Multi-Objective Solid-State NMR

Description

The vast majority of solid-state NMR experiments‭ ‬in existence today‭ ‬is based on development and analyses by average Hamiltonian theory‭ (‬AHT‭) [‬1‭]‬.‭ ‬After formulation of AHT in the late‭ ‬60th the limitations of this approach quickly became apparent too.‭ ‬Especially presence of multiple anisotropic spin-interactions together with simultaneous spin-modulation by radio-frequency pulses and sample rotation at the magic-angle,‭ ‬present multi-parameter scenarios that are mostly intractable by means of analytical calculations by AHT.‭ ‬These limitations brought computer based search and‭ ‬optimisation techniques,‭ ‬like optimal control theory‭ (‬OCT‭) [‬2‭]‬,‭ ‬to the scene.‭ Optimisation,‭ ‬especially when conducted by computers requires measures‭ (‬figures of merit‭) ‬by which to quantitatively assess the quality of a solid-state NMR pulse sequences.‭ ‬Historically spectroscopists use terms like:‭ ‬efficient,‭ ‬broadbanded,‭ ‬selective,‭ ‬sensitive,‭ ‬robust,‭ ‬high-resolution,‭ ‬low-power,‭ ‬etc.‭ ‬for this purpose.‭ ‬While it is relatively straight forward to quantitatively asses NMR experiment with respect to a single one of these qualities,‭ ‬optimisation with respect to two or more simultaneously,‭ ‬poses a considerable challenge due to the different characters of these qualities. Here we demonstrate how such multi-objective optimisation problems can be tackled in an intuitive fashion by using the concept of Pareto optimality in connection with a genetic algorithm optimisation approach.‭

Period	25 Apr 2016
Event title	NMR Valtice, 2016, 31th Central European NMR Meeting
Event type	Conference
Location	Czech RepublicShow on map