Systematic errors in the determination of the spectroscopic g-factor in broadband ferromagnetic resonance spectroscopy: A proposed solution

A theoretical and experimental study of the influence of small offsets of the magnetic field (δH) on the measurement accuracy of the spectroscopic g-factor (g) and saturation magnetization (M_s) obtained by broadband ferromagnetic resonance (FMR) measurements is presented. The random nature of δH generates systematic and opposite sign deviations of the values of g and M_s with respect to their true values. A δH on the order of a few Oe leads to a ∼10% error of g and M_s for a typical range of frequencies employed in broadband FMR experiments. We propose a simple experimental methodology to significantly minimize the effect of δH on the fitted values of g and M_s, eliminating their apparent dependence in the range of frequencies employed. Our method was successfully tested using broadband FMR measurements on a 5 nm thick Ni₈₀Fe₂₀ film for frequencies ranging between 3 and 17 GHz.