Unified Multilayer Diffusion Model and Application to Diffusion Experiment in Porous Media by Method of Chambers

Abstract

Diffusion coefficient is an important parameter for examining contaminant transport in the environment. Chamber methods (with or without external mixing devices) are the most popular methods for measuring effective diffusion coefficients in porous media (Deff) through air or water. The objectives of this paper were to apply simplified and unified analytical methods for both perfectly mixed and nonmixed (one- or two-) chamber systems and to examine how mixing affects the estimation of Deff. An analytical solution for a multilayer transient diffusion model was applied to the chamber methods without external mixing. By increasing the diffusion coefficient in reservoirs (D1 and D3) more than 10 times from the value for air or water (D0), the model was sufficient to approximate the well-mixed condition and, consequently, can be used to model transient diffusion in chamber systems with external mixing devices. We demonstrated that at long time Deff was related to the first eigenvalue (beta1) of a transcendental equation, which provided a quick method for determining Deff accurately from experimental data. The error caused by using the well-mixed approximation can be significant for a single-chamber system when there are no external mixing devices. This error increased rapidly with decreases in the experimental duration. A good fit for the concentration versus time curve could not be obtained forthe well-mixed solution, especially when sampling ports were near the boundary (x=0) and interface (x = l1). The proposed solutions are useful when the reservoir or chamber methods are used for measuring Deff and have wide applications in predicting contaminates transport in porous media and groundwater.