Frequency-Switching Inductive Link: Enhancing Power Transfer Efficiency for Wireless, Miniaturized and Distributed Neural Implants.

In recent years, wireless implants have undergone significant development, becoming valuable biomedical instruments for monitoring and treating neurological and mental disorders. Among the others, inductive power transfer is the most promising method for powering multiple, distributed and miniaturized implants. However, achieving high efficiency within the safety-limited specific absorption rate is challenging. High frequencies (above a few hundred MHz) enable miniaturized receivers and higher efficiencies, but lower frequencies (below a few GHz) are needed to avoid heat and power dissipation. The recently introduced frequency-switching inductive link represents a significant innovation in wireless power transfer. This system exploits an active resonator to switch frequencies from a lower (13.56 MHz) to a higher one (433.92 MHz), enabling both miniaturization and good efficiency within safety limits. Power is supplied by an external transmitter through a 2-coil low-frequency link and transferred to 1024 miniaturized implants via a 3-coil high-frequency link with 37 passive hexagonal resonators. Despite its initial promising prospects, several aspects should be investigated. This master's thesis aims at a comprehensive understanding of the frequency-switching approach, thus including all the components and all the non-resonant contributions in the efficiency equation. In this way, power transfer efficiency without first-order approximations can be achieved. To this end, analytical equations were derived from the complete impedance matrix and the efficiency was calculated in MATLAB. The values of the additional coupling coefficients were extracted from the complete finite element modelling in ANSYS and electrical equivalent circuital simulations were conducted in Cadence to compare this results with analytical calculations. A parametric analysis was carried out to construct the matching network, with the goal of optimizing the power transfer when considering all the components within the frequency-switching link. In this complete scenario, the efficiency of the low-frequency link remains consistent at 21.8% with and without approximations. However, for the high-frequency link, the efficiency decreases by almost two orders of magnitude compared to the case with approximations, primarily due to non-resonant contributions. After optimizing the matching networks almost 15 dB in terms of S21 were recovered, resulting in a difference of only 2 dB from the approximated case. The overall power transfer efficiency of 0.019%, obtained by multiplying the efficiencies of each block, represents a promising alternative compared to state-of-the-art three-coil inductive links. To conclude, considering all the components of frequency switching is fundamental, as a strong coupling exists between non-resonant coils. This approach, which is novel compared to current literature, not only makes the performance of frequency switching more representative, but also allow for more precise efficiency calculations in a generic N-coil system. Indeed, both the derived equations and the network optimization algorithm used in this work can be applied to any multi-coil system, making it essential for optimizing inductive links in miniaturized and distributed neural implants.