Analytical data transfer between a Fourier transform and a dispersive NIR instrument

Importance of the Topic

NIR spectroscopy is a key tool in quality control of lubricating oils and many other products. Maintaining consistent calibration models when replacing instrumentation is critical to avoid costly and time-consuming method redevelopments. Demonstrating reliable transfer of calibration data between different NIR technologies enhances laboratory flexibility and reduces downtime.

Objectives and Study Overview

The study aimed to transfer existing calibration models for acid number, kinematic viscosity and moisture content from an FT-NIR instrument to a dispersive Metrohm NIRS XDS RapidLiquid Analyzer. The goal was to preserve analytical performance without remeasuring the full sample set or rebuilding methods from scratch.

Methodology

The data transfer involved four main steps:

• Conversion of FT-NIR spectra from wavenumber to wavelength scale and from transmittance to absorbance.
• Interpolation of converted spectra to match the resolution and wavelength grid of the dispersive analyzer.
• Selection of a transfer set of 30 samples covering the calibration range and computation of a transfer model using piecewise direct standardization (PDS) over the 1120–2100 nm region.
• Import of transferred data into Vision Air Complete software, development of quantification models and adjustment via simple slope/bias correction, followed by validation on an independent sample set.

Instrumentation

The source instrument was an FT-NIR analyzer operating from 800 to 2500 nm (12 500–4000 cm⁻¹) with 16 cm⁻¹ resolution and 32 scans per spectrum. The target system was the Metrohm NIRS XDS RapidLiquid Analyzer, measuring samples in 8 mm disposable glass vials in transmission mode over 400–2500 nm at 40 °C. Data acquisition and chemometric modelling employed Vision Air 2.0 Complete.

Main Results and Discussion

Comparison of transferred FT-NIR spectra and spectra acquired on the dispersive analyzer showed minimal differences, confirming successful alignment. Validation yielded strong correlations between reference and NIR-predicted values:

• Acid number: SECV 0.80 mg KOH/g, SEP 0.84 mg KOH/g
• Moisture: SECV 0.012 %, SEP 0.018 %
• Viscosity at 40 °C: SECV 2.1 cSt, SEP 2.9 cSt

Regression plots demonstrated high accuracy and precision across all quality parameters.

Benefits and Practical Applications

The transfer approach eliminates the need for complete remeasurement of calibration samples and full method redevelopment, saving both time and resources. Laboratories can switch between instrument types or suppliers with minimal disruption and maintain continuous quality control operations.

Future Trends and Opportunities

Advances in calibration transfer algorithms, integration of cloud-based chemometric services and expansion to real-time, online process monitoring will further streamline instrument changes. Application of this workflow to a broader range of matrices and the adoption of more automated transfer protocols are promising directions.

Conclusion

This work demonstrates that FT-NIR calibration models can be efficiently and accurately transferred to a dispersive NIR platform within hours. The procedure preserves analytical performance, simplifies instrument replacement and supports robust quality control without extensive redevelopment.