CAREER: On-Chip Terahertz Electronic Frequency Combs

保持电子信号生成、传感和处理的指数级增长，对于迎接即将到来的医疗保健、环境监测、自动驾驶汽车/机器等无处不在的传感器时代的新挑战至关重要。特别是，如果低成本电子设备的工作频率可以扩展到太赫兹（THz）范围，前所未有的宽带将使许多新的应用成为可能，如非电离成像、分子识别、高分辨率雷达和超高速数据链路。虽然在过去的十年中，硅太赫兹集成电路在输出功率和效率方面有了显著的提高，但它们对高质量因数共振的依赖使得当前的太赫兹传感器窄带，无法充分利用可用的广谱进行宽范围气体传感和高精度雷达测距。为了突破这一限制，本课题研究了一种基于电子太赫兹频率梳的新技术。通过集成电路芯片上的并行信号处理，采用该技术的气体传感器和成像雷达将获得更好的频谱覆盖和能量效率。这项研究工作也将与首席研究员的教育事业目标相结合，通过创建新课程和通过麻省理工学院的外展项目吸引代表性不足的学生和K-12学生，促进高度跨学科的研究。本提案的目标是利用硅半导体制造工艺的集成能力，并使用太赫兹频率梳技术将信号传感负载分配到一组窄带，精确控制的太赫兹电路单元。它保持了高能效，同时以可扩展的方式覆盖了宽带宽。基于两种太赫兹频率梳的传感器将在该计划下进行研究。首先，旋转模式分子传感器将演示使用均匀分布的频率梳。频率梳将无缝覆盖超过100ghz的带宽，并将芯片级太赫兹光谱仪的频谱扫描速度提高200倍。通过片上频率校准技术，光谱仪将以十亿分之一级别的频率精度扫描频谱。其次，将演示使用非均匀梳状和压缩感知的成像雷达。利用一组非线性分布波长对目标距离进行量化，并采用误差校正算法，使雷达具有低信噪比和低功耗的特点。太赫兹梳状雷达还将集成准光学配置的低成本传感器阵列，从而实现具有电子扫描和精细分辨率的实时3D成像。通过这些综合研究，该方案将建立并行在芯片级宽带信号传感和处理中的优势。

Maintaining the exponential growth of electronic signal generation, sensing, and processing is essential to meet the new challenges of the upcoming era featuring ubiquitous sensors for healthcare, environment monitoring, autonomous vehicles/machines, etc. In particular, if the operating frequency of low-cost electronics can be extended into the terahertz (THz) regime, the unprecedented wide bandwidth will enable numerous new applications, such as non-ionizing imaging, molecular identification, high-resolution radar, and ultra-high-speed data link. Although in the past decade THz integrated circuits in silicon have improved remarkably in output power and efficiency, their reliance on high-quality-factor resonance makes the current THz sensors narrowband and fails to fully capitalize on the available broad spectrum for wide-range gas sensing and high-precision radar ranging. To break such a limit, this project investigates a new technique based on electronic THz frequency comb. Through parallel signal processing on the integrated circuit chip, gas sensor and imaging radar using the proposed technique will achieve significantly better spectral coverage and energy efficiency. This research effort will also be integrated with the principal investigator's educational career goal of promoting highly-interdisciplinary studies through the creation of new courses and engaging underrepresented and K-12 students through MIT's outreach programs.The objective of this proposal is to leverage the integration capability of the silicon semiconductor fabrication process and use THz frequency comb technique to distribute the signal-sensing load to an array of narrowband, precisely-controlled THz circuit units. It maintains high energy efficiency while covering a wide bandwidth in a scalable fashion. Sensors based on two types of THz frequency comb will be investigated under this program. First, a rotational-mode molecular sensor will be demonstrated using an evenly distributed frequency comb. The frequency comb will seamlessly cover more than 100 GHz of bandwidth and increase the spectral scanning speed of chip-scale THz spectrometer by a factor of 200. Through an on-chip frequency calibration technique, the spectrometer will scan the frequency spectrum with one-part-per-billion level frequency precision. Second, an imaging radar will be demonstrated using a non-uniform comb and compressive sensing. By quantizing the target distance with a set of nonlinearly distributed wavelengths and by using an error-correction algorithm, the radar can operate with low signal-to-noise ratio and low power consumption. The THz comb radar will also be integrated with a low-cost sensor array in a quasi-optical configuration, so that real-time 3D imaging with electronic scanning and fine resolution can be realized. Through these comprehensive studies, the program will establish the advantages of parallelism in chip-scale wide-band signal sensing and processing.