FLUOVIAL 19F NMR & Machine Learning Platform
CASC4DE has developped FLUOVIAL, efficient 19F NMR and Machine Learning techniques for analysis, detection, characterisation, quantification and even identification of fluorinated compounds from NMR spectra. In particular, this platform is suitable for PFAS pollutants (Per- and/or PolyFluoroAlkyl substances classified as “emerging contaminants”).
Thus our FLUOVIAL platform combines:
- a Fluorine NMR spectral database including spectra from various known fluorinated species families,
- a Machine Learning software for an automated analysis of 19F NMR datasets.
Interested in detecting fluorinated compounds, characterising, quantifying and identifying them on your samples ?
The FLUOVIAL platform represents a preliminary step towards comprehensive fluorinated molecule recognition. While it offers fast and automated analysis through Machine Learning, it is currently limited to a specific panel of fluorinated compounds and can identify individual molecules only in isolation, not within mixtures.
Indeed Casc4de has developed a non-specific and non destructive analytical method without previous separation or purification on samples.
Typically, for fluorinated molecules, 19F NMR allows to distinguish:
- aromatic Φ-F,
- -CF3, -O-CF3
- aliphatic -CF2-, -CHF-…
For PFAS compounds, additional information may be available through NMR experiments such as estimation of :
- average length of chain,
- polydispersity,
- functionalization of per- and polyfluoroalkyl chains.
Fluorinated NMR datasets are preprocessed implementing an automatic pipeline based on Plasmodesma.
The work was partly supported by innovation funding programs from ANR, the French National Research Agency and Alsace Innovation agency.

Nowadays fluorine appears in many compounds: pesticides, pharmaceuticals, veterinary drugs and PFC polymers (PFAS,…). Therefore it is found in many emergent environmental contaminants, in particular persistent organic pollutants (POP). Detection by 19F NMR of Fluorine in milk (top left), soil (middle left) and effluent water (bottom left), showing organic fluorine. In particular PFOA molecules in soil and water; the DOSY performed on the water sample shows two molecular populations (bottom right). Current detection limit is 9 ppb (top right).
The results associated with this project were presented at GERM (Milan, Italy) in September 2022 and at the scientific meeting on endocrine disruptors organized by ANSES in June 2024 (Paris, France).
As a result of this work, CASC4DE was able to join the IPANEMA project, which is partially supported by ADEME and involved working with CNRS, INRAE, Gustave Eiffel University, and the company VALGO. Our job is to develop and evaluate a non-targeted fluorine magnetic resonance (NMR) quantification technique that may be used in tandem with mass spectrometry (MS) to characterize PFAS mixes found in environmental matrices.