Ultratrace Analysis of Solar (Photovoltaic) Grade Bulk Silicon by ICP-MS

Significance of the Topic

Solar power is poised to play a major role in future energy production, with projections indicating that photovoltaic systems could supply a substantial fraction of electricity demand by mid-century. The purity of silicon used in solar cells has a direct impact on device efficiency, since ultratrace impurities such as boron and phosphorus critically influence carrier lifetimes and electrical performance. Achieving and verifying impurity levels well below 10 ppb in solid silicon requires highly sensitive and robust analytical techniques.

Objectives and Study Overview

This work presents a quantitative method for detecting 31 elements in photovoltaic-grade bulk silicon down to low ppb and ppt levels using an Agilent 7500cs ICP-MS with an Octopole Reaction System (ORS). Key goals include:

• Developing sample preparation strategies that eliminate silicon matrix interferences while retaining volatile boron and enabling accurate phosphorus measurement.
• Validating recovery of all target analytes in complex HF/HNO₃ and H₂SO₄ digestions.
• Establishing detection limits in both solution and solid matrices.
• Demonstrating method performance across 13 real polysilicon samples.

Methodology and Instrumentation

Sample Preparation
Two distinct digestion protocols were optimized:

• General trace analysis: Silicon pieces were cleaned with HF, dissolved in HF/HNO₃, then spiked with H₂SO₄ to drive off residual Si by high-temperature evaporation. This step ensures complete removal of Si and eliminates Si-based polyatomic interferences on phosphorus.
• Boron-specific protocol: To prevent loss of volatile boron species, H₂SO₄ was omitted and evaporation was conducted at lower temperature. Matrix removal remains sufficient for boron quantification, but temperatures must be carefully controlled to avoid loss of Nb, Ta, and W near dryness.

Calibration
Calibration standards (0–1 ppb) were prepared in a mixed acid matrix matching the sample composition (0.34% HNO₃ and 0.33% H₂SO₄ by weight). A multielement spike enabled simultaneous quantification of all analytes.

Instrumentation
An Agilent 7500cs ICP-MS equipped with ORS and helium collision gas achieved interference reduction. Sample introduction employed a Micro Flow 100 nebulizer, an inert PFA spray chamber, sapphire-injected torch, and platinum sampling and skimmer cones. Operating parameters were optimized for three plasma conditions (cool, normal, ORS He mode) to target different analyte subsets. Integration times and gas flows were adjusted to balance sensitivity and interference removal.

Key Results and Discussion

Recovery Tests
Spike recoveries at 5 ppb in silicon digests ranged between 85% and 120% for all elements under both digestion protocols. Complete matrix removal with H₂SO₄ was confirmed by monitoring residual ²⁸Si in solution. Phosphorus determination at mass 47 (PO⁺) under cool plasma yielded background equivalent concentrations around 20 ppt, free from Ti interferences. Boron recoveries were excellent without H₂SO₄, provided evaporation was halted before full dryness.

Detection Limits
Three-sigma detection limits in solution spanned from tens of ppt for most metals to ~20 ppt for boron, translating into sub-ppb to low-ppb limits in the solid sample (0.3 g to 15 mL dilution). Phosphorus DL in the solid was approximately 850 ppt, and boron DL near 100 ppt.

Quantitative Analysis of Polysilicon Samples
Analysis of 13 different polysilicon blocks revealed:

• Boron levels between 6 and 100 ppb.
• Phosphorus ranging from 10 to 1 000 ppb across samples.
• High inhomogeneity for Fe, Mn, and Ni, with Fe up to 1 700 ppb in some blocks.
• Most other trace elements present at low-ppb or sub-ppb levels.

These results underscore the need for stringent quality control during silicon production.

Benefits and Practical Applications

This method provides:

• Rapid and reliable quantification of a broad suite of elements critical to solar cell performance.
• Detection capabilities meeting industry requirements for ultratrace impurity control.
• Reproducible sample preparation workflows suitable for routine QC labs.

Implementation in semiconductor and photovoltaic manufacturing enables tighter process monitoring and improved device yields.

Future Trends and Applications

Advancements likely include:

• Integration of collision/reaction gases beyond helium to further suppress interferences on challenging isotopes.
• Automated sample handling with robotic platforms to increase throughput and minimize contamination.
• Extension of ultratrace methods to next-generation materials such as perovskites and thin-film semiconductors.
• Data integration with machine learning models to predict performance impact of specific impurity profiles.

Conclusion

A robust ICP-MS-based protocol for ultratrace analysis of photovoltaic-grade silicon has been established. The dual digestion strategies ensure accurate quantification of volatile boron and interference-prone phosphorus, while ORS and optimized plasma conditions deliver ppt-level sensitivity for a wide element range. The approach supports stringent impurity specifications essential for high-efficiency solar cells.

Reference

• Junichi Takahashi. Ultratrace Analysis of Solar (Photovoltaic) Grade Bulk Silicon by ICP-MS. Agilent Technologies Application Note, October 2008.
• Scientific American. “Solar Grand Plan,” January 2008.