) Mykrolis (Millipore) Ken Adrian, Budge Johl, Geanne Vasilopoulos & Gregg Conner Rohm and Haas Electronic Materials (Team) Levitronix® – Provided Pump for Evaluation Compass and SEMVision are trademarks of Applied Materials, Inc. Klebosol is a trademark of AZ Electronic Materials. Planargard is a trademark of Mykrolis Corp. .