Thermodynamic Eigenrate Decomposition for Drug Release from Electrospun Fibrous Matrices: Plain-Polymer, Cyclodextrin-Complexed, and Experimentally Informed Formulations

Abstract Diffusion, swelling, and erosion act on different time scales in electrospun matrices, making release hard to predict based on formulation variables. The Modified Multicomponent Interactive Release (M-MIR) model assigns each mechanism a first-order depletion rate in reciprocal seconds, so that the three become commensurable and their normalized ratios serve as mechanistic weights. These follow from Hansen solubility parameters, Flory–Huggins interaction parameters and partition coefficients, and are settled before any release datum is consulted. A Flory–Rehner criterion switches the swelling pathway off for non-swelling polymers, a mass-balance-constrained burst term completes the early-time description, and a two-scale analysis identifies the mat half-thickness, not the fiber radius, as the rate-controlling diffusion length. Three tiers cover plain matrices, cyclodextrin inclusion complexes, and systems where measured swelling and mass-loss functions replace the constitutive sub-models. The model is tested on two published systems lying on opposite sides of the swelling criterion. For naproxen in poly(ε-caprolactone) (PCL) nanofibers, plain and β-cyclodextrin (CD)-inclusion complexed (IC), an interaction parameter of 9.75 from tabulated data alone closes the gate, fixing the diffusion weight at unity with nothing adjusted; the description ranks first for the plain formulation and resolves complexation into a 65% fall in effective diffusivity opposed by a 2.45-fold rise in the fitted transport constant, which a single empirical rate cannot separate. For ciprofloxacin and rutin released together from crosslinked poly(vinyl alcohol) (PVA)/chitosan membranes, the gate is open, with the weights being swelling-dominated at 0.58 to 0.69, and the description ranks first on all three profiles. The two tests are not equivalent: the closed gate is predicted from tabulated data, whereas the open gate is established from measured swelling and corroborated against independent determinations, with the solubility-parameter route being unreliable for water-rich media. Throughout, the fitted burst fraction orders correctly with solubility and composition, and every parameter remains non-negative and mechanistically readable, which the closest classical competitor does not achieve. The framework reduces to the classical release equations as limiting cases.