U-Shaped Spin-Wave Transducers: Excitation and Design Rules for Performance

The need for more efficient and high-performance technologies to keep up with the demands of the modern world requires innovative solutions beyond conventional electronic technologies. Spin-wave (SW) technology promises this future to electronics through magnonic devices, but challenges to large-scale adoption persist due to SW transduction in these systems. Here, a micromagnetic simulation approach followed by a regression analysis, providing a perspective on optimizing the SW transduction for a specific case, is presented. Using the micromagnetic simulations in conjunction with analytical methods offers insight into this magnonic system and the influence of the transducer parameters on the SW excitation. The approach is validated by performing several different simulations involving parameter studies of the transducer, performing a regression analysis on one such study, and deriving appropriate design rules to support the heuristically derived ones. A key component of this approach is the use of Gaussian-process (GP) regression as a surrogate modeling framework to extract physically interpretable design rules from high-dimensional micromagnetic simulation data, enabling efficient exploration of coupled parameter effects that are not accessible through conventional single-parameter studies. The results from these studies show that independently varying the parameters of the transducer is markedly different from interdependently varying them, as delineated by the design rules derived by employing regression analysis. The approach thus demonstrated can be advanced further by modifications for more accurate trends or even the inclusion of other non-trivial physics in the studies.