Fibrillar networks formed during solution processing play a key role in the high performance of modern organic solar cells (OSCs). Yet their formation is intrinsically non-equilibrium and largely stochastic, leading to discontinuous domains and mismatched donor/acceptor interfaces that limit charge transport. Here, we show that the evolution of fibrillar networks can be guided through molecular design. By designing a cycloalkoxy-functionalized acceptor, O6R-4F, we create a molecular navigator that preferentially localizes at donor/acceptor interfaces, suppresses excessive self-aggregation of L8-BO-C5 acceptor, and promotes coordinated donor/acceptor aggregation and crystallization during film formation. This results in finer, more interconnected fibrillar networks with improved phase separation. As-cast D18: L8-BO-C5: O6R-4F devices achieve a power conversion efficiency of 20.9% without post-deposition treatment. The same approach also improves morphology and performance in chemically distinct donor/acceptor systems, demonstrating that controlling intermolecular interactions provides a general strategy for directing fibrillar network formation in solution-processed organic semiconductors. Controlling the formation of fibrillar networks during solution processing is important for organic solar cells. Lai et al. design a cycloalkoxy-functionalized acceptor to guide fibrillar network formation, achieving a power conversion efficiency of 20.9% without any post-deposition treatment.
We investigate the thermodynamic performance of a quantum Otto heat engine modulated by stochastic resetting, utilizing an exact discrete-time collision framework of a qubit coupled to a hierarchical non-Markovian environment. By evaluating the exact ensemble-averaged dynamics, we demonstrate that state-selective resetting functions as a targeted ‘non-Markovian eraser’ that repeatedly severs instantaneous system–memory correlations. Consequently, resetting dramatically dampens persistent, memory-induced heat-flux oscillations in strong-memory regimes, whereas its impact on effectively Markovian relaxation remains virtually negligible owing to the intrinsically weak environmental memory effects in this regime. By implementing an optimal stroke-dependent resetting protocol, the work extraction is monotonically maximized, displaying an approximately linear scaling in the effectively Markovian regime and a pronounced nonlinear boost under strong memory effects. Crucially, the effective efficiency remains close to the conventional Otto benchmark within numerical accuracy under effectively Markovian dynamics, whereas in the non-Markovian regime it increases monotonically with the resetting rate by actively erasing parasitic, memory-induced heat absorption. This work establishes state-selective resetting as a potent control strategy for managing environmental memory and optimizing nonequilibrium quantum thermal machines in complex open systems.
