Nuclear magnetic resonance (NMR) spectroscopy is a widely-used tool for molecular characterisation. It relies on a property of certain atomic nuclei called spin in order to gain information about molecules. A bewildering number of NMR experiments exist, each of which provides different structural and dynamical characterisations of chemical samples. All of these experiments require a range of radiofrequency (RF) pulses to be suitably arranged in ‘sequences’ in order to organise the nuclear spins and, ultimately, generate informative spectra.
The Baldwin group have recently published a method that draws heavily on quantum information methodology from the group of Jonathan Jones (Department of Physics). The result is a new method (called seedless) that create RF pulses for magnetic resonance experiments. With this technology, pulses – carefully tailored to specific samples and hardware – can be produced in just a few seconds. The result is that in almost all experiments they were able to get more signal, by compensating for non-ideal features in NMR hardware. The method also enables some new applications that previously were not possible, such as ultra-broadband excitation of 19F nuclei.
The group of Ray Freeman led the way in Oxford, producing optimised RF pulses for NMR in the 1980s. This paper updates this earlier work for modern instruments. With seedless, the team propose a new paradigm for magnetic resonance, where optimal pulses for experiments are created to precisely match the characteristics of individual samples. Owing to its computational efficiency, this process takes just a few seconds on a laptop, and so can be done ‘on-the-fly'. The team believe that the success of this approach indicates that there is still room for non-artificial intelligence in algorithm design and chemical analysis.
The software can be downloaded from https://seedless.chem.ox.ac.uk, and you can read more in Nature Communications.
