| The permeation of organic solvents through polymeric chemical protective clothing (CPC) depends on the equilibrium solubility, S, and diffusion coefficient, D, of the solvent in the CPC material. Predictions of the breakthrough time (BT) and steady-state permeation rate (SSPR), therefore, depend on obtaining accurate estimates of S and D. This study describes the development and testing of models of S and D, and their use in predicting BT and SSPR assuming Fickian diffusion. Immersion tests and permeation tests were conducted for 53 solvents with samples of butyl, natural, neogrene, and nitrile rubber gloves. Two S models were investigated: one based on Hansen's three-dimensional solubility parameters (3DSP) used in the context of Flory-Rehner (FR) theory, and the other based on solvatochromic parameters (SCP) and linear solvation energy relationships (LSER). Improved methods for the determination of polymer 3DSPs were developed. Models based on 3DSP-FR and on the SCP-LSER methods provided solubility estimates within a factor of two of experimental values in 76% and 68% of the cases, respectively. Regression models of D based on empirical correlations with solvent properties such as molecular weight, molar volume, viscosity, and the solvent-CPC Flory interaction parameter, yielded estimates of D consistently within 50% of the experimentally derived values. Estimates of BT and SSPR were determined by using the modeled S and D values in Fickian diffusion equations. Most modeled BT values were within 5 min of experimental values for BTs {dollar}<{dollar} 20 min, within 10 min for BTs ranging from 20-60 minutes, and within a factor of l.5 for BTs {dollar}>{dollar} 1 hr. Modeled SSPR values were within a factor of two and of three of experimental values for 62% and 77% of cases, respectively. These models provide sufficient accuracy for rationally selecting appropriate CPC when experimental performance data are not available. Initial work toward modeling BT and SSPR for binary mixtures entailed modeling component solubilities for 10 binary mixtures at three relative concentrations with each glove material. The 3DSP-FR based model and a second model based on Flory-Huggins theory provided estimates of S for each component within a factor of two of experimental in 61% and 68% of the tests, respectively. |
