Dynamic Electrical Pathway Tuning in Neuromorphic Nanowire Networks

Abstract

Neurobiology-inspired phenomena such as winner-takes-all competition and critical dynamics have been recently reported to arise in neuromorphic nanowire networks. These are unique systems where interactions between memristive elements creates emergent conductance pathways between discrete electrodes. This mode of operation can offer substantial advantages to create a truly concomitant plastic-static system for integration in neuromorphic devices. However, critical aspects such as pathway controllability and stability are yet to be explored. In this study, pathway formation in self-assembled neuromorphic networks formed by Ag nanowires decorated with TiO 2 nanoparticles is investigated. Direct visualization of pathway formation through a neuromorphic network is attained using the lock-in thermography technique. Using this technique, it is demonstrated that how networks preserve information from previously used pathways through increased local junction connectivity. This effect directly reshapes subsequent formation of pathways whenever the spatial location of the electrodes is dynamically changed. Combining these results with conventional current-voltage measurements, which show that the network electrically acts as a volatile switching memristor, a unique interaction between short-term and long-term memory arises. This produces unexpected collective dynamical states of potentiation and inhibition of network conductance whenever different spatiotemporal signals are dynamically fed to the network.

Neuromorphic nanowire networks: principles, progress and future prospects for neuro-inspired information processing

“…A distinctive feature observed in neuromorphic NWNs is the emergence of a first 'winner takes all' (WTA) optimal transport path concomitant with network activation [46,48,55,[73][74][75], which occurs when connected junctions between electrodes collectively switch from HRS to LRS (Figure 6(a)). Under a DC bias, a NWN responds like a single unipolar resistive switch, with an abrupt shift in conductance from a low to a high state (G off → G on , with G on ≫ G off ) [74].…”

“…The controllability of emergent current paths in NWNs was investigated by the NIMS-MANA group using self-assembled networks of Ag nanowires decorated with TiO 2 nanoparticles [75]. Using lock-in thermography to visualize dynamic electrical pathway formation (Figure 7), they found that previously established pathways influence the formation of subsequent pathways, thus demonstrating a collective memory effect arising from the connected memristive junctions.…”

Information dynamics in neuromorphic nanowire networks

“…Previous studies have shown that NWNs exhibit complex, neuromorphic structural properties such as small-worldness, modularity and recurrent feedback loops 24 . In response to electrical stimulation, NWNs exhibit emergent collective neural-like dynamics 25 – 31 , including a “winner-takes-all” (WTA) electrical transport path 18 , 27 – 30 , 32 , 33 . This feature emerges above a threshold voltage when connected edge junctions collectively switch to a low resistance state, thereby activating the network into a highly conducting state.…”

Planar and Transparent Memristive Devices Based on Titanium Oxide Coated Silver Nanowire Networks with Tunable Switching Voltage

“…In order to gain further insight on the main factors governing the switching mechanism that takes place in our system, we evaluated the relationship between network density and the power needed to induce the switching. The dependence of the electrical power generated with the applied voltage for [34] AgNW/AgO x core−shell Planar 100 ≈0.1-3 Du et al, 2017 [15] PI/AgNWs/TiO 2 /Pt Vertical 200 1.1 Yeom et al, 2017 [18] AgNWs/AgO x /AgNWs Planar 5×10 5 0.5-20 Wan et al, 2018 [19] AgNW and TiO 2 NPs Planar ->100 Li et al, 2020 [17] AgNWs/TiO 2 core-shell Vertical 10 6 0.1 Kim et al, 2020 [35] AgNWs/TiO 2 Planar 10 6 0.16-5.2 This work different network densities is plotted in Figure 5b. Remarkably, we observed that the threshold switching takes place at similar electrical power values, regardless of the network density used, represented by the color dots in Figure 5b.…”

“…[6] The combination of metallic nanowire networks with oxide thin films has been quite successful in the creation of stable transparent and flexible electrodes for many different applications, such as transparent heaters, [7,8] photovoltaics, [9] efficient lighting, [10] or smart windows. [11] Different metallic networks based on silver nanowires (AgNWs) have also been applied in memristive devices in both planar [12][13][14][15][16][17] and vertical configurations. [18][19][20] These nanowire-based devices present two different switching mechanisms: conductive filament formation and tunneling mechanism.…”