Designing optimal loop, saddle, and ellipse-based magnetic coils by spherical harmonic mapping

Abstract

Adaptable, low-cost, coils designed by carefully selecting the arrangements and geometries of simple primitive units are used to generate magnetic fields for diverse applications. These extend from magnetic resonance and fundamental physics experiments to active shielding of quantum devices including magnetometers, interferometers, clocks, and computers. However, finding optimal arrangements and geometries of multiple primitive structures is time-intensive and it is challenging to account for additional constraints, e.g. optical access, during the design process. Here, we demonstrate a general method to find these optimal arrangements. We encode specific symmetries into sets of loops, saddles, and cylindrical ellipses and then solve exactly for the magnetic field harmonics generated by each set. By combining these analytic solutions using computer algebra, we can use numerical techniques to efficiently map the landscape of parameters and geometries which the coils must satisfy. Sets of solutions may be found which generate desired target fields accurately while accounting for complexity and size restrictions. We demonstrate this approach by employing simple configurations of loops, saddles, and cylindrical ellipses to design target linear field gradients and compare their performance with designs obtained using conventional methods. A case study is presented where three optimized arrangements of loops, designed to generate a uniform axial field, a linear axial field gradient, and a quadratic axial field gradient, respectively, are hand-wound around a low-cost, 3D-printed coil former. These coils are used to null the background in a typical laboratory environment, reducing the magnitude of the axial field along the central half of the former's axis from $\left(7.8\pm0.3\right)$ $\mu$T (mean $\pm$ st. dev.) to $\left(0.11\pm0.04\right)$ $\mu$T.