Department of Materials Science and Engineering, Pohang University of Science and Technology (POSTECH), Pohang 37673, Republic of Korea.
ACS Appl Mater Interfaces. 2020 Jun 17;12(24):27467-27475. doi: 10.1021/acsami.0c01775. Epub 2020 Jun 2.
The aim of the neuromorphic computing is to emulate energy-efficient and smart data-processing ability of the biological brain, which is achieved by massively interconnected neurons and synapses. The strength of a connection between two neurons is modified by homosynaptic and heterosynaptic plasticity. As current research in the neuromorphic device is mainly focused on emulating homosynaptic plasticity, complex biological functions are not easy to mimic because they require both homosynaptic and heterosynaptic plasticity. We demonstrate the use of a liquid crystal-carbon nanotube (LC-CNT) composite as a resistive switching material that can emulate both the homosynaptic and heterosynaptic functions of biological neurons. The LC-CNT composite undergoes resistance change by CNT alignment and aggregated wire formation subjected to an applied electric field. A two-terminal device that exploits this mechanism achieves analog switching and homosynaptic potentiation. In a multiterminal device structure, the modulatory interneuron could tune the synaptic properties to perform heterosynaptic functions such as heterosynaptic potentiation, heterosynaptic facilitation, and synaptic weight normalization to emulate complex biological functions of a brain. Artificial synapses that exploit this multifunctionality of the LC-CNT composite have uses in next-generation neuromorphic devices.
神经形态计算的目标是模拟生物大脑的节能和智能数据处理能力，这是通过大量相互连接的神经元和突触实现的。两个神经元之间连接的强度通过同突触和异突触可塑性来改变。由于当前神经形态器件的研究主要集中在模拟同突触可塑性上，因此复杂的生物功能不容易模拟，因为它们需要同突触和异突触可塑性。我们展示了使用液晶-碳纳米管 (LC-CNT) 复合材料作为电阻开关材料，该材料可以模拟生物神经元的同突触和异突触功能。LC-CNT 复合材料通过施加电场导致 CNT 排列和聚集的金属丝形成来实现电阻变化。利用这种机制的两端器件可实现模拟开关和同突触增强。在多端器件结构中，调节性中间神经元可以调节突触特性，以执行异突触功能，如异突触增强、异突触易化和突触权重归一化，从而模拟大脑的复杂生物功能。利用 LC-CNT 复合材料这种多功能性的人工突触可用于下一代神经形态器件。
