Trace Impurity Elemental Analysis of Magnet-Grade Neodymium–Iron– Boron by ICP-OES

Trace impurity elemental analysis of magnet‑grade NdFeB by ICP‑OES using FACT spectral deconvolution

Significance of the topic

Rare‑earth permanent magnets based on Nd2Fe14B are critical materials for electric vehicles, wind turbines, electronics and aerospace. Small amounts of substituting rare earth elements (REEs) and other impurities can disrupt the magnetic phase, degrade performance, and reduce manufacturing yield. Reliable quantification of REE impurities at low concentrations is therefore essential for quality control, materials optimization and process control in high‑value magnet production.

Objectives and overview of the study

• Demonstrate a practical ICP‑OES method for accurate quantification of 13 rare‑earth impurity elements in NdFeB magnet material.
• Show how Agilent’s FACT (Fast Automated Curve‑fitting Technique) real‑time spectral deconvolution removes severe REE‑to‑REE and matrix spectral interferences that commonly compromise ICP‑OES results.
• Evaluate method performance via calibration linearity, detection limits, spike recovery and measurement of a commercial NdFeB sample digest.

Methodology and sample preparation

• Sample digestion: ~0.1 g NdFeB dissolved in 5 mL freshly prepared aqua regia (HNO3:HCl = 1:3 v/v) heated to 150 °C until complete dissolution, then diluted to 50 mL with ultrapure water.
• Standards and spikes: multi‑element REE standard (Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Sm, Tb, Th, Tm, Y, Yb) used for calibration; single‑element Fe, Nd and Pr standards used to build FACT interferent models and validate deconvolution.
• Blanks and detection limit calculations: 10% aqua regia blanks measured 11 times. IDL defined as 3×SD of blanks; MDL adjusted by sample dilution factor.
• Spike recovery tests: known spike levels added to digests prior to analysis to verify accuracy.

Used instrumentation

• Agilent 5800 ICP‑OES controlled by Agilent ICP Expert software (v7.7.3) with IntelliQuant and FACT capabilities.
• Sample introduction: SeaSpray concentric glass nebulizer, double‑pass glass cyclonic spray chamber, fully demountable torch with 1.8 mm i.d. injector.
• Operational parameters (representative): axial viewing mode; plasma gas flow 14 L/min; auxiliary gas 1.0 L/min; nebulizer gas 0.7 L/min; RF power 1300 W; read time 10 s; 3 replicates; uptake delay 13 s; stabilization 12 s; pump speed 12 rpm.
• Consumables and standards: peristaltic pump tubing, multi‑element 10 μg/mL REE standard in 5% HNO3, high‑concentration Fe, Nd, Pr single standards.

Selection of analytical wavelengths and spectral correction strategy

• IntelliQuant screening was used to identify major matrix elements (Fe, Nd, Dy, Pr, B) and recommend optimal analytical lines by a star‑rating system to avoid obvious matrix interferences.
• FACT spectral deconvolution models were built from blanks and single‑element standards to mathematically separate closely spaced emission lines (example: deconvolution of La 408.671 nm from overlapping Nd 408.680 nm).
• FACT background correction was applied to 11 REEs; fitted background correction (FBC) was used for a small number of lines (e.g., Lu 261.541 nm and Yb 328.937 nm in this study).

Main results and discussion

• Calibration and linearity: All 13 REE impurity elements showed excellent linearity across the chosen calibration range (0.05–0.5 mg/L) with correlation coefficients >0.999.
• Sensitivity and detection limits: Instrument detection limits (IDLs) for the set of REEs were all below approximately 5 μg/L (ppb); method detection limits (MDLs) adjusted for the sample dilution were in the sub‑to‑low mg/kg range for the digested NdFeB matrix (e.g., MDLs ranged from ~0.10 to 2.40 mg/kg depending on element).
• Accuracy: Spike recoveries for the 13 REEs were between 90% and 111%, demonstrating effective correction of spectral overlap and good method accuracy in the complex NdFeB matrix.
• Sample results: Quantified concentrations in the tested NdFeB digest (mg/kg) included notable levels such as Ce ~157 mg/kg, Dy ~180 mg/kg, Gd ~25 mg/kg, Sm ~21 mg/kg, La ~14 mg/kg, Tb ~4.6 mg/kg; several elements (Eu, Tm) were below the MDL.
• Practical performance: The FACT approach successfully resolved near‑coincident REE emission lines and allowed robust quantification without requiring alternative techniques (e.g., ICP‑MS) or extensive wavelength reselection.

Benefits and practical applications of the method

• Improved accuracy for REE impurity analysis in NdFeB by real‑time spectral deconvolution reduces false positives/negatives caused by overlapping REE and matrix lines.
• High sensitivity (IDLs <5 ppb) and good method precision make the approach suitable for QC laboratories monitoring trace impurities that affect magnetic performance and production yield.
• IntelliQuant automation accelerates method development and line selection, lowering analyst time and potential for human error.
• The workflow is accessible to laboratories with ICP‑OES instrumentation and does not require switching to more costly or slower techniques when spectral interferences are managed by FACT.

Future trends and possibilities for application

• Broader adoption of real‑time spectral deconvolution and intelligent software (like IntelliQuant) will streamline ICP‑OES workflows for other complex matrices where multi‑element spectral overlap is an issue (e.g., battery materials, critical metals recycling, geological samples).
• Integration with automated sample preparation and QA/QC routines can enable higher throughput quality control for magnet manufacturers and recyclers.
• Continued improvements in modeling algorithms and reference library expansion will further reduce limits of detection and expand the number of resolvable overlaps, narrowing the performance gap between ICP‑OES and ICP‑MS for certain applications.

Conclusion

The Agilent 5800 ICP‑OES equipped with FACT spectral deconvolution and IntelliQuant screening provides a robust, sensitive and accurate method for quantifying 13 rare‑earth impurity elements in NdFeB magnet material. Calibration linearity (R>0.999), instrument detection limits below ~5 μg/L and spike recoveries of 90–111% demonstrate that FACT effectively mitigates complicated REE spectral interferences, enabling effective QC and compositional analysis of magnet‑grade materials.

References

1. Part II: XB/T 617 – Chemical Analysis Methods for Neodymium–Iron–Boron Alloys, Chinese standard (Part II).
2. Real‑Time Spectral Correction of Complex Samples Using FACT Spectral Deconvolution Software, Agilent application note 5991‑4837EN.
3. Agilent IntelliQuant Software: For greater sample insight and simplified method development, Agilent publication 5994‑1516EN.