Charge-Unified Semiconductor Switching Theory

Semiconductors and downstream applications underpin the electronic, information, energy and industrial systems sustaining modern and future society, their sustainability is an urgent global priority, particularly with global electricity generation projected to rise more than 2.5-fold by 20501. However, since 19472, semiconductor switching has remained largely empirical or phenomenological in temporal evolution, with the physical nature of circuit elements, particularly semiconductors unclear; and fundamental unifications absent across macroscopic and microscopic insights, charge- and energy- conservation frameworks, and equivalent-circuit formalisms. These limitations fragment domains across the semiconductor value chain and constrain their sustainability. Here we present Charge-Unified Semiconductor Switching Theory (CUSST) that reveals the charge-mediated nature of circuit elements, particularly semiconductors, reveals switching inertia and the dynamical nature of switching, achieves these long-missing unifications, and establishes a unified formulation across conceptual, mechanistic, formal and analytical levels, with enhanced simplicity through unifications. The demonstrated implications include helping to generalize circuit theory and extend conservation laws, and guiding new theoretical systems such as fundamental modelling. For example, a CUSST-based switching-energy-loss model (errors: 0.88–11.60%) achieves a 17-fold average error reduction compared to the conventional model (errors: 34.41–80.05%); CUSST enables unprecedented causal-mechanistic interpretability of switching waveforms as manifestations of underlying switching dynamics. Furthermore, its prospective implications include informing domains across the semiconductor value chain, while unifying these separate domains towards a unified discipline, including through future full-chain causal feedback, future cross-domain research and co-design for optimal sustainability. CUSST may help to identify directions across disciplines such as semiconductor materials3, chip design4, packaging5, reliability6,7, thermal engineering8,9; downstream applications such as power electronics; and potentially extend to broader systems, including communication, computing and integrated circuits.10,11