Determination of transition metals at ppt levels in high-purity water and SC2 (D-clean) baths

Importance of the Topic

Metal residues in semiconductor cleaning agents and ultrapure rinse waters can compromise device performance and yield. Monitoring transition metals at picogram-per-liter levels ensures optimal removal of contaminants during wafer processing. Robust analytical methods capable of detecting trace metals at ppt concentrations are essential for quality assurance in semiconductor manufacturing.

Study Objectives and Overview

This work presents an ion chromatography approach to quantify sub-nanogram-per-liter concentrations of transition metals in high-purity water and SC2 (D-clean) baths. The method employs a Dionex IonPac CS5A column for metal separation using strong and moderate chelating eluents, followed by post-column derivatization with PAR (4-(2-pyridylazo)resorcinol) and visible detection at 530 nm. Preconcentration on a trace cation concentrator column enhances sensitivity for ultra-trace analysis.

Methodology and Instrumentation Used

Sample preparation includes acidification to 2 mM HCl to stabilize metal ions and prevent hydroxide precipitation. Trace levels (< 2 µg/L) are concentrated on a sulfonated resin concentrator at 2 mL/min for 5–15 minutes, depending on sample volume.

• Separation column: IonPac CS5A (2 × 250 mm) with CG5A guard (2 × 50 mm)
• Concentrator: IonPac TCC-2 (3 × 35 mm)
• Eluent: 7 mM PDCA, 66 mM KOH, 5.6 mM K₂SO₄, 74 mM formic acid, 0.3 mL/min
• Post-column reagent: 0.06 g PAR in Pari-diluent, 0.15 mL/min
• Detection: UV/Vis at 530 nm
• Flow path: PEEK metal-free tubing, microbore (0.125 mm i.d.) to minimize dead volume

Alternative configuration: Thermo Scientific ICS-5000+ HPIC system.

Main Results and Discussion

Preconcentration of a 1 µg/L standard from 30 mL yielded baseline-resolved peaks for seven transition metals with sharp, reproducible retention times. Analysis of high-purity water blanks (10 and 30 mL) delivered iron levels of approximately 45 ng/L, with area and retention time RSDs below 2%. Examination of SC2 bath samples (30 mL) revealed Fe³⁺, Cu²⁺, and Zn²⁺ concentrations of 80, 75, and 106 ng/L, respectively. Higher iron recovery in the bath compared to water indicates metal contributions from bath reagents.

Benefits and Practical Applications

This method enables reliable detection and quantification of transition metals at ppt levels, supporting stringent contamination control in semiconductor cleaning processes. The combination of microbore columns, strong chelation, and post-column derivatization provides high sensitivity and minimal background. Rapid preconcentration and short run times improve laboratory throughput for routine QA/QC.

Future Trends and Potential Applications

Advances may include integration with automated sample handling to reduce contamination risk, coupling with high-resolution detectors for lower detection limits, and exploration of alternative complexing agents to expand metal coverage. Miniaturized flow cells and improved column chemistries could further enhance sensitivity, enabling real-time monitoring of process waters and cleaning baths.

Conclusion

The described ion chromatography method, featuring PDCA-based separation, PAR derivatization, and preconcentration on a sulfonated concentrator, delivers ppt-level analysis of transition metals in ultra-pure water and SC2 baths. The approach offers reproducible performance, low detection limits, and compatibility with semiconductor quality control requirements.

Reference

1. Moreau WM. Semiconductor Lithography: Principles, Practices, and Materials. Plenum Press; 1988.
2. SEMI. Suggested Guidelines for Pure Water used in Semiconductor Processing. Doc 2796; 1998.
3. Dionex. 2-mm Transition Metal System with Postcolumn Delivery Installation and Troubleshooting Manual, P/N 031355.