Modular Dual Polarity Py-GC-IMS Enhanced by ML/DL Algorithms for Chemical and Biological Threat Detection (TeChBioT)

Description

Rapid and reliable identification of biological and chemical warfare agents (BWAs and CWAs) after deliberate releases or natural outbreaks is critical for effective response in military and civilian contexts. First responders are often confronted with unknown hazards, requiring field‑deployable, broad-spectrum detection capabilities. Current technologies typically address either CWAs or BWAs, creating operational gaps and delaying decision‑making.

Ion mobility spectrometry (IMS), alone or coupled with gas chromatography (GC), is widely used for detection of volatile CWAs. However, the low volatility of certain CWAs, besides the molecular complexity and reduced volatility of BWAs, limits its broader applicability.

To overcome these limitations, the TeChBioT consortium developed a modular dual‑polarity-IMS system with a temperature control up to 140 °C, operable as a stand‑alone instrument or coupled to a high‑temperature GC. The platform is enhanced with machine learning (ML) and deep learning (DL) algorithms to improve compound identification. Using this system, five CWA simulants and eight CWAs, including low‑volatility compounds, were detected with sensitivities in the pptV range, even in the presence of complex background interferences such as gasoline vapors. To evaluate the device under realistic conditions, the prototype was mounted on a “Ziesel” unmanned ground vehicle during outdoor exercises, including aerosol‑release scenarios. In both stationary and mobile modes, it reliably detected a CWA simulant.

For BWA detection, the GC‑IMS system was coupled with a custom pyrolyzer. Rapid pyrolysis of bacterial and viral suspensions (520 °C, 10 s) generated characteristic volatile profiles dominated by membrane‑derived fatty acids. ML and DL models enabled proof‑of‑concept discrimination between Gram‑positive and Gram‑negative bacteria, and between enveloped and non‑enveloped viruses, with accuracy and F1-scores exceeding 80%. Field exercises showed that current bacterial sensitivity (≃10¹² CFU/measurement) could further be improved by sample pre‑concentration via filtration.

Thus, TeChBioT system establishes the feasibility of a unified, deployable IMS‑based platform for integrated CWA and BWA detection under operationally realistic conditions.

Période	21 mai 2026
Titre de l'événement	7th International conference CBRNE Research & Innovation
Type d'événement	Une conférence
Emplacement	Arcachon, FranceAfficher sur la carte
Degré de reconnaissance	International