The Analysis of Beer Components Using FT-NIR Spectroscopy

Importance of the topic

Beer is a chemically complex, globally produced beverage whose quality and consistency depend on multiple interrelated parameters (ethanol content, color, density, refractive index and more). Traditional analyses use multiple targeted techniques (distillation, densimetry, refractometry, photometry, chromatography), which are time-consuming, require sample preparation and specialized instruments. Rapid, multi-component screening methods are therefore valuable in brewery quality control and process monitoring. This study demonstrates the application of Fourier-transform near-infrared (FT-NIR) spectroscopy combined with chemometrics as an efficient alternative for simultaneous quantification of several routine beer attributes.

Objectives and study overview

The primary aim was to evaluate whether a single FT-NIR spectrum, acquired via a transflectance fiber-optic probe, can be used to predict multiple beer parameters reliably and rapidly. Specifically, the work developed and validated PLS regression models to predict alcohol content, color (EBC scale), refractive index and specific density from 27 beer standard samples, comparing FT-NIR predictions with conventional reference methods.

Methodology

• Samples: 27 beer standard samples measured without any sample preparation.
• Spectral acquisition: FT-NIR spectra were collected from 10,000 to 4,000 cm-1 at 8 cm-1 resolution, co-averaging 32 scans and using a Spectralon reference background. Single-sample acquisition time for quantitative prediction was about 25 seconds.
• Sampling accessory: Measurements performed with a SabIR fiber-optic probe combined with a transflectance accessory to interrogate liquid beer directly.
• Chemometrics: Spectra were mean-centered and multiplicative scatter correction (MSC) was applied. Partial Least Squares (PLS) regression models were built in TQ Analyst. Different spectral subregions and preprocessing were chosen per analyte (first or second derivatives, and Norris derivative smoothing where appropriate).
• Validation: Models were evaluated using cross-validation; PRESS diagnostics and RMSE metrics (RMSEC, RMSECV, RMSEP) were used to assess fit and predictive performance.

Instrumentation

• FT-NIR spectrometer: Thermo Scientific Antaris FT-NIR Method Development Sampling (MDS) System.
• Sampling probe: Thermo Scientific SabIR fiber-optic probe with transflectance accessory.
• Reference material: Spectralon reference for background scans.
• Software: RESULT for spectral acquisition and Thermo Scientific TQ Analyst for chemometric model development.
• Key acquisition parameters: 10,000–4,000 cm-1 spectral range; 8 cm-1 resolution; 32 co-averaged scans; MSC pathlength correction; derivative and Norris smoothing applied selectively in PLS models.

Main results and discussion

The PLS models achieved excellent agreement with conventional reference methods across all four target parameters. Correlation coefficients were very high (≥0.996 for specific density and >0.998 for alcohol, color and refractive index), indicating strong linear relationships between FT-NIR predictions and reference values. Prediction errors were low: RMSEP values were small relative to typical analytical tolerances (for example, alcohol RMSEP on the order of 0.02–0.08 units depending on the parameter), and RMSECV and RMSEC indicated robust calibrations. PRESS plots exhibited the expected decrease to a minimum and stabilization, supporting model validity and appropriate latent variable selection.
Using targeted spectral subregions and derivative pretreatments improved sensitivity to the analytes of interest and reduced interferences from overlapping absorptions and baseline effects. The transflectance probe facilitated direct measurement of undiluted beer, eliminating sample preparation steps that traditionally add time and potential variability.

Benefits and practical applications

• Multi-analyte capability: A single FT-NIR spectrum provides rapid, simultaneous quantification of multiple routine beer quality parameters, replacing several separate assays.
• Speed and throughput: Approximately 25 seconds per sample enables high-throughput screening suitable for in-process monitoring and quality control.
• Minimal sample preparation: Direct transflectance measurement reduces labor, consumables and risk of sample alteration.
• Cost and efficiency: Consolidation of analyses into one instrument lowers operational costs and accelerates decision-making in production environments.
• Robustness: Chemometric models validated by cross-validation and PRESS diagnostics demonstrated reliable predictive performance for the tested standards.

Future trends and potential uses

• Expansion of calibration libraries: Increasing the diversity and number of calibration samples (different beer styles, seasonal batches) will improve model generalizability for routine brewery deployment.
• Process analytical technology (PAT) integration: Inline or at-line FT-NIR probes could enable continuous monitoring of fermentation and blending steps, supporting real-time process control.
• Broader analyte panels: Extension to sugars, organic acids, bitterness components and off-flavor markers through targeted spectral and chemometric development.
• Transfer and standardization: Development of model transfer protocols and standardized spectral preprocessing to allow deployment across multiple sites and instruments.
• Machine learning enhancements: Combining PLS with newer machine learning approaches may further improve prediction accuracy and robustness against matrix variability.

Conclusion

FT-NIR spectroscopy with a transflectance fiber-optic probe and PLS chemometrics provides a fast, accurate and robust method for simultaneous determination of key beer quality parameters. The approach reduces analysis time and sample handling, shows excellent correlation with conventional reference methods, and is well suited for brewery quality control and process monitoring. Adoption of FT-NIR can streamline workflows and improve operational efficiency while maintaining analytical confidence.

References

• Narimoto KM, de Oliveira AM. The Analysis of Beer Components Using FT-NIR Spectroscopy. Application Note 51892. Thermo Fisher Scientific; 2010. AN51892_E 03/10M.
• Thermo Scientific Antaris FT-NIR Method Development Sampling System documentation and SabIR fiber-optic probe specifications (Thermo Fisher Scientific).