Discrimination of Soft Drinks using a Chemical Sensor and Principal Component Analysis

Significance of the Topic

Understanding volatile fingerprinting of soft drinks is crucial for quality control, authenticity verification, and product consistency in the beverage industry.
Combining chemical sensors with multivariate analysis streamlines complex data interpretation and accelerates decision making.

Objectives and Study Overview

This study evaluates a chemical sensor based on quadrupole mass spectrometry coupled with principal component analysis (PCA) to discriminate cola beverages from different vendors and packaging.

Methodology and Experimental Design

• Sample Selection: Three cola brands (A, B, C) with brand C in both plastic bottles and aluminum cans; eight replicates per sample type.
• Headspace Sampling: 5 mL aliquots equilibrated at 80 °C for 20 min in 10 mL vials; direct injection into the mass selective detector without chromatographic separation; mass range m/z 46–150 scanned over ~0.75 min.
• Comparative Analysis: Brand C samples also analyzed by GC/MS after 10× dilution and one-hour stir bar sorptive extraction (Twister) with thermal desorption.

Instrumentation Used

• Gerstel ChemSensor 4440 integrating a headspace unit with Agilent 5973 mass selective detector and Pirouette PCA software.
• Agilent 6890 gas chromatograph with 5973 MSD and Gerstel Thermal Desorption Autosampler.
• Stir bar sorptive extraction (Twister, Gerstel) for volatile enrichment.

Key Results and Discussion

PCA captured over 90 % of data variance within the first three principal components, producing distinct clusters by brand and packaging.
Brand C in plastic bottles separated clearly along PC1 from all canned samples, while PC2 differentiated brands A and B.
Hierarchical cluster analysis using Euclidean distance confirmed four clusters corresponding to sample types.
GC/MS–Twister results revealed higher abundances of terpenes and p-cymene in canned samples, indicating possible light-sensitive degradation in plastic bottles.

Benefits and Practical Applications

• Rapid, non-targeted profiling for authenticity screening and quality monitoring without chromatographic steps.
• Efficient differentiation of similar products in production and regulatory environments.

Future Trends and Potential Applications

Advancements may include integration of sensor arrays with machine learning for enhanced specificity and predictive modeling.
Development of portable electronic nose platforms for in-line beverage monitoring and field inspections.

Conclusion

The combination of headspace chemical sensing and PCA offers a fast, reliable approach for discriminating soft drink samples by brand and packaging.
Complementary GC/MS with stir bar extraction identifies key volatiles responsible for observed differences, improving method interpretability.