ConspectusSelf-assembly is a central paradigm in chemistry, enabling complex molecular architectures to emerge from simple components. Traditionally, self-assembly processes have been classified as thermodynamically controlled─yielding the most stable product at equilibrium─or kinetically controlled, where outcomes are dictated by individual rate constants associated with irreversible or effectively irreversible steps. However, many contemporary self-assembled systems, particularly coordination-driven assemblies and dynamic supramolecular architectures, are constructed entirely from reversible elementary reactions. In such systems, it remains unclear how robust pathway selection and high-yield product formation can emerge without invoking intrinsic irreversibility.In this Account, we introduce network-controlled self-assembly, a framework in which assembly outcomes are governed not by isolated rate constants or static pathways but by the temporal organization of reaction flux within a fixed, fully reversible reaction network. We show that even when all elementary reactions are reversible, time-dependent redistribution of flux─termed temporal flux organization─can generate quasi-irreversible behavior at the level of pathways and products, thereby enabling selective and high-yield self-assembly.This framework is established through quantitative experimental analysis and numerical modeling of metal–organic cage (MOC) formation, combining QASAP (Quantitative Analysis of Self-Assembly Process) and NASAP (Numerical Analysis of Self-Assembly Process). QASAP reconstructs the hidden ensemble of intermediates from experimental time-series data, while NASAP translates this information into explicit reaction networks, quantitative rate constants, and time-resolved analyses of pathways and fluxes. In a first case study of a M6L4 truncated tetrahedron, we analyze pathway competition between triangular (M3L3) and square (M4L4) intermediates within a fully reversible coordination network. Flux analysis reveals that, as assembly proceeds, reaction flux becomes progressively organized in time such that one pathway dominates, resulting in quasi-irreversible pathway selection without any irreversible elementary steps.In a second case study of an M6L4 square-based pyramid, we demonstrate network-controlled yield amplification through catalytic modulation. Using an identical network topology, a catalyst (ReO4–) selectively accelerates late-stage elementary reactions, reorganizing temporal flux, suppressing intermediate accumulation, and generating a strong forward flux toward the final product. As a result, near-quantitative product formation emerges at the network level despite complete molecular reversibility.This Account synthesizes a progression from quantitative kinetic analysis (QASAP), through explicit reaction-network modeling (NASAP), to a network-level understanding of quasi-irreversibility, culminating in temporal flux organization as a design principle for reversible self-assembly. These studies establish temporal flux organization as a unifying principle that links microscopic kinetics to macroscopic self-assembly outcomes in fully reversible reaction networks. By reframing self-assembly as a problem of time-resolved flux flow rather than static pathways or isolated rate constants, network-controlled self-assembly elevates reaction flux from a descriptive quantity to a design variable, enabling predictive and rational control of pathways and yields across coordination assemblies, supramolecular polymers, and other dynamic reaction networks.
Shûichi Hiraoka (Mon,) studied this question.