Chromatography-Based Analysis for Smarter Plant-Based Milk Sensors

As plant-based milks become more popular, new challenges arise with food safety and how these drinks perform when you make them, especially when it comes to preventing allergic reactions from cross-contamination. Normally, quality checks only happen early and in the factory. Unfortunately, this doesn't help protect individual consumers or ensure the drink still tastes right at the exact moment they're pouring it. To solve this, a research team has introduced a smart, built-in sensor system, combining a type of electronic nose (MOX sensors) with fast, lightweight artificial intelligence (AI) that can analyze the drink's smell and make safety and quality decisions right on the spot. The team tested this technology on four common plant-based milks (oat, almond, soy, and coconut), characterizing the complete collection of volatile organic compounds (VOCs) using gas chromatography-mass spectrometry coupled with solid-phase microextraction (GC-MS-SPME) as a reference method and compared it against standard lab methods. The results show that our electronic nose can quickly and safely identify the unique "fingerprint" of each drink without altering it. A paper based on this work was published in Sensors.¹

Connecting food safety and consumer experience requires the ability to interpret complicated food matrices in real time, and translating the chemical information obtained into actionable, on-device decisions.² Among the different chemical layers accessible for such real-time sensing, the collection of VOCs available for analysis are a particularly informative interface between product composition, processing history, and functional behavior.³ VOCs reflect the nature of raw materials as well as the technological treatments applied during production. For plant-based milk beverages, whose formulation process frequently involves complex emulsions and stabilizing systems, volatile profiles can provide indirect yet meaningful information on both compositional identity and technological performance.⁴

According to the researchers, linear discriminant analysis demonstrated clear discrimination among beverage types based on their volatile signatures, supporting the use of MOX sensor arrays as functional descriptors of compositional identity and process-related variability. Beyond beverage classification, the proposed framework is designed to support future implementation of (i) screening for anomalous volatilomic patterns potentially compatible with accidental cow's milk carryover in shared preparation settings and (ii) adaptive tuning of preparation parameters (for example, foaming-related settings) in smart beverage systems.¹

“The results,” write the authors of the paper,¹ “highlight the role of embedded volatilomic intelligence as a unifying layer between personalized risk-aware screening and sensory-oriented process control, paving the way for intelligent food-processing appliances capable of autonomous, real-time adaptation at the point of consumption.”

The researchers report that future study will concentrate on the expansion of the experimental framework to include controlled dilution series to evaluate robustness under varying concentration levels, batch-to-batch variability assessment to strengthen model generalization, calibrated mixture experiments for the detection of cross-contamination between plant-based and dairy milk matrices, and validation under real-world operational conditions to support practical deployment in smart food-processing environments. The integration of lightweight embedded implementation and real-time validation under operational conditions will also be explored to support practical deployment in smart food-processing environments.¹

Read More on Similar Topics

Advances in Chromatographic Techniques for the Extraction and Characterization of Plant-Based Bioactive Compounds

References

1. Poeta, E.; Sberveglieri, V.; Núñez-Carmona E. Embedded MOX-Based Volatilomic Sensing for Real-Time Classification of Plant-Based Milk Beverages. Sensors (Basel) 2026, 26 (6), 1976. DOI: 10.3390/s26061976
2. Haenlein, G.F.W. Goat Milk in Human Nutrition. Small Rumin. Res. 2004, 51, 155–163. DOI: 10.1016/j.smallrumres.2003.08.010
3. Durgun, M. Real-Time Milk Quality Control Using Multi-Spectral Sensing and Edge Computing: Advancing On-Site Detection of Milk Components with XGBoost. Appl. Sci. 2024, 14, 10916. DOI: 10.3390/app142310916
4. Corvino, A.; Khomenko, I.; Betta, E. et al. Rapid Profiling of Volatile Organic Compounds Associated with Plant-Based Milks versus Bovine Milk Using an Integrated PTR-ToF-MS and GC-MS Approach. Molecules 2025, 30, 761. DOI: 10.3390/molecules30040761