Near-Infrared Spectroscopy: Quantitative analysis according to ASTM E1655

Importance of the topic

Near-infrared spectroscopy (NIRS) offers rapid, non-destructive quantitative analysis for a wide variety of materials, from polymers and fuels to food and pharmaceuticals. Its ability to operate offline, at-line, or inline makes it essential for real-time process monitoring, quality assurance, and raw-material verification in industrial and research laboratories.

Objectives and Study Overview

This white paper outlines the standardized workflow defined by ASTM E1655 for developing robust multivariate NIR calibration models. It guides users through sample selection, spectral measurement, chemometric model building, validation, and routine implementation, minimizing common pitfalls and clarifying critical steps.

Methodology and Instrumentation

The workflow comprises:

• Calibration Set Selection: Include representative samples spanning the full analyte range and all relevant sources of variation (e.g., moisture, particle size, geographical origin). For feasibility studies, use 30–50 samples; for final models, ensure at least 24–6×(factors+1) samples depending on latent variables.
• Spectral Data Acquisition: Use stable scanning-type NIR instruments in diffuse reflectance for solids and transmission for liquids. Fix all acquisition parameters (temperature, sampling accessories) across calibration, validation, and routine runs. Monitor instrument performance per ASTM E275 or pharmacopeia standards.
• Chemometric Model Development: Apply preprocessing (scatter correction, derivatives) and select spectral regions of interest. Employ PLS1 regression for single analytes, choosing the number of latent factors based on minimizing cross-validation error without overfitting. Detect spectral outliers by leverage (Mahalanobis distance) and residual outliers by standardized residual tests.
• Validation: Use an independent set covering ≥95% of each analyte range, with sample count ≥4×(factors+1) or 20. Compare standard error of prediction (SEP) to calibration errors, assess bias by t-test, and verify precision across the range with replicate measurements and χ2 testing.
• Routine Analysis and Maintenance: Implement the validated model in production, monitor performance with quality-control samples per ASTM D6122, and update models via slope/bias adjustments or full recalibration when new sample variation arises.

Main Results and Discussion

Following ASTM E1655 ensures consistent method performance: calibrations typically achieve comparable calibration and validation errors, robust bias control, and reliable precision across matrices. Outlier management improves model stability, and guided software workflows streamline development for end-users.

Benefits and Practical Applications

Key advantages include:

• Fast, simultaneous quantification of multiple analytes without consumables.
• Flexible deployment modes—offline QC, at-line checks, or inline process control.
• Reduced laboratory workload and higher sample throughput.
• Enhanced reproducibility through standardized protocols and software guidance.

Future Trends and Opportunities

Advances in sensor miniaturization, machine-learning algorithms for feature selection, and real-time adaptive calibration promise broader adoption. Integration with process analytical technology (PAT) frameworks and cloud-based data analytics will further enable predictive quality control and remote monitoring.

Conclusion

Standardized NIR method development according to ASTM E1655 provides a clear, structured approach to achieve reliable, high-throughput quantitative analysis across diverse applications. Adhering to best practices in sample selection, spectral acquisition, chemometrics, and validation ensures models that are accurate, robust, and easily maintained.

References

1. Encyclopedia of Analytical Chemistry, John Wiley and Sons, 2014.
2. Burns, D.A., Ciurczak, E.W., Handbook of Near-Infrared Analysis, CRC Press, 2007.
3. ASTM E1655-05(2012), Standard Practices for Infrared Multivariate Quantitative Analysis, ASTM International, 2012.
4. Wilhelm, R., “Know your type of standards,” ASTM, 2000.
5. ASTM D6342-12(2017)e1, Determining Hydroxyl Number of Polyols by NIR Spectroscopy, ASTM International, 2017.
6. ASTM D2699-16e1, Test Method for Research Octane Number, ASTM International, 2016.
7. ASTM E275-08(2013), Measuring Performance of UV-Vis Spectrophotometers, ASTM International, 2013.
8. United States Pharmacopeia 39, 2016.
9. European Pharmacopoeia, 9th Edition, 2016.
10. ASTM D6122-15, Validation of Multivariate Infrared Analyzer Systems, ASTM International, 2015.