Raman gas spectroscopy for detection of bacterial contamination in milk

Traditional microbiological methods for detecting clostridium spores in milk are slow and lack specificity for these bacteria. Clostridia are well known to cause so-called late blowing defects in cheese, due to their high production of gas. This doctoral work focused on developing an automated instrument based on Raman spectroscopy for the non-invasive detection of hydrogen and other gases (CO2, N2, O2, H2O) in the headspace of sealed vials. The prototype operates within the framework of the standard Most Probable Number (MPN) method for spore enumeration but overcomes its main limitations by providing specificity for clostridia and significantly reducing measurement times, nearly halving the duration of experimental campaigns. The developed system integrates a vial loader, instrument thermalization, and dedicated software for automated sample handling and data analysis. The software includes algorithms to optimize camera integration times, vial positioning, and spectral acquisition. For the first time, Raman spectroscopy was employed to monitor headspace gas dynamics, enabling the tracking of hydrogen production, which exhibited an exponential trend consistent with theoretical expectations. The instrument was made portable and ready for use in the relevant industrial environment (TRL6). Although optimized for the detection of clostridia in milk, the instrument is highly versatile, allowing the simultaneous monitoring of multiple gases in vials in a non-invasive, contactless, and non-destructive manner. Its portability and flexibility make it suitable for potential future applications beyond the dairy sector.