Could you explain the significance of your article to the non-specialist?

Nuclear magnetic resonance is the most extensively used characterisation tool in chemistry and materials science. Its spectral parameters constitute a challenging and phenomenologically rich object of theoretical and computational investigation. The article covers the current state-of-the-art in the calculation of NMR parameters by electronic structure and simulation methods, with a particular emphasis on realistic, experiment-oriented modelling where factors such as relativistic, thermal and intermolecular interaction- induced effects are taken into account.

What has motivated you to conduct this work?

The rapid increase in computer power and the simultaneous development of efficient computational algorithms have opened the way to increasingly realistic modelling of magnetic resonance through the last two decades. This way theoreticians have been able to make themselves essential to experimental research programs, not only in virtually all chemical and materials scientific subdisciplines, but also in biology and medicine. The field is important, multidisciplinary and conceptually challenging.

Where do you see this work developing in the future?

As the methodology is developed, experimentalists can increasingly adopt routine NMR parameter calculations to their normal workflow of spectral registration, analysis and reporting. At the same time, the frontiers of the field move from basic phenomenology demonstrated on small molecules to systems modelling, where large length and time scales are tackled. This means that not only the molecules of interest but also their environment are routinely included in the computations, and the time evolution (dynamics) of the nuclear framework is followed. First-principles theory of NMR relaxation develops. Periodic models will supersede cluster models of continuous solids.

Are there any particular challenges facing future research in this area?

NMR parameters are determined by hyperfine interactions between nuclei and the electrons. These interactions are energetically small and their spatial origin is very localised. These facts make it challenging to develop coarse-grained, preferably non-atomistic modelling paradigms where large length- and timescales can be recovered, for NMR properties. Density-functional theory is indisposable as the electronic structure tool for large systems, also in the field of NMR calculations. Despite its success in energetics and structures of molecules and solids, for magnetic properties the currently used approximate exchange-correlation functionals do not constitute universally applicable, reasonably accurate methods. For properties such as spin- spin and hyperfine coupling tensors, large errors prevail for some nuclei, whereas for others the predictive power of DFT is already very satisfactory. Systematic, first principles development of accurate exchange-correlation functionals that also work for hyperfine interactions throughout the Periodic Table, is a major challenge.