On-Chip Integration of Micro-supercapacitor in VLSI Design for Power Management in Artificial Intelligence Processors and Memory Chips: A Review of Methods and Materials

Abstract

Supercapacitors can act as an instant energy source to quickly supply electric power to any connected system because they are energy storage devices with high power densities. This feature of the supercapacitor can be used to reduce power fluctuations and access latency in cache memories in the VLSI (very-large-scale integration) design of quickly operating artificial intelligence (AI) processors and memory chips. High-performance computing with the best data throughput is produced by this high-power stability and instantaneous power delivery. This article reviews various thin film micro-supercapacitor integration techniques, fabrication processes, and electrode and electrolyte materials, applicable for on-chip integration of micro-supercapacitors on AI processors and memory chips. This review extends the discussion of the thin film versions of supercapacitors in planar and vertically stacked configurations and their relative advantages in mediating ionic conduction. On-chip fabrication of micro-supercapacitor by laser micropatterning, laser surface roughening, carbonization by pyrolysis, laser-induced reduction, and carbon MEMS to scribe electrode patterns with different materials are reviewed. For the on-chip integration of micro-supercapacitors, this examines various fabrication techniques including photolithography, such as monolithic integration, heterogeneous integration, and 3D stacking. The synthesis and implementation of various carbon-based, transition metal oxide-based, and conducting polymer-based electrode materials in both their pure and composite forms are reviewed.