CAREER: Enhanced Sensing and Spectral Analysis for Zero-to-Ultralow Field Nuclear Magnetic Resonance

This research will advance a form of zero-field nuclear magnetic resonance (ZF NMR) detection spectroscopy for high-information chemical analysis. While conventional, high-field NMR spectroscopy typically relies on expensive, homogenous, superconducting magnets, in this approach, samples would be briefly polarized and then measured in a shielded, near-zero-field environment using compact atomic quantum sensors, enabling a path toward low-cost instruments that could operate in parallel rather than one sample at a time. By overcoming sensitivity and analysis bottlenecks, the project would help unlock distributed chemical fingerprinting for applications such as faster reaction screening and quality control, and it could simplify measurements in water-rich samples where conventional NMR often requires complex suppression methods. The effort would also contribute to workforce development by training undergraduate and graduate researchers in research instrumentation, quantum sensing, and quantum information methods, and by engaging local high school and community college students through lab visits and hands-on exposure to modern chemical measurement science. Technically, the research will focus on four complementary thrusts: (1) sensitivity improvements intrinsic to the laboratory ZF NMR instrument; (2) development of machine-learning tools for prediction of J-couplings and zero-field NMR spectra; (3) quantum-computing-assisted algorithms for scalable spectral inference (Hamiltonian learning); and (4) zero-field-compatible optical hyperpolarization. In the first thrust, the team will pursue a set of instrumentation advances to improve sensitivity, including increasing sample–sensor coupling, deploying small arrays of optically pumped magnetometers (OPMs) to expand the effective detection volume and enable common-mode noise rejection, reducing environmental magnetic noise through improved shielding and low-noise electronics, and implementing compact high-field prepolarizers together with faster sample shuttling to minimize polarization loss. These improvements will be quantitatively benchmarked across representative small molecules using standardized sensitivity metrics. To enable scalable spectral interpretation, the project will build an experimental library of zero-field spectra for a broad range of small organic molecules and use these data to develop machine-learning models capable of predicting J-couplings and full ZF NMR spectra directly from molecular structure. In parallel, spectral analysis will be formulated as a parameter-estimation problem in which a candidate J-coupling Hamiltonian is iteratively refined to reproduce experimentally measured time-domain signals or spectra. Hybrid quantum–classical algorithms will be developed to accelerate the simulation and observable estimation required for this refinement as the size of coupled spin networks grows beyond the regime where exact classical simulation is practical. Finally, the team will develop reproducible triplet-based optical polarization sources and engineer a rapid crush-and-dissolve transfer workflow that delivers enhanced nuclear polarization into liquid analytes. Integration of this workflow with ZF NMR detection will allow systematic characterization of polarization enhancement factors, polarization lifetimes, and their effects on spectral resolution and repeatability across a wide panel of analytes. 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. NSF Award ID: 2544648 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Ashok Ajoy | Institution: University of California-Berkeley, BERKELEY, CA | Award Amount: $575,078 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2544648 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2544648.html