Determination of Ultratrace Impurities in Semiconductor Photoresist Using ICP-MS/MS

Significance of the Topic

The rapid scaling of integrated circuit fabrication to process nodes of 7, 5 and 3 nm and the roadmap toward 2 nm devices impose stringent purity requirements on all process chemicals. Photoresist is a key material in photolithography and must be free of metal contaminants at sub-ppt levels to avoid defects in circuits containing hundreds of millions of transistors per square millimeter.

Objectives and Study Overview

This application note evaluates the performance of the Agilent 8900 ICP-MS/MS for ultratrace analysis of 20 metallic contaminants in IC-grade photoresist. Key aims include

• Achieving sub-ppt detection limits for all critical elements
• Assessing simple sample preparation by dilution in PGMEA
• Demonstrating interference removal in a high carbon organic matrix
• Verifying accuracy by spike recovery and dilution tests

Used Instrumentation

The following configuration was used

• Agilent 8900 ICP-MS/MS #200 Semiconductor module
• 200 μL/min MicroFlow PFA nebulizer and quartz spray chamber
• Organic solvent sample introduction kit with solvent-resistant tubing
• Fifth gas controller for O2/Ar addition to prevent carbon buildup
• Pt-tipped interface cones and s-lens for high ion transmission

Methodology

Photoresist samples were diluted 10× in semiconductor grade PGMEA (G5, <10 ppt total metal), using ultrapure water from an UPW system. Calibration standards at 200, 500 and 1000 ppt were prepared in PGMEA via N-Methyl-2-Pyrrolidone intermediate with 2 percent HNO3. Analyses were performed in MS/MS mode with a multitune method using NH3, H2 and O2 reaction gases under warm plasma conditions (1000 W, extended sampling depth, increased carrier gas and makeup gas flows) to decompose carbon species and resolve polyatomic interferences.

Key Results and Discussion

Calibration curves for interfered elements showed excellent linearity and precision. Background equivalent concentrations and detection limits were below 1 ppt for all analytes except Pb, reflecting trace solvent contamination. Spike recoveries for a 0.1 ppb addition ranged from 91 to 110 percent. A comparison of 10× and 20× dilutions agreed within ±20 percent for 11 out of 15 measured elements, confirming matrix tolerance without internal standards. A one-hour stability test on a 10× diluted sample gave RSDs below 5 percent for analytes above 20 ppt and below 12 percent even at ultratrace levels.

Benefits and Practical Applications

• Sub-ppt detection meets current and future semiconductor purity requirements
• Minimal sample preparation reduces contamination risk and increases throughput
• Effective removal of carbon and sulfur based interferences in photoresist matrices
• No need for internal standards simplifies routine QC in fab environments
• High stability ensures consistent long-run performance

Future Trends and Prospects

• Lowering detection limits further to support emerging node requirements
• Expanding analyte panels to include novel dopants and contaminants
• Integrating inline monitoring and automation for real-time process control
• Applying AI-driven data analytics for faster decision making
• Developing advanced cell gas chemistries to tackle new interferents

Conclusion

The Agilent 8900 ICP-MS/MS configuration delivers the sensitivity, selectivity and robustness required for ultratrace elemental analysis in high carbon photoresist solvents. Simple dilution in PGMEA combined with tailored reaction cell operation provides reliable sub-ppt detection, excellent accuracy and long-term stability, supporting the semiconductor industry’s escalating purity demands.

References

• Agilent Applications of ICP-MS Measuring Inorganic Impurities in Semiconductor Manufacturing publication 5991-9495EN