UltraHighResNMR : Ultra-high resolution proton NMR at no cost. Applications to the investigation of dynamics and interactions in solvate ionic liquids
Description of the PhD project
Ionic liquids are materials with unique properties (thermal stability, negligible vapor pressure…) as solvents for environment-friendly chemistry and efficient energy storage. In particular, solvate ionic liquids (SILs), in which the cation is chelated, are promising candidates to be included as components in a wide range of battery systems. By tailoring the composition of a SIL, its properties can be finely tuned, making SILs promising candidates for safer and more environmentally friendly battery electrolytes.
As the composition of ionic liquid samples becomes more complex, with SILs and water in solvate ionic liquids (WISILs) their investigation by proton nuclear magnetic resonance becomes more challenging. Proton nuclear magnetic resonance is a powerful method to determine the structure, dynamics and
interactions of a variety of molecular systems. We have shown that it is possible to characterize quantitatively dynamics and interactions in mixtures by proton high-resolution relaxometry. This label free approach is broadly applicable and can be applied to a broad variety of systems, from biological fluids to formulations of biological therapeutics. In the UltraHighResNMR project, we will expand high resolution relaxometry to the investigation of ionic liquids and their more complex forms, SILs and WISILs. Unfortunately, proton spectra of mixtures often suffer from spectral crowding, which limits our ability to derive molecule and site-specific information. In the past two decades, pure-shift NMR has been proposed as a solution to the resolution limit. By applying homonuclear decoupling methods, each multiplet in the proton spectrum is replaced by a single narrow peak, increasing the spectral resolution. We have combined pure-shift NMR with high-resolution relaxometry at the cost of 90% of
the signal, making this an impractical solution. Hansen and coworkers have recently shown that convolutional neural networks could be trained to predict pure-shift spectra from a small number ofsensitive spin-echo experiments. We have successfully introduced a new architecture, joint time frequency WaveNet, to train neural networks to predict clean high-resolution relaxometry proton spectra from vibration-distorted experiments. Here, we will expand this architecture to predict pure shift high-resolution relaxometry : a minute increase of 1 to 5% of experimental time will allow us to predict pure-shift spectra with full intensity and calibrated uncertainties. This method will be applied to a variety of ionic liquids, mostly (water in) solvate ionic liquids, including the more environment friendly compositions, where the anion is not fluorinated. We will use pure-shift high-resolution
relaxometry to investigate the nature of interactions, rotational, and translational diffusion, as well as site-specific dynamics in (WI)SILs, which are promising candidates for electrolytes in batteries. Access to relaxation measurements from 100 µT to 21 T will provide us with a comprehensive picture of molecular dynamics from picoseconds to a few microseconds. Experimental data from relaxometry will be interpreted with advanced models of relaxation (collaboration with Danuta Kruk) and molecular dynamics simulations (collaboration with Ralf Ludwig). Our approach will be generally applicable to all soft materials that are amenable to proton NMR and will open a new window on the molecular properties of soft matter.
Keywords
Pure-shift NMR ; Convolutional neural networks ; high-resolution relaxometry ; ionic liquids
Research Unit, UMR number and acronym
ENS - Physical Chemistry and Chemistry of Life (CPCV) UMR 8228
Description of the research Unit/subunit
The research unit CPCV (Physical Chemistry and Chemistry of Life) at ENS has a very broad scope in chemistry, with research encompassing synthesis, theoretical chemistry, spectroscopy, electrochemistry, soft matter in a biological context, biophysical chemistry, etc. The NMR team at CPCV is specialized in the development of new instruments and methods in NMR with applications to a broad range of materials and molecular systems, from biological macromolecules to inorganic materials, biological fluids to polymerization catalyzers.
Name of the supervisor
Fabien Ferrage (Fabien.Ferrage@ens.psl.eu)
Name of the co-supervisor
Guillaume Bouvignies (guillaume.bouvignies@ens.psl.eu)
3i Aspects of the proposal
The PhD project will take place in the unique environment of the NMR team at ENS, including a brand new 900 MHz system equipped with a world unique relaxometry prototype. The PhD project will be supervised by both Fabien Ferrage and Guillaume Bouvignies. Both have supervised PhD students, who have pursued successful carriers in research, industry and teaching. They have already co-supervised two PhD theses. Supervisors will be available for informal discussions daily. Individual meetings (30 mins to 1h) are organized every Friday to discuss the project, progress during the week and planning for the week ahead. Three PhD students will defend their PhD theses in 2026 so that each supervisor will supervise only one PhD student in 2027, in addition to the UltraHighResNMR student.
Intersectoriality
The UltraHighResNMR project is based on a 15-year-long collaboration between the NMR team at CPCV and the company Bruker Biospin, which is the world leader in NMR instrumentation. Bruker and the ENS team have worked together to design, build, and exploit new classes of instruments that couple high-resolution high-field NMR and low magnetic fields. The experimental part of the project will take place on a unique field-cycling system that is one of the fruits of this long-term collaboration. The UltraHighResNMR project will provide vastly enhanced tools to analyse the experiments recorded on this field-cycling system and expand the scope of this instrument, opening up a new class of analytical tools to the soft matter community.
International
The team has a longstanding collaboration with Pr. Snoeijer at the University of Twente, including the preliminary study of breath figures, which is relevant for the proposed thesis. He is a leading expert in (soft) wetting dynamics and is working on the development of predictive models for nucleation on polymeric surfaces, a core challenge in this project. It is therefore natural that the candidate spends a period of time at Pr. Snoeijer’s lab.
Interdisciplinarity
This project is clearly positioned at the interface of chemistry, physics, and artificial intelligence. The materials under investigation are ionic liquids, soft materials which can be used for energy storage. The experimental approach, NMR spectroscopy, is clearly positioned at the interface of physics and
chemistry, so are molecular dynamics simulations. The methods that will be developed are based on machine learning, more precisely convolutional neural networks, making use of original architectures designed for signal processing.
Expected profile of the candidate
This interdisciplinary project would be a good match for a student with initial training in at least one – and ideally two – of the three domains : chemistry, physics, or artificial intelligence. Some experience
with coding will be considered a strong plus. Experience in NMR would be useful but not necessary (we can train students). Above all, we are looking for a candidate who is curious to learn and explore new fields. We strongly value independence and creativity. .
Important dates
Call for applications : from September 1st to October 31st 2026
Eligibility check results : November
3i Committee evaluation results : December
Interviews from the shortlisted candidates with the Selection Committee : January 2027
Start of the PhD : March 1st 2027