Search for a command to run...
The shift to sixth-generation (6G) systems may profoundly transform connectivity beyond wireless. This transition will facilitate the interconnection of the entire globe to a cyber-physical continuum [9]. To achieve this vision, which includes holographic communication in real-time, a tactile internet and widespread autonomous intelligence, unprecedented performance metrics will be needed, e.g., Tbps data rates, less than 100 μs deterministic latencies and high-density connection of devices [1], [9]. To achieve these targets, migration of sub-terahertz (sub-THz) and THz frequency bands (0.1–10 THz) is needed [1]. But traditional silicon-based Very Large-Scale Integration (VLSI), which is the core feature of the modern digital world, has reached a performance limit [4]. This paper argues that the synergistic co-integration of computational VLSI and high performance photonic devices is the critical solution required or heterogeneous integration. Such is the subject of the present paper, which explores the differences between the computational and photovoltaic properties of silicon and the III-V compound semiconductors [5], [6]. A detailed classification, or taxonomy, of integration methods is presented covering the entire range of methods, from monolithic epitaxial growth to very advanced techniques such as wafer bonding and micro-transfer printing techniques to 2.5D/3D System-in-Package (SiP) architectures [7], [14]. The use of photonic assisted THz transceivers, reconfigurable optical beamforming networks and integrated sensing and communication (ISAC) modules hints at the potential of these platforms [2], [13]. The discussion goes below the surface of some major challenges related to multiphysics issues such as thermal cross talk, power delivery integrity, high frequency signal fidelity at electro-optical interfaces and the urgent need for electronic-photonic design automation (EPDA) toolchains.
DOI: 10.1117/12.3108506