Characterization of Surface Metal Contamination on Silicon Wafers Using Surface Metal Extraction Inductively Coupled Plasma Mass Spectrometry (SME- ICP-MS)

Importance of the Topic

As semiconductor features continue to shrink and integration scales increase, even ultra-trace metal contaminants on silicon wafer surfaces can severely impact device performance and yield. Rapid, accurate surface analysis techniques are therefore essential to identify and control sources of metallic impurities in real time during wafer fabrication.

Study Objectives and Overview

This work evaluates the combination of Surface Metal Extraction (SME) with Inductively Coupled Plasma Mass Spectrometry (ICP-MS) for characterizing trace metals on silicon wafers. The goals were to achieve detection limits well below projected 2009 industry specifications, enable complete wafer preparation and analysis in under 20 minutes, and demonstrate suitability for inline production monitoring.

Methodology and Instrumentation

The SME procedure dissolves the native or thermal oxide layer on the wafer with hydrofluoric acid, then collects released metals by scanning a micro-volume (250 µL) droplet across the surface. The droplet is recovered and introduced directly into an Agilent 7500s ICP-MS equipped with:

• MicroFlow nebulizer (20 µL/min uptake) and Peltier-cooled spray chamber
• ShieldTorch System (STS) to eliminate argon-based polyatomic interferences
• Soft-extraction mode with twin extraction lens design for enhanced signal-to-background ratio
• 27.12 MHz high-temperature plasma generator and Omega II ion lens for flat mass response

Calibration standards ranged from 0 to 500 ppt in 5% HF/6% H₂O₂. A synthetic SME matrix containing 0.59 ppm Si, prepared from high-purity silicon, was used to assess spike recoveries and matrix effects.

Main Results and Discussion

Detection limits for a suite of semiconductor-relevant elements remained in the sub-ppt to single-ppt range even in the high silicon matrix, comfortably meeting or exceeding 2009 roadmap requirements for 450 mm wafers. Key findings include:

• 10–100-fold improvement in signal-to-background ratios using soft-extraction mode
• Virtually complete removal of polyatomic interferences via STS without sensitivity loss
• No significant difference in detection limits between calibration matrix and synthetic SME matrix
• Spike recoveries of 91–115% across all tested elements, indicating negligible plasma suppression or transport losses

Benefits and Practical Applications

This SME-ICP-MS approach offers:

• Rapid turnaround (under 20 minutes) for wafer preparation and analysis
• Multi-element capability (up to 40 elements) in a single micro-droplet
• Ultra-trace detection limits (sub-ppt) for real-time production monitoring
• Simplified workflow without internal standards, matrix matching, or complex collision cell technologies

Future Trends and Potential Applications

Further developments may include integration with automated wafer handling systems, adaptation to next-generation 300 mm and 450 mm wafer formats, and extension to new substrate materials or advanced dielectric layers. Improvements in nebulizer throughput and plasma source design could push detection limits even lower, supporting evolving industry requirements.

Conclusion

The SME-ICP-MS technique on the Agilent 7500s platform delivers rapid, highly sensitive, and interference-free analysis of surface metal contaminants on silicon wafers. Its performance surpasses projected semiconductor industry needs for 2009 and provides a practical solution for inline quality control and yield optimization.

References

• Sakata K, Yamada N, Sugiyama N. Spectrochimica Acta Part B. 2001;56:1249.
• Tye CT, Sakata K. Agilent Technologies ICP-MS Journal. 2000;8:7.
• Tye CT, Gutierrez A. Agilent Technologies ICP-MS Journal. 2000;7(2).
• Lian H, Nicoley B, Howard A, Radle M. Semiconductor International. July 2001.