Phase variation in the locked state of mutually synchronized spin torque nano-oscillators.

Spin torque nano-oscillators (STNOs) have been shown to efficiently "lock" to electrical signals whether these are from external sources, self-reflected signals, or other STNOs via mutual synchronization. Traditionally, the locked state of an STNO has been considered digital, either "ON" or "OFF." In this report, we show how the phase in the locked state can vary significantly as a function of the applied dc to the STNO with the phase strongly affecting the emitted power. This phase dependence demonstrates the analogue nature of the locked state, offering an extra degree of freedom when designing STNO-based neural networks. [ABSTRACT FROM AUTHOR]