In an era where the rapid rise of artificial intelligence is accompanied by exponentially increasing energy costs, a promising approach involves harnessing ambient thermal noise as an ultra-low-power computing resource.
This strategy, which biology appears to already exploit in the brain — where noise helps neurons explore and decide — has prompted researchers to design noisy nano-components capable of emulating neurons within electronic chips dedicated to computing. To achieve this, magnetic memory-type nano-neurons have been developed: superparamagnetic tunnel junctions (SMTJs).
SMTJs consist of a free magnetic layer and a fixed magnetic layer, separated by an insulator. The relative orientation of magnetization in these layers, parallel or antiparallel, corresponds to two metastable states separated by an energy barrier.
In this study, the specific design allows SMTJs to be highly sensitive to ambient thermal noise, unlike conventional applications (memories and sensors). Indeed, simple thermal fluctuations can randomly reverse the magnetization of the free layer. Thus, these SMTJs behave as stochastic binary neurons whose advantage is consuming very little energy. The shorter the average waiting time between magnetic reversals, the higher the computing speed.
A team from
CEA-Irig/SPINTEC has experimentally demonstrated waiting times between magnetization reversals in perpendicular magnetization superparamagnetic tunnel junctions miniaturized to 50 nm in diameter, purely induced by thermal fluctuations. The measurement requires very low currents to observe changes in the orientation of the free magnetic layer on the scale of a few nanoseconds, a timescale never before observed in these systems.
(a) Schematic of a magnetic tunnel junction. The free layer can be parallel (P) to the fixed layer (green), or anti-parallel (AP) (purple).
(b) Time evolution of the voltage in an SMTJ showing the waiting times between reversals on the scale of a few nanoseconds between the P (-2 mV) and AP (+2 mV) states.
(c) Schematic of the energy landscape associated with magnetization reversal.
© CEA
The average waiting times measured are much shorter than predictions from conventional models, which researchers theoretically interpret as a significant contribution from entropy, increasing the probability of crossing the energy barrier separating the magnetic states of the SMTJ. Entropy reflects the number of magnetic configurations accessible to the system.
In perpendicular magnetization SMTJs, the intermediate states where magnetization gradually switches from parallel to antiparallel orientation (and vice versa) are numerous. This contributes to high entropy by increasing the number of different ways to transition between states.
Under the effect of thermal energy alone, perpendicular magnetic tunnel junctions with diameters of only a few tens of nanometers randomly switch from one state to another, with ultra-short average waiting times on the order of nanoseconds.
By capitalizing on these fluctuations as a magnetization reversal mechanism, this work paves the way for implementing stochastic elements for ultra-low-power neuromorphic computing.