An experimental investigation of NAPL pool dissolution enhancement by flushing

Experiments coupled with mathematical modeling are used to further elucidat e the quantification of NAPL pool dissolution and its enhancement by water flushing. The experiments were performed using glass bead-packed column rea ctors with small pools of neat toluene or a toluene in dodecane mixture (to luene mole fraction, X-tol approximate to 0.02 or 0.09). Experimental quasi -steady-state toluene dissolution fluxes were determined by a mass-balance approach. The results of these experiments are used to quantify the impact of flushing on the NAPL pool dissolution flux, J, for wide ranges of two of the key controlling parameters-equilibrium concentration, C-S, and average pore water velocity, nu(x). In addition, the experimental dissolution flux data are used to evaluate predictions made using independently obtained mo del parameters and analytical mathematical models incorporating the two bas ic approaches used for mathematically describing the interphase distributio n of NAPLs: the local equilibrium (LE) approach; and the mass transfer limi ted, or nonequilibrium (NE) approach. The data from the experiments with a neat toluene pool demonstrate the expected trend of increasing J as nu(x), increased from approximately 2 to 30 m/day. The LE model and the NE model w ith an average mass transfer coefficient, k(1) = 4.76 m/day, were able to d escribe the neat toluene pool data for nu(x) < 18 m/day reasonably well. Ho wever, for nu(x) > 18 m/day, the NE model with k(1) = 4.76 m/day provides a better description of the data, suggesting that the equilibrium boundary c ondition may become invalid for very high velocities. The experiments with the binary toluene in dodecane pools also show the expected trend of increa sing J with increasing nu(x). The LE model describes the binary pool data r easonably well for the entire range of nu(x) studied (approximately 0.1 to 10 m/day), and the NE model predictions with the average k(1) = 4.76 m/day determined in the neat-pool studies do not deviate significantly from the L E model in this nu(x) range, although small deviations occur for > 10 m/day . (C) 1999 Elsevier Science B.V. All rights reserved.