Rapid analysis of multiple components in tobacco using the Antaris II FT-NIR Analyzer

Rapid analysis of multiple components in tobacco using the Antaris II FT-NIR Analyzer — Summary

Significance of the topic

FT-NIR spectroscopy offers a rapid, low-cost, and non-destructive alternative to traditional wet-chemical assays for routine quality control in the tobacco and cigarette industries. Because tobacco materials are highly heterogeneous by type, origin and leaf position, manufacturers require fast multi-constituent feedback from raw materials and process intermediates. FT-NIR enables simultaneous quantification of many analytes from a single spectrum with minimal sample preparation, reducing analysis time, reagent use and operator training compared with manual chemical methods or reagent-based automated systems such as FIA.

Objectives and overview of the study

The application note evaluated FT-NIR (Thermo Scientific Antaris II) for quantitative analysis of 16 tobacco components. The goals were to develop and validate calibration models for each component using representative, naturally occurring tobacco samples from multiple types and producing regions in China, and to compare FT-NIR performance (accuracy, robustness, speed, and operational advantages) against conventional wet-chemical methods.

Methodology

• Sample set: ~800 tobacco leaf samples (milled to powder) representing different tobacco types, origins and processing states (including aged/fermented leaves). Calibration sets per analyte ranged from several hundred to ~785 samples; validation sets were ~39–56 samples per analyte.
• Spectral acquisition: Diffuse reflectance spectra collected on the Antaris II FT-NIR with an integrating-sphere solid sample module over 10,000–3,800 cm-1 at 8 cm-1 resolution. Seventy scans per sample with rotation of the sample cup; typical acquisition time ~1 minute per sample.
• Reference data: Wet-chemical analyses performed by the Chinese Tobacco Research Institute and factory analytical centers according to national standard methods; these reference values were used for calibration.
• Chemometrics: Spectra pretreated routinely with Multiplicative Scatter Correction (MSC), first derivative and Norris smoothing. Partial Least Squares (PLS) regression (TQ Analyst software) was used for all calibrations. Outlier screening employed statistical tests (Dixon or Chauvenet). Model selection used RMSEC, RMSECV and RMSEP and inspection of residual distributions and RMSECV vs PLS factor curves to avoid under- or over-fitting.

Instrumentation

• Thermo Scientific Antaris II FT-NIR analyzer with integrating-sphere solid sampling module.
• TQ Analyst chemometric software (for PLS model development and validation).

Main results and discussion

• Sixteen analytes were modeled successfully with good predictive performance overall. Correlation coefficients for calibration models ranged approximately from 0.91 to 0.99, indicating strong relationships between NIR predictions and reference values.
• Representative RMSEP values (examples): nicotine 0.170 (% absolute), total sugars 1.17, reductive sugars 0.92, total nitrogen 0.0882, potassium 0.186, chlorine 0.0529, total volatile acids 0.00530, total volatile bases 0.0205, sulfate 0.159, starch 0.56, cellulose 0.00855, polyphenols 2.7 (largest RMSEP in the set), total petroleum ether extracts 0.00420, neutral petroleum ether extracts 0.00361, ash 0.945, water-soluble ash bases 0.226.
• PLS model complexity varied by analyte (PLS factors typically 9–24). Spectral regions used differed per analyte to capture relevant absorption features; common preprocessing (MSC, first derivative) improved calibration stability by reducing scatter and baseline effects.
• Models showed robustness given large, diverse calibration sets drawn from real-world tobacco variability. Outlier detection and validation helped ensure predictive reliability. For some analytes (e.g., polyphenols), prediction errors were larger, reflecting either larger reference-method variability or weaker direct NIR absorbance signatures.
• Compared with wet chemistry and FIA, FT-NIR provided drastic reductions in analysis time (<1 minute per sample for multiple components), eliminated reagent consumption and hazardous waste, reduced sample-preparation errors, and lowered operating costs and training needs.

Benefits and practical applications

• Rapid at-line or laboratory QC screening for raw tobacco, blends and intermediates to support blending decisions, process control and supplier acceptance.
• Simultaneous multi-component results from a single scan reduce per-sample cost and turnaround time versus sequential wet-chemical assays.
• Non-destructive measurement and minimal sample handling improve reproducibility and reduce operator-induced variability.
• Instrument transferability using FT-IR/FT-NIR technology facilitates multi-site deployment and centralized model management.

Future trends and possibilities for use

• Expansion of calibration libraries with larger, geographically and temporally diverse sample sets to improve robustness and extend applicability to new tobacco types and derived products.
• Implementation of at-line or in-line FT-NIR sensors in processing lines for real-time monitoring and automated process control.
• Advanced chemometrics and machine-learning approaches (nonlinear methods, domain adaptation) to improve prediction for constituents with weak NIR signals or complex matrices.
• Integration with digital quality-management systems for trend analysis, supplier performance tracking and predictive maintenance of processing operations.
• Miniaturized or portable NIR instruments and hyperspectral imaging as complementary tools for spatially resolved quality mapping of leaves and blends.

Conclusions

The study demonstrates that FT-NIR spectroscopy using the Antaris II platform can reliably quantify a wide suite of tobacco components with rapid throughput, low operating cost and minimal sample preparation. Well-developed PLS models built from large, representative calibration sets produced prediction errors acceptable for routine QC for most analytes. FT-NIR is a practical and cost-effective replacement or screening tool for many traditional wet-chemical assays in tobacco laboratories, enabling faster decision-making and reduced reagent waste.

References

• McClure WF. Spectral characteristics of tobacco in the near infrared region from 0.6 to 2.6 microns. Tobacco Science. 1968;12:232–235.