Anisotropic Black Phosphorus Synaptic Device for Neuromorphic Applications

Abstract: Synapses are functional links between neurons, through which "information" flows in the neural network. These connections vary significantly in strength, typically resulting from the intrinsic heterogeneity in their chemical and biological properties. Such heterogeneity is fundamental to the diversity of neural activities, which together with other features of the brain enables functions ranging from perception and recognition, to memory and reasoning. Realizing such heterogeneity in synaptic electronics is critical towards building artificial neural network with the potential for achieving the level of complexity in biological systems. However, such intrinsic heterogeneity has been very challenging to realize in current synaptic devices. Here, we demonstrate the first black phosphorus (BP) synaptic device, which offers intrinsic anisotropy in its synaptic characteristics directly resulting from its low crystal symmetry. The charge transfer between the 2-nm native oxide of BP and the BP channel is utilized to achieve the synaptic behavior. Key features of biological synapses such as long-term plasticity with heterogeneity, including long-term potentiation/depression and spike-timing-dependent plasticity, are mimicked. With the anisotropic BP synaptic devices, we also realize a simple compact heterogeneous axon-multi-synapses network. This demonstration represents an important step towards introducing intrinsic heterogeneity to artificial neuromorphic systems.

• The paper introduces a novel BP-based synaptic device that leverages black phosphorus anisotropy to mimic heterogeneous biological synapses.
• It employs a POx/BP heterostructure to achieve synaptic functionalities like long-term potentiation, depression, and spike-timing-dependent plasticity.
• Results reveal orientation-dependent carrier mobility and ionization activation energies of ~0.16 eV, underscoring robust neuromorphic performance.

Anisotropic Synaptic Devices Using Black Phosphorus for Neuromorphic Applications

The paper introduces an innovative approach to synaptic electronics by employing black phosphorus (BP)-based devices, revealing a significant step toward mimicking the inherently heterogeneous nature of biological synapses in artificial neural networks. By leveraging the intrinsic anisotropic electronic properties of BP, this research outlines the development of synaptic devices specifically designed to harness the low crystal symmetry inherent in BP, thereby capturing essential features characteristic of biological systems.

Key Highlights

This research demonstrates that BP can be used as an effective channel material in synaptic devices due to its anisotropic characteristics, resulting in heterogeneity in synaptic behavior. The primary focus is on achieving synaptic functionalities such as long-term potentiation and depression, as well as spike-timing-dependent plasticity (STDP), features crucial for emulating cognitive functionalities seen in the brain. The BP synaptic devices feature a PO_x/BP heterostructure, which leverages the trapping and releasing of electrons in the phosphorus oxide to achieve synaptic characteristics reflecting variations in synaptic weights typical of biological synapses.

One of the notable strengths of the BP devices is their variable response dependent on crystal orientation. Devices oriented along the x- (armchair) and y- (zigzag) directions demonstrate different synaptic weight changes, attributed to mobility variations in the crystal structure, with the x-direction exhibiting higher carrier mobility. This anisotropy is used to emulate variable synaptic strengths, providing a robust model that closely resembles the diversity in connection strengths found in the human brain.

Detailed Observations

The BP synaptic devices demonstrate both positive and negative synaptic responses to applied electric pulses, with variations dependent on crystal orientation. The differential response has been extensively characterized through temperature-dependent measurements, revealing ionization activation energies of 0.16 eV, pointing to the nuanced control these devices can achieve in mimicking biological synaptic plasticity.

An integrated axon-multi-synapse network built from these BP devices exhibits heterogeneous synaptic responses, a model that efficiently captures the complex connectivity seen in neural architectures. The structure promotes a simultaneous yet anisotropic response across multiple synaptic devices, capable of emulating simultaneous excitatory and inhibitory synaptic activities.

Implications and Future Directions

The findings from this paper underscore the potential of BP-based devices in neuromorphic computing applications. The ability to tune synaptic responses based on intrinsic material properties opens new avenues for creating compact, efficient, and complex neural networks, essential for advancing brain-inspired computing technologies. The anisotropic properties of BP provide an innovative methodology for introducing variants to the homogeneous approaches currently privileging resistive RAM (RRAM) and field-effect transistor (FET) technologies.

Future developments could explore integrating these devices into larger-scale networks, potentially expanding their application to more complex machine learning tasks. Moreover, the long-term stability and reproducibility of these synaptic weights over different environmental conditions and extended periods remain critical for practical deployment, necessitating further experimental validation.

In conclusion, the paper represents a vital contribution to neural engineering by introducing BP as an effective material for constructing anisotropic and heterogeneous synaptic devices, bridging gaps between artificial and biological synaptic architectures in terms of functionality and dynamics.