E2CDA: Type II: A new non-volatile electrochemical transistor as an artificial synapse: device scaling studies

使计算能力超越摩尔定律的限制的一种方法是建立模仿大脑的计算机体系结构。事实上，大脑能够非常高效地执行某些类别的复杂计算，例如模式识别，尽管它的固有速度很慢，但使用的功率很低。这项提议探索了一种新的设备，它展示了在大脑中发现的突触的一些有用的特性，例如超低功率开关。将详细研究其工作机制，并将制造和微型化器件阵列，以提供迈向类脑集成电路的第一步。这些设备通常被称为神经形态电子学，这是一个日益引起人们兴趣的领域，因此，在该奖项中开展的研究有可能影响微电子行业的未来，并培养出一支为这一新兴领域培训的劳动力。最后，该设备的光学特性在操作时会改变颜色，这将允许开展一项将艺术和科学联系起来的推广活动，最终目标是吸引更多不同的学生群体进入电子设备材料科学领域。将研究一种基于聚合物半导体的模仿突触功能的新设备，称为ENODE(电化学神经形态有机设备)。展示了具有超低开关能量(10pJ)、亚毫伏开关电压、以非挥发方式保持500个单态的ENOD。为了优化ENODE的结构，我们将研究这种低开关能量和低电压的物理来源。此外，ENOD将由一系列不同的材料制造，利用固态技术将开关速度从1 kHz提高到MHz范围，从而实现更高水平的集成。最后，使用传统微电子学中使用的相同工艺，将制造ENODE阵列，以测试这些新器件的重复性和缩放规律。

One way to move computational power beyond the limits imposed by Moore's Law is to build computer architectures that mimic the brain. Indeed the brain is able to perform certain classes of complex computations, such as pattern recognition, very efficiently and using low power in spite of its inherent low speed. A novel device that exhibits some of the useful properties of the synapses found in the brain, such as ultra low-power switching, is explored in this proposal. Its working mechanism will be studied in detail and arrays of devices will be fabricated and miniaturized in order to provide a first step towards brain-like integrated circuits. Neuromorphic electronics, as these devices are often called, are a growing area of interest, therefore the research developed in this award has the potential to impact the future of the microelectronics industry and to produce a workforce trained for this emerging area. Finally, the optical properties of the device, which changes color upon operation, will allow to develop an outreach activity connecting art and science with the ultimate goal of attracting a more diverse student population to the area of materials science of electronic devices.A new device based on polymeric semiconductors that mimics the functions of synapses, called ENODe (electrochemical neuromorphic organic device), will be studied. ENODes with ultra-low switching energy (10 pJ), sub-mV switching voltage, retaining 500 individual states in a non-volatile fashion were demonstrated. The physical origin of such low switching energies and voltages will be investigated with the goal of optimizing the architecture of the ENODe. Furthermore, ENODes will be fabricated with an array of different materials that may enable a higher level of integration making use of solid-state technology to increase the switching speed beyond 1 kHz into the MHz range. Finally, using the same process used in traditional microelectronics, arrays of ENODes will be fabricated to test the reproducibility and the scaling laws of these new devices.