No abstract

No abstract

To gain perspective and insight into the performance of permeable reactive barriers containing granular iron metal, it is useful to compare the degradation kinetics of individual chlorinated solvents over a range of operating conditions. Pseudo first‐order disappearance rate constants normalized to iron surface area concentration (kSA) recently have been reported for this purpose. This paper presents the results of further exploratory data analysis showing the extent to which variation in kSA is due to initial halocarbon concentration, iron type, and other factors. To aid in preliminary design calculations, representative values of kSA and a reactive transport model have been used to calculate the minimum barrier width needed for different ground water flow velocities and degrees of halocarbon conversion. Complete dechlorination of all degradation intermediates requires a wider treatment zone, but the effect is not simply additive because degradation occurs by sequential and parallel reaction pathways.

[1] Numerical models were used to evaluate how aquifer and barrier heterogeneity affect influent and effluent concentrations for permeable reactive barriers (PRBs). Spatial variability in the reaction rate constant k r and hydraulic conductivity K P of the PRB and hydraulic conductivity of the aquifer (in terms of variations in the mean m lnK , standard deviation s lnK , and correlation scale l of the logarithm of hydraulic conductivity (ln K)) were considered. Spatial variability of k r and K P was found to change influent and effluent concentrations by less than an order of magnitude. Spatial continuity in hydraulic conductivity parallel to flow, described by the correlation length l x , has a modest effect, with greater continuity yielding higher effluent concentrations. Decreasing the hydraulic conductivity of the aquifer (i.e., decreasing m lnK ) does not affect influent concentrations but decreases the median effluent concentrations and broadens the distribution of effluent concentration due to increased residence time in the PRB. Increasing the variability of the aquifer hydraulic conductivity (i.e., increasing s lnK ) decreases the median influent concentration and broadens the distribution of influent concentrations due to additional dispersion caused by greater heterogeneity. Greater variability in aquifer hydraulic conductivity also results in a higher median effluent concentrations and a broader distribution of effluent concentration. Comparison of effluent concentrations predicted using a one-dimensional deterministic model and the threedimensional numerical model shows that longer residence times and lower effluent concentrations are usually predicted by the one-dimensional model. The results indicate that designers should carefully consider factors of safety used for design, perhaps opting for a more conservative approach until more guidance on designing amidst aquifer heterogeneity is available.