NMR Imaging of Non-Markovian Diffusion Transport

With the support of the Chemical Measurement and Imaging Program in the Division of Chemistry, Professor Louis Bouchard at the University of California Los Angeles will pioneer a measurement strategy poised to transform routine nuclear-magnetic-resonance (NMR) spectra into nanoscopic probes of molecular motion inside pores thousands of times narrower than a human hair. By analyzing the intrinsic shape of the NMR signal rather than relying on highly specialized and inaccessible hardware that generates extremely strong magnetic-field gradients, the project would allow researchers to quantify surface and bulk properties of porous materials, estimating pore dimensions, viscosity, and surface adhesion in catalysts, energy-storage devices, filtration media, and even human lungs. The effort would engage graduate, undergraduate, and high-school students in simulations, bench experiments, and open-source code development. Outreach activities, including hands-on demonstrations in local middle schools, would introduce K-12 learners to molecular imaging and highlight career pathways in chemical measurement science. Technically, the research team would integrate atomistic molecular-dynamics simulations with a generalized Langevin equation whose memory kernel is obtained from Green–Kubo time-correlation integrals, generating time-dependent diffusion coefficients and velocity-autocorrelation functions that preserve the molecular “memory” of prior collisions, surface encounters, and long-ranged hydrodynamic interactions in confined gases and liquids. These computed diffusion fingerprints would then be benchmarked against variable-temperature NMR spectra collected under carefully calibrated weak residual gradients in well-characterized porous media, including inverted-colloidal-crystal scaffolds, mesoporous silica, and single-wall carbon nanotubes, to ensure that the model reproduces experimental linewidths across a wide range of temperatures, pore diameters, and interaction strengths. By expressing the Green–Kubo response as separable functions of temperature, pore geometry, and intermolecular-potential depth, the investigators would feed those parameterized functions into an evolutionary optimization engine that iteratively adjusts the simulated spectrum until it converges on the measured lineshape, thereby solving the inverse problem of extracting nanometer-scale pore size, viscosity, and surface adhesion energies directly from a single NMR scan. Cross-validation against electron microscopy and gas-adsorption benchmarks would quantify uncertainties and refine the model. Once validated, the framework would be generalized to frequency-encoded pulse sequences and embedded in diffusion-weighted MRI protocols, furnishing clinicians, materials chemists, and energy researchers with a physics-based route to interpret non-Markovian transport in biological tissue, catalytic reactors, battery electrodes, and filtration membranes without relying on empirical calibrations, thus opening new avenues for quantitative molecular imaging and rational microstructural design. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.