Sample-Induced Resistance Estimation in Magnetic Resonance Experiments: Simulation and Comparison of Two Methods

Signal-to-noise ratio estimation in magnetic resonance experiments requires the knowledge of sample-induced resistance value, where the sample can be protein solutes, cell suspensions, plants, animals, portions of human body or saline solution phantoms. Many authors studied sample–coil interaction using homogeneous infinitely long cylinders, spheres or half-space as approximations of the sample geometry. However, in real magnetic resonance experiments, both sample shape and dimensions can be very different with respect to these models. This paper describes and compares two different methods developed by the authors for sample-induced resistance estimation, both useful for predicting the performance of radio-frequency coils strictly coupled to the sample, where the knowledge of a sample–coil interaction model permits to estimate the different noise contributors. The main goal of our research is testing the proposed algorithms and finding their limitations by comparing their performances for a simple case which uses a sample simplified geometry. The first method, based on the magnetostatic approach, employs vector potential calculation and can be easily implemented for simple coils and sample geometries. The second method uses finite-difference time-domain algorithm and permits to simulate systems with various geometries, without approximations in sample and coil geometries. Comparison with experimental data, performed on three homebuilt surface coils each of them successively tuned at three different frequencies, demonstrated the differences in accuracy of the developed methods.