Photoelectrochemical (PEC) cells offer a direct route to solar-driven fuels and chemicals, yet practical devices still operate far below their thermodynamic limits owing to coupled losses spanning light absorption, carrier recombination and transport, interfacial charge transfer, and semiconductor–electrolyte matching. These losses are intrinsically entangled, and a general framework for quantifying and comparing these losses across real devices remains lacking, leaving PEC optimization largely empirical.

Here we introduce, to our knowledge, the first unified analytical model that treats both built-in junction (BIJ) and semiconductor–electrolyte junction (SEJ) photoelectrodes within a single, physically meaningful set of parameters while retaining the distinct interfacial physics of each architecture. By fitting experimental current–voltage data, the model decomposes device performance into thermodynamic, optical, recombination/transport, interfacial charge-transfer, and parasitic loss channels, and maps each directly onto concrete optimization strategies such as surface passivation, co-catalyst integration, contact optimization, and nanostructuring. Energy flows are visualized through Sankey diagrams, giving an intuitive picture of how incident solar energy is absorbed, dissipated, or converted into chemical output.

We validate the model against state-of-the-art devices spanning solar water splitting, CO₂ reduction, NH₃ synthesis, and solar redox flow batteries, achieving high-quality fits and physically consistent loss attributions. The analysis reveals a systematic divergence between material classes: photovoltaic-grade absorbers such as Si and perovskite in BIJ devices are dominated by bulk recombination and transport losses, whereas non-photovoltaic materials such as BiVO₄ and Ta₃N₅ in SEJ devices are limited primarily by interfacial reaction kinetics. We further propose a semiconductor–electrolyte matching criterion beyond simple energy-level alignment and construct efficiency maps that define material-selection windows, identifying a moderate reaction-potential window of approximately 0.7–1.1 V as a universal operating regime for high-efficiency PEC systems.

Together, these capabilities help researchers see where energy is lost and turn that insight directly into better device designs, supporting a shift from empirical optimization toward mechanism-informed rational design of viable solar fuels technologies. The model is accompanied by an open, ready-to-use toolkit for efficiency-map generation, J–V fitting with loss decomposition, and interface-matching analysis.