Fixed-Control Accessibility-Loss Events in Open-System Decoherence: Detection, Window Scaling, and Benchmark Subclasses

We identify a regime of open quantum system dynamics in which the observed coherence time exhibits a discontinuous transition arising from selective loss of accessibility of an effective decoherence channel. Within the Gaussian influence-functional formalism, normalization of the Keldysh weight matrix leads to a parameter-independent coherence-time ratio R ≡ τₚost/τₚre = 4/3. The result follows entirely from standard open-system theory and introduces no modification of quantum mechanics. The predicted transition differs fundamentally from known dynamical-decoupling fixed points, occurring in time at fixed control parameters and remaining independent of pulse number. Because the framework yields explicit falsification criteria, the predicted ratio provides a direct experimental probe of multi-channel decoherence structure. Our analysis shows that this value is structurally fixed by the normalized geometry of the Keldysh representation rather than by microscopic fine-tuning, recurring across related mechanisms (e. g. , the DD fixed point of Paper A and the companion analysis in Paper C). The central result is derived from a three-channel decoherence model with normalized weights (w₁, w₂, w₃) = (1, 1/2, 1/2), arising from the Keldysh rotation of the Feynman–Vernon influence functional in the Gaussian vacuum-dominated limit. When Channel 3 loses dynamical accessibility at transition time t*, the total effective decoherence weight drops from Wₚre = 2 to Wₚost = 3/2, producing the ratio R = 4/3 without retuning any external control parameter. The predicted transition is distinguished from dynamical-decoupling fixed points by three signatures: (i) a temporal discontinuity at fixed control parameters rather than a plateau across a pulse-number window, (ii) independence of the ratio from pulse number N, and (iii) absence of global filter-function reshaping in the noise spectrum. Explicit falsification criteria (F1–F5) are provided. A concrete experimental realization using NV-center magnetometry is described, including channel identification, readout timescales, and control experiments. This is Paper B in a series. Paper A (Shioiri, Zenodo 2026) identifies a related R = 4/3 fixed point arising from pulse-sequence symmetry in dynamical-decoupling noise spectroscopy. Both papers share the same Keldysh normalization structure Mₙorm = diag (1, 1/2, 1/2) but represent distinct physical mechanisms. V2: This version includes a strengthened operational toy-model section, a more explicit experimental protocol, and an expanded Appendix clarifying the origin of the normalized Keldysh weight structure. V3: This version strengthens the presentation of the decoherence-transition framework by clarifying its status as a normalized effective-channel description rather than a universal theorem of arbitrary open-system dynamics. The assumptions behind additivity, channel accessibility, and the vacuum-inspired Keldysh normalization are now stated more explicitly, and the discussion of NV-center realizations has been revised to distinguish the strict vacuum-dominated limit from the experimentally relevant effective Gaussian regime with thermal corrections. The operational binary-switch picture has also been sharpened, including the role of finite switching-time smearing and the effective nature of channel independence. In addition, the appendix has been refined so that the equal-weight assignment (1, 12, 12) (1, 12, 12) (1, 21, 21) is presented as a symmetry-motivated normalized input under isotropy and minimal rank, rather than as a microscopic uniqueness statement. These revisions do not alter the central result: within the present normalized Gaussian effective model, selective loss of one decoherence contribution produces a discontinuous coherence-time jump with the structurally fixed ratio R=4/3R = 4/3R=4/3. Version 4 refines the operational formulation of fixed-control accessibility-loss events in open-system decoherence. It clarifies f (Δ) f () f (Δ) as the primary observable, keeps RRR as a derived quantity only in the common single-rate regime, tightens the finite-window averaging argument, and makes the benchmark subclass assumptions explicit. Figures and notation have also been corrected and standardized.