Modeling the Temperature Dependence of N-Methylpyrrolidone Permeation Through Butyl- and Natural-Rubber Gloves

Abstract

This paper describes the temperature dependence of N-methylpyrrolidone (NMP) permeation through gloves used in microelectronics fabrication facilities. One type of butyl-rubber glove (North B161), two types of natural-rubber gloves (Edmont Puretek® and Ansell Pacific White®), and a natural rubber/nitrile/neoprene-blend glove (Pioneer Trionic®) were tested at four temperatures from 25–50 °C using the ASTM F739-85 permeation test method. The butyl-rubber glove showed no breakthrough after four hours of exposure at any temperature. The variations with temperature of measured breakthrough times (BT) and steady-state permeation rates (SSPR) for the other gloves were described well by Arrhenius relationships, with BT values decreasing by factors of 7–10 and SSPR values increasing by factors of 4–6 over the temperature range studied. Extrapolation to 70 and 93°C, the temperatures at which degreasing is often performed, yielded BT values of <2 min and <0.5 min, respectively, in all cases. With the exception of the butyl-rubber glove, following an initial exposure at 25°C and air drying overnight, low levels of NMP vapor were detected off-gassing from the inner surfaces of the gloves. Experimental results were then compared to those expected from several permeation models. Estimates of the equilibrium solvent solubility, S, were calculated using a model based on three-dimensional solubility parameters. Estimates of the solvent diffusion coefficient, D, were obtained from correlations with either the solvent kinematic viscosity or the product of the Flory interaction parameter, χ, and the solvent molar volume. Combining these values of D and S in Fickian diffusion equations gave modeled BT estimates that were within 23% of experimental values over the temperature range examined. Modeled SSPR values were within 50% (typically within 25%) of experimental values. Another model based on a generalized Arrhenius relationship also provided useful but generally less accurate estimates of the changes in BT and SSPR values with temperature.