Characterization of Trace Impurities in Silicon Wafers by High Sensitivity Reaction Cell ICP-MS

Significance of the Topic

The semiconductor industry requires ultra-pure silicon wafers to achieve high yields and reliable device performance. As device dimensions shrink and integration increases, ppt-level metal contaminants can cause defects, speed degradation, or yield loss. Accurate trace metal characterization in silicon matrices is therefore essential for process control and cost reduction.

Objectives and Study Overview

This study evaluated a high-sensitivity reaction cell ICP-MS method for quantifying trace metal impurities in silicon wafers, even with silicon concentrations up to 2000 ppm. Key aims included assessing the removal of silicon-based polyatomic interferences, meeting SEMI detection requirements for 36 elements, and validating method robustness for small sample volumes typical of semiconductor analysis.

Methodology

Sample preparation began with HF etching of silicon wafers to remove surface deposits, followed by complete digestion of bulk silicon in a sealed HF/HNO₃ mixture at 60 °C. Digests were diluted with ultrapure water to produce a 0.2 % Si solution. External calibration was performed using matrix-matched standards spiked at 0, 20, 60, and 100 ppt. No internal standards were used to minimize contamination risk and simplify sample preparation. Method validation included 50 ppt spike recovery experiments and a two-hour short-term stability study.

Used Instrumentation

• Agilent 7500cs ICP-MS equipped with an Octopole Reaction System (ORS)
• MicroFlow nebulizer (MFN-100) and Peltier-cooled PFA spray chamber
• 2 mm platinum injector torch with ORS ShieldTorch (STS)
• Reaction gases: hydrogen (reaction mode) and helium (collision mode)
• Operating conditions: 1600 W RF power, carrier gas 0.45 mL/min, He gas flow 5.0 mL/min

Main Results and Discussion

Detection limits (3σ) and background equivalent concentrations (BEC) for all SEMI-specified elements in the 2000 ppm Si matrix were at sub-ppt to tens of ppt levels. Notable findings:

• ORS cell modes effectively attenuated silicon-based interferences on Ti, Ni, Cu, Zn and plasma-based interferences on K, Ca, Fe, allowing on-mass detection.
• 50 ppt spike recoveries ranged from 79 % to 113 %, complying with SEMI’s 75–125 % criteria.
• Short-term stability over two hours showed %RSD typically below 5 %, confirming method robustness without internal standards.

Benefits and Practical Applications

• Rapid analysis of trace metals at ppt levels directly in high-silicon matrices in about five minutes per sample.
• Small sample volumes (<350 µL) compatible with surface metal extraction (SME) or liquid drop decomposition (LDD).
• Simplified workflow by eliminating internal standards and enabling single-run acquisition of all elements.
• Automated switching of reaction/collision modes improves throughput and reduces operator intervention.

Future Trends and Applications

Emerging reaction gases and improvements in ICP-MS cell design may further lower detection limits and expand interference removal capabilities. Integration with microfluidic sample preparation and in-line monitoring systems in semiconductor fabrication could enable real-time contamination control. The approach can be adapted for other wafer materials and advanced device architectures requiring ever-more stringent purity standards.

Conclusion

The Agilent 7500cs ORS ICP-MS method demonstrated reliable ppt-level detection of 36 elements in a 2000 ppm silicon matrix. Effective interference removal, compliance with SEMI specifications, robust stability without internal standards, and rapid analysis make this approach well-suited for semiconductor quality control and research applications.