Fig. 6: Probabilistic bit with double-free-layer stochastic MTJs.
From: Connecting physics to systems with modular spin-circuits
a Self-consistent magnet and transport model combined with transistors to model a probabilistic bit. b Time-dependent spin currents are produced from the transport model that goes into the sLLG modules. We show the x-axis component of spin–currents for magnets, 1 and 2. c Histogram and time fluctuations for the \(\cos (\theta )\) between mz components of magnet 1, 2 for the double-free-layer sMTJ. Slight anti-parallel tendency is due to the dipolar coupling which is not completely overcome by thermal fields. d Resistance of the sMTJ is measured while the voltage is swept from −0.5 to 0.5 V over 1 ms. The discrete data points are average resistances over 500 ns showing the roughly bias-independent characteristics of the device. e The drain node (\({{\mathsf{V}}}_{{\mathsf{D}}}\)) and the output of the inverter (\({{\mathsf{V}}}_{{\mathsf{OUT}}}\)) are measured while the input (\({{\mathsf{V}}}_{{\mathsf{IN}}}\)) is swept from −0.3 V to 0.3 V over 1 ms. The output of the inverter shows the binary stochastic neuron behavior. f Digital output fluctuations over time for the probabilistic bit output at different bias conditions for \({{\mathsf{V}}}_{IN}\).
