The Analysis of Beer Components Using FT-NIR Spectroscopy

Importance of the topic

Beer is a chemically complex, widely consumed fermented beverage whose quality depends on multiple compositional and physical attributes (alcohol content, color, refractive index, density, bitterness, etc.). Routine monitoring of these parameters is essential for consistent product quality and process control in breweries. Conventional analytical workflows typically require multiple separate instruments and time‑consuming sample preparation. This application note demonstrates how Fourier Transform Near‑Infrared (FT‑NIR) spectroscopy, combined with chemometric modeling, can provide a fast, reagent‑free, simultaneous analysis of several key beer parameters, enabling more efficient QA/QC and process monitoring.

Objectives and study overview

The study aimed to evaluate the feasibility, accuracy and robustness of FT‑NIR spectroscopy for quantitative determination of multiple beer attributes using a single spectrum per sample. Specific targets were alcohol content, color (EBC scale), refractive index and specific density. A calibration set of 27 beer standards was used to develop Partial Least Squares (PLS) models and performance was assessed by cross‑validation and comparison with conventional reference methods.

Methodology and instrumentation used

Key experimental details:

• Instrument: Thermo Scientific Antaris II FT‑NIR Method Development Sampling (MDS) system.
• Sampling accessory: SabIR fiber optic probe with transflectance module (direct, no sample preparation).
• Spectral range: 10,000–4,000 cm⁻¹.
• Acquisition parameters: 8 cm⁻¹ resolution, 32 co‑averaged scans, Spectralon reference background; single‑sample measurement time ≈ 25 s.
• Software and chemometrics: Thermo Scientific RESULT for acquisition and TQ Analyst for calibration; spectra mean‑centered and multiplicative scatter correction (MSC) applied.

Preprocessing and spectral regions used per analyte (calibration strategy):

• Alcohol: 5,500–4,000 cm⁻¹, first derivative applied.
• Color (EBC): 9,900–4,100 cm⁻¹, no derivative.
• Refractive index: 7,000–4,000 cm⁻¹, first derivative plus Norris derivative smoothing (segment 5, gap 2).
• Specific density: 7,162–4,099 cm⁻¹, second derivative plus Norris derivative smoothing (segment 3, gap 2).

Main results and discussion

PLS calibration models demonstrated excellent correlation between FT‑NIR predictions and conventional reference measurements. Key performance metrics (calibration summary):

• Alcohol — correlation coefficient 0.99886; RMSEC 0.0430; RMSEP 0.0247; RMSECV 0.383.
• Color — correlation coefficient 0.99983; RMSEC 0.0315; RMSEP 0.0793; RMSECV 0.187.
• Refractive index — correlation coefficient 0.99965; RMSEC 0.132; RMSEP 0.172; RMSECV 0.367.
• Specific density — correlation coefficient 0.99619; RMSEC 0.188 × 10⁻³; RMSEP 0.306 × 10⁻³; RMSECV 0.341 × 10⁻³.

PRESS (Predicted Residual Error Sum of Squares) plots from cross‑validation showed the expected decrease to a minimum and subsequent stabilization, indicating well‑behaved and robust chemometric models. Overall, FT‑NIR provided low prediction errors and high linearity versus reference methods for all measured attributes. The transflectance probe enabled direct measurement of liquid beer without dilution or other sample treatment, delivering results in roughly 25 seconds per sample.

Benefits and practical applications

Advantages demonstrated in this work include:

• Rapid, simultaneous quantification of multiple beer quality attributes from a single spectrum, reducing analysis time versus traditional multi‑instrument workflows.
• No sample preparation or reagents, lowering per‑sample cost and operator time.
• High accuracy and robustness of PLS models when built from representative calibration sets.
• Potential for at‑line or online implementation (probe‑based measurements) to support process analytical technology (PAT) and real‑time QA/QC.

Practical uses include routine brewery QC, process monitoring during fermentation and blending, and raw material or finished product screening to ensure consistency in alcohol levels, color and density.

Future trends and potential applications

Emerging and recommended directions for expanding this approach:

• Scale‑up of calibration libraries to cover broader recipe and seasonal ingredient variability to improve model transferability between breweries and product types.
• Integration of inline/online NIR probes for continuous process monitoring and closed‑loop control of fermentation and blending operations.
• Application of advanced machine learning or hybrid chemometric approaches to improve robustness against matrix effects and to extend predictions to additional analytes (e.g., sugars, organic acids, higher alcohols).
• Standardization and validation protocols to facilitate regulatory acceptance and routine deployment in QA environments.
• Development of calibration transfer strategies and multivariate maintenance plans to minimize drift and reduce the need for frequent recalibration.

Conclusion

This application demonstrates that FT‑NIR spectroscopy using a transflectance fiber probe and PLS chemometrics can accurately and rapidly quantify multiple important beer parameters without sample preparation. The Antaris II FT‑NIR system produced high‑quality calibrations with low prediction errors and robust cross‑validation behavior, confirming FT‑NIR as a viable alternative to many conventional analytical methods for brewery QA/QC. Implementation can deliver faster turnaround, cost savings and support for at‑line process control.

Reference

Thermo Fisher Scientific. Application Note AN51892 (0321). The Analysis of Beer Components Using FT‑NIR Spectroscopy. Thermo Scientific Antaris II MDS system; SabIR fiber optic probe; TQ Analyst software. Source language of original document: EN