Priyo Heru Adiwibowo, Muhammad Zohri, Lalu Ahmad Didik Meiliyadi, Muslizainun Mustapha
The reliability of photovoltaic (PV) modules is a critical concern for the sustainable deployment of solar energy, as degradation and failure phenomena significantly affect their performance and lifespan. This study aims to review failure modes by component and identify future research directions using a bibliometric approach. A comprehensive literature search and keyword co-occurrence analysis were conducted to map dominant degradation mechanisms, environmental triggers, and emerging trends in reliability research. The results indicate that glass breakage, backsheet cracking and delamination, solar cell fractures, and junction box or diode failures are the most prevalent failure modes, often accelerated by thermal cycling, moisture ingress, and electrical stress. Material interactions, such as encapsulant degradation producing corrosive by-products, further exacerbate reliability issues. Bibliometric visualization reveals three major research clusters: material aging and durability, system optimization, and advanced diagnostic techniques, with machine learning (ML) and recycling emerging as trending topics. These findings highlight a shift toward predictive maintenance, AI-driven fault detection, and sustainable end-of-life (EOL) management. This review underscores the need for integrated multi-stress testing, improved material formulations, and intelligent monitoring frameworks to enhance PV module resilience. By consolidating current knowledge and identifying research gaps, the study provides a roadmap for developing durable, cost-effective, and environmentally responsible PV technologies. ©2026 The authors.
Department of Mechanical Engineering, Universitas Negeri Surabaya, Kampus Unesa 1, Surabaya, 60231, Indonesia; Department of Physics Education, Universitas Islam Negeri Mataram, Mataram, 83116, Indonesia; Department of Applied Physics, Faculty of Science and Technology, Universiti Kebangsaan Malaysia, Bangi, 43600, Malaysia
