Behavior of SiLECT 6000, Direct STI, CMP Slurry During Local Recirculation with Frequent System Shut Downs

 

 

Richard Viscomi, Matt Federau, Mark Diaz: Cabot Microelectronics, Aurora, IL
Leon Han, Ted McGill: ChemFlow Systems Inc., San Jose, CA

 

 

 

Background

 

The goal of this evaluation was to test the suitability of the ChemFlow CFB for the distribution of Cabot Microelectronics’ SiLECT6000 slurry and to determine the effect of frequent system shut downs (and subsequent restarts) on the stability of the abrasive particles.

 

SiLECT™ 6000, a single pack, ceria based Direct STI CMP slurry supplied by Cabot Microelectronics was pre-agitated using a Yamada DP-20 diaphragm pump prior to loading into a ChemFlow CFB slurry delivery system. Once in the system, the slurry was recirculated through an internal tool recirculation loop plumbed with ¾-inch PFA tubing for approximately 5 days at a flow rate of 2 gallons per minute.  Throughout the course of testing, the slurry was subjected to frequent system stoppages with extended periods of inactivity where particle stratification would be a possibility.

 

The CFB has the capacity to blend up to 3 different chemical components with a make-up rate of 1.5 gallons per minute and a maximum flow rate of 2 gallons per minute.  Blend accuracy is achieved through the patented U-Tube technology, which unlike other systems, does not rely upon flow meters or other moving parts.  Blended chemical is gravity transferred from an internal tank to the dispense vessels and from there, pressurized to the polishing tools.

 

The main objective of this study was to: 1) determine the viability of the CFB to distribute SiLECT™ 6000; 2) evaluate the effects frequent system shutdowns (on slurry stability) with no slurry recirculation for extended periods of time.  Slurry was circulated within the system for 8 hours per day and then the system was deactivated for the remainder of that day.  The system would be re-activated again on the following day according to the above schedule.  Slurry samples were collected at the beginning and end of each day for analytical and CMP polish testing.

 

 

Results Summary

 

·        Under the above conditions (recirculation for 8 hours followed by stagnation for 16 hours) SiLECT™ 6000 slurry exhibited no significant changes in the normal slurry health parameters: pH, density, and wt. % solids. There was only minor, visible settling of the ceria particles, despite the extended periods of no slurry circulation.

 

 

·        SiLECT™ 6000 contains suspended colloidal ceria particles. This slurry was found to have much slower settling characteristics compared to most ceria based slurries, which settle very rapidly. Recirculation at 2 gpm was sufficient to fully resuspend any ceria particles that may have settled during the period of time when the system was idle.

 

·        SiLECT™ 6000 requires an initial pre-use mixing in the supply container with an external drum/tote impeller mixer.  After the initial pre-use agitation, re-circulation should be sufficient to maintain slurry homogeneity, while continuous slow speed mixing in the daytank is not necessary.  Very minimal settling of the abrasive particles was visually observed after a 48-hour simulated fab shutdown. Wt % solids measurements confirmed that settled abrasive particles redisperse quickly (in < 5 minutes) as soon as the flow in the distribution loop was restarted.

 

·        Agitation at 1725 rpm for 20 minutes using a 0.33 hp mixer equipped with a 3.4” A-310 impeller in a 55-gallon drum of slurry is recommended to redisperse the ceria particles before first use of the slurry. This particle dispersion can be maintained by intermittent mixing of the drum at 1725 rpm for 5 minutes every 4 hours.  Alternately, a bellows pump at 3 gpm for at least 10 minutes every 4 hours may be employed to maintain particle dispersion after initial agitation by impeller mixer.

 

·        The particle distribution of the ceria in SiLECT™ 6000 did not change throughout the course of the experiment.  This demonstrates that any particles, which settled during the system inactivity, did not form hard agglomerates, which could lead to increases in defectivity during polishing.

 

Data

 

 

 

Available upon request.

 

 

 

 

Conclusion

 

This handling test went according to plan and no setbacks were observed. The slurry parameters remained stable despite the multiple, extended system shut-downs. The ChemFlow CFB showed the capability of handling and dispensing SiLECT™ 6000, its small footprint make it suitable for R&D develoment work or as a point-of-use production tool capable of delivering SiLECT™ 6000 to CMP polishers.

 

 

 

 


About ChemFlow Systems

 

ChemFlow Systems is a supplier of simple u-tube technology for slurry and chemical blending systems in the semiconductor, disk drive, and biotech industries. They offer simple solutions for complex problems. They are located in the San Jose area and serve a world market. Information about ChemFlow Systems, Inc. can be found at: www.chemflowsys.com, or e-mail at:.

customerservice@chemflowsys.com

 

 

About Cabot Microelectronics

 

Cabot Microelectronics is the world's leading developer, manufacturer and supplier of high-performance polishing slurries for chemical mechanical planarizations--formally known as integrated circuit (IC) devices--and data storage components.  The Company is also an emerging supplier of specialized CMP polishing pads for use with its slurries; www.cabotcmp.com.