Atomic-Scale Electronics

点击翻译按钮获取中文摘要

The progression of CMOS transistors and technology is expected to reach 5nm gate length by the year 2030. At these dimensions and beyond, many of the properties of bulk materials, including the bandgap, the energy band diagrams, and carrier mobilities, undergo dramatic changes and tunneling dominates. Coincidentally, room-temperature tunneling-based coupled quantum-dot (QD) devices, envisioned for quantum computing (QC), will also become feasible at these dimensions. Although integration density has continued to follow Moore's law due to feature size shrinking and the introduction of 3D transistor structures such as the FinFET, MOSFETs are fast approaching their scaling limits. This will affect the future evolution of both computing and wireless communication, the pillars of today's revolution in data science, machine learning and the internet of things, which rely on ever-increasing computational power and data rates, and on ambient sensing. To continue computing speed and functionality improvement beyond the end of CMOS, it is critical that a variety of channel materials and qubit structures suitable for integration in a production silicon platform be thoroughly investigated for large volume and low-cost, scalable QC qubits and integrated circuits. This proposal addresses (i) the continued scaling of computational power by exploring novel atomic-scale quantum-computing (QC) hardware in CMOS foundry processes, which can be integrated on the same die with classical CMOS logic and millimeter-wave electronics, and (ii) transceiver architectures exploiting the un-chartered 140-300 GHz frequency band for qubit spin control/readout, ambient sensing, instrumentation, and wireless communications at over 100 Gb/s. The proposed qubit ICs consist of coupled-QD electron and hole spin qubits, placed in the atomic-scale channel of multi-gate n- and p-MOSFETs, and of 60-240GHz spin control and spin readout circuits integrated on the same die. As a radical breakthrough, the fabricated qubits will feature mode energy level splitting on the order of 0.25-1 meV corresponding to Rabi frequencies in the 60-240GHz range, suitable for operation at 312 degrees Kelvin, two orders of magnitude higher than today's qubits. The tuned mm-wave circuits allow for 10-20ps spin control pulses which help to filter out wideband thermal noise and make the proposed qubits and QC logic more tolerant of short spin coherence and relaxation times. Thermal noise filtering may lead to even higher temperature operation for a given energy-level splitting.