Abstract

In this work, we present a nanoscale solid state structure, which is a 3D-array of tunnel-coupled arsenic dopants in silicon with a system of metallic electrodes leading to them. The structures of eight metal electrodes were fabricated on the inhomogeneously in depth doped with arsenic silicon surface, four of which converge to a region 50 nm in diameter, and four to a region of 200 nm. After removal of a thin highly conducting upper silicon layer, single-electron transport in an array (reservoir) of arsenic impurity atoms located between the electrodes is demonstrated. The Coulomb blockade was \({\sim}100\) mV at a temperature of 4.2 K. The proposed structure can be used as a reservoir neural network, where single impurity atoms act as neurons, and electrodes will act as input and output terminals of the device, and also be used to configure the neural network. The operating temperature of such devices can be significantly increased due to the relatively small effective size of impurity arsenic atoms in silicon (3–5 nm).

中文翻译：

---

基于硅中掺杂原子无序系统的用于实现固态量子器件的多电极系统

摘要

在这项工作中，我们提出了一种纳米级固态结构，该结构是硅中隧道耦合砷掺杂剂的3D阵列，并带有通向它们的金属电极系统。八个金属电极的结构是在不均匀深度掺杂砷硅表面上制造的，其中四个会聚到直径50 nm的区域，另外四个会聚到200 nm的区域。在去除薄的高导电性上硅层之后，展示了位于电极之间的砷杂质原子的阵列（储层）中的单电子传输。库仑封锁为\（{\ sim} 100 \）mV在4.2 K的温度下。拟议的结构可用作储层神经网络，其中单个杂质原子充当神经元，电极将充当设备的输入和输出端子，也可用于配置神经网络。由于硅中杂质砷原子的有效尺寸相对较小（3-5 nm），因此此类设备的工作温度会大大提高。