Abstract

The mutual synchronization of spin-torque nano-oscillators (STNOs) arranged in a large array improves the individual properties of each STNO, increasing the output-power and reducing the linewidth of the emitted signal. These improvements are required to implement STNOs in wireless communication schemes [1] and furthermore open applications for different oscillator based neuromorphic computing schemes [2]. Many theoretical and experimental studies considered single STNO devices [3], while experimentally the synchronization of a large number of STNOs remains a challenging task [4]. This is because the synchronized state and the phase-locking depend on the geometry, device resistance, input signals, coupling mechanism, among others [5].As a first step to guide experiments on the synchronization behavior of STNO arrays, we have undertaken a theoretical study, combining numerical and analytical approaches based on the spin-wave formalism [3]. The STNO devices considered here are of circular cross-section (100nm in diameter) with a perpendicular polarizer and an in-plane magnetized free layer under a strong out-of-plane field. This field is larger than the saturation field. The free layer magnetization is orientated out-of-plane leading to an out-of-plane precession mode that is advantageous to induce strong dipolar coupling between STNOs, which is the coupling mechanism considered here. We derived analytically the coupled equations of amplitudes and phases where the latter are similar to the Kuramoto-phase equation, but limited to a finite number of devices. Solving for the stationary states, this model provides analytical expressions for the locking-range, the stationary amplitudes, and the frequency of oscillations of the synchronized state. These solutions depend on the array geometry, number N of STNOs in the array, separation between STNOs, distribution of the DC current, as well as the different magnetic material parameters, and the external magnetic field. In this work, we investigate the solution of these equations for two specific array geometries: a straight line and a ring structure (N=3,6 and 10), shown in figure 1b.Our results show that the locking-range depends on the separation among the STNOs (because this scales the coupling strength), but it is independent of the angle among them. Furthermore, we show that the locking-range is larger when the STNOs are placed in a ring structure (or triangle for N=3) as compared to a straight-line structure. Finally, the most interesting result is that for the ring structure, we identify two dynamic modes that represent a standing wave for the STNO phases along the ring structure (observed for N>5). These two dynamics modes are shown in Fig. 2 and are called here (a) the in-phase mode, with identical phases for all STNOs, and (b) the out-of-phase mode, where the phase difference between two adjacent STNOs is 2π/N. We study how these states and their existence range are modified upon variation of the current density through the different STNOs (identical and non-identical currents), the STNO separation, and the number of STNOs in the ring. The analytical results are in good agreement to the numerical solutions of the Landau-Lifshitz-Gilbert (LLG) equation, as long as the separation among the STNOs is above a certain critical value. For smaller separations, the dipolar interactions become too strong and non-linear effects would require to go beyond the first-order approximations used to solve the amplitude and phase equations. The results reported here represent a first step to validate the phase equations and to define their limitations, and can be used to address now the Kuramoto equations to find solutions for large scale structures (N very large).We acknowledge financial support in Chile from FONDECYT 1200867, and Financiamiento Basal para Centros Científicos y Tecnológicos de Excelencia AFB 180001. D. M.-A. acknowledges Postdoctorado FONDECYT 2018, 3180416. M. A. Castro acknowledges Conicyt-PCHA/Doctorado Nacional/2017-21171016. The project was supported in part by ERC Advanced Grant MAGICAL No. 669204 **

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call