This paper systematically refines several key theoretical issues within the Constrained Quantum Geometry framework. First, based on the physical picture of the coherent volume acting as a momentum–energy absorber, the fusion branch selection mechanism is revised: the dominance of the ⁴He channel arises from the coherent volume carrying away recoil momentum and energy through collective excitations, thereby eliminating the quantum cost of γ-photon emission; when the coherent volume is insufficient, the system is forced to select particle channels to naturally satisfy conservation laws. Second, the confinement volume in the geometric probability is corrected from a cube to a sphere with diameter equal to the confinement scale L, yielding Pgeom = 8rN³/L³, which brings the theoretical value for carbon‑based systems closer to experimental inference. Third, the progressive activation nature of the refresh rate is clarified: the effective refresh rate starts from the ground‑state level and increases with the number of activated LCUs to a saturation value, which is the result of a dynamic balance between coherence establishment and decoherence rates; whether its upper limit can be raised requires experimental verification. Fourth, the concept of temperature in the local thermal fusion branch is abandoned and replaced with a pure energy description: the 3–4 MeV energy released via the particle channels gives neighboring particles keV‑level kinetic energy, thereby triggering secondary fusion via the Motion‑Penetration mechanism, coexisting with the Existence‑Refresh branch in low‑coherence regions. Under the refined theory, without any adjustable parameters, the single‑LCU fusion rate for carbon‑based systems is 1. 54×10⁻⁴ s⁻¹ and for metal‑based systems is 1. 92×10⁻⁵ s⁻¹, in excellent agreement with experimental median values (1. 7×10⁻⁴ and 2. 1×10⁻⁵ s⁻¹).
锦义 林 (Tue,) studied this question.