Abstract: With the sustained growth of global photovoltaic (PV) installed capacity,early-service modules are gradually entering their end-of-life phase,making the resource recovery of retired crystalline silicon PV modules an urgent engineering challenge in the circular economy.PV glass and silicon materials constitute two core high-value components;However,current mainstream recycling processes predominantly resort to downcycling both,resulting in a massive waste of high-purity silicon and premium quartz resources.On the basis of delineating high-value utilization pathways for recovered silicon and PV glass,this paper systematically investigates the impurity removal mechanisms of various wet chemical systems,with a focus on the challen ges of interfacial metal impurity adhesion and the transformation of their occurrence states induced by module dismantling and pretreatment.The performance differences among four types of systems—traditional inorganic strong acids,alkaline etching,or ganic acid complexation,and novel deep eutectic solvents (DES)—are compared in terms of metal removal efficiency,etching selectivity,and their capacity to preserve the intrinsic structures of the silicon lattice and the silica network.The results show that although traditional strong acid systems deliver high impurity removal efficiency,they non-selectively damage both the sili con substrate and the glass matrix while posing significant environmental risks. In contrast,DES,serving as a green and mild medium,exhibit high selectivity,low corrosivity,and excellent recyclability,enabling synergistic and non-destructive purifica tion of silicon and glass,and thus represent the most promising technological direction for future development.This paper chartsa pathway for transforming the PV recycling paradigm from simple precious metal extraction toward the synergistic high-purity regeneration of silicon and glass,providing a theoretical basis and engineering reference for the low-carbon,high-quality recy cling of retired photovoltaic materials.