Journal of Environmental Engineering

Studies were conducted to examine the effect of solution chemistry, defined here as pH and ionic strength, on the permeability of negatively charged polymeric nanofiltration membranes. Water permeation through the membrane was demonstrated to decrease at conditions of low pH and high ionic strength in the absence of organic macromolecules. The reduction in membrane permeability was attributed to a compaction of the membrane matrix resulting from charge neutralization at the membrane surface and electric double layer compression. An uncharged model organic macromolecule (polyethylene glycol) was used to quantify the effects of solution chemistry on membrane compaction and solute rejection capabilities of the charged membrane. Studies of membrane permeability and rejection were then repeated with solutions containing natural organic matter (NOM), enabling concurrent evaluation of the effects of electric double layer compression as well as changes in both membrane structure and the apparent macromolecular size of charged NOM macromolecules.

A method was developed to coat anthracite and sand filter media with synthetic manganese oxides. The two-stage method used a stirred tank reactor followed by a recirculating flow fixed-bed reactor. Treated anthracite developed encapsulating coatings of manganese oxides up to 30-μm thick, at loadings over 4 mg/g⁻¹ extractable Mn²⁺. Media were characterized by BET surface area analysis and electron microscopy. Laboratory-scale column tests confirmed the efficient Mn-removal capability of the synthetically coated media in adsorption and oxidation modes. This method may help improve the removal of Mn²⁺ from drinking water and hazardous effluents by filtration with oxide-coated media.

A process model for soluble Mn²⁺ removal by oxide-coated media filtration was developed, incorporating a modified Gnielinski mass transfer correlation for boundary layer transport, hydrodynamic dispersion, the Freundlich isotherm, and a modified Keinath correlation for intracoating surface diffusion. The available fraction of equilibrium sorption capacity was the calibration parameter used for the sorptive process mode. For the process mode of continuous regeneration by free chlorine, the sorbed Mn²⁺ oxidation rate constant was an additional calibration parameter, and the two parameters were uncorrelated. The calibrated model was predictive for laboratory column effluent profiles under dynamic influent boundary conditions of sorption followed by desorption over ranges of pH, flow rate, ionic strength, and (Ca²⁺). The model also satisfactorily accounted for oxidative Mn²⁺ removal over a range of process conditions for both synthetic and “naturally” coated media. Model structure was suggested by results of fundamental physical and chemical characterization of filter media from water treatment plants.

An integrated model for computing the height variation of the mudline and sediment-suspension interfaces of a slurry exhibiting zone settling behavior is proposed. The model accounts for the dynamic interaction of the two dominant components of a batch test, the settling of solids in the suspension region, and the compression of the material that has already been deposited at the bottom of the cylinder. The settling of solids is modeled by the Kynch theory, while the compression process is represented by a self-weight nonlinear finite strain model. A graphical procedure that involves use of the slurry flux curve in conjunction with the Kynch model is developed to calculate the flux of the solids that settle atop the sediment layer. The equation governing the compression process, which is a nonlinear transient partial differential equation of the parabolic type, is solved numerically by using the finite-element method for the spatial domain, while for the time domain a finite-difference scheme is used. Its solution provides the compaction rate and the velocity of the upward moving liquid that is squeezed out from the sediment layer. The coupling of the two components is accomplished by continuously accounting for the effect of the newly deposited sediment on the thickness of the sediment layer and the attendant release of upward moving liquid and its influence on the settling of solids within the suspension. The validity of the model is illustrated by using data reported in the literature for a mineral clay.

Mass-transfer and pressure-drop packing performance correlations are used together with tower design equations and detailed cost models to compare the effectiveness of cross-flow and counterflow air stripping towers over a wide range of contaminant volatility. Cross-flow towers are shown to offer a significant economic advantage over counterflow towers when stripping low volatility contaminants primarily due to savings in energy costs. These savings increase as contaminant volatility decreases and as water flow rate increases. A further advantage of the cross-flow configuration is that it extends the feasible operating range for air stripping as cross-flow towers can accommodate higher air-to-water flow ratios than conventional counterflow towers. Finally it is shown that the optimized least-cost design for both counterflow and cross-flow towers varies with Henry's law constant, water flow rate, and percent removal, but that the optimum is virtually insensitive to other cost and operating variables. This greatly simplifies the tower design procedure.

The design of anaerobic treatment systems for sulfate containing high-strength wastewater requires the understanding of interactions between sulfate reducers and other anaerobes. The objectives of this research were to study the effects of the feed propionate:sulfate ratio on the interactions among the sulfate-reducing bacteria (SRB), acetogens, and methanogens, and to determine the kinetics of anaerobic biological reactions for propionate degradation under sulfate-limited and sulfate-rich conditions. Both continuous and batch experiments were conducted to meet these objectives. In sulfate-limited chemostats with propionate as the sole carbon source, the effluent sulfate concentration was governed by the propionate utilizing SRB and was below 10 mg/L. In sulfate-rich systems the effluent sulfate concentration was governed by acetate utilizing SRB and could not be maintained below 400 mg/L without controlling the sulfide concentration. Under these sulfate-rich conditions, a decrease in SRT from 30 to 13 d decreased the extent of acetate utilization by SRB. The propionate utilizing SRB outcompeted the propionate utilizing acetogens in the chemostats. This was attributed to lower Monod-half velocity coefficient, K_s, for propionate utilization by SRB (4.5 mg-Pr/L), than the K_s for propionate utilization by acetogens (20.4 mg-Pr/L). The rate constants (k) for SRB and acetogens in terms of mg of propionate/mg of volatile suspended solids (VSS) per day were approximately 5.2 and 3.1 mg-Pr/mg-VSS-d, respectively. In sulfate-rich conditions, propionate was utilized by SRB, and the intermediate acetate was utilized by both SRB and methanogens.

Proper design of a soil vapor extraction (SVE) treatment system at a site contaminated with volatile organic compounds (VOCs) hinges on the effective characterization of vapor flow through the subsurface. The collection and analysis of soil vacuum data during vapor extraction provides the basis for this characterization. Although the collection of soil vacuum data is routinely performed across the industry at VOC-contaminated sites, a systematic and quantitative method of analyzing vacuum data typically is not employed. Thus a comprehensive analysis is presented for the characterization of vapor flow through the subsurface. This methodology includes: (1) the interpretation of vacuum data using statistical analysis to infer site stratigraphy and soil permeability; (2) diagnosis of vapor flow “short circuits”; and (3) flow path mapping. The statistical analysis includes the determination of the 95% confidence intervals of the collected SVE data. The confidence intervals are used to determine if deviations of individual data points from the linear trend are statistically significant. Once a region of homogeneity is delineated through the statistical analysis, flow net construction is performed. Case studies are presented to demonstrate how vacuum data is used to characterize subsurface vapor flow and to discuss the SVE design implications of these characterizations.

Liquid-liquid extraction was investigated for use with surfactant enhanced subsurface remediation. A surfactant liquid-liquid extraction model (SLLEM) was developed for batch equilibrium conditions based on contaminant partitioning between micellar, water, and solvent phases. The accuracy of this fundamental model was corroborated with experimental results (using naphthalene and phenanthrene as contaminants and squalane as the extracting solvent). The SLLEM model was then expanded to nonequilibrium conditions. The effectiveness of this nonequilibrium model was corroborated with experimental results from continuous flow hollow fiber membrane systems. The validated models were used to conduct a sensitivity analysis evaluating the effects of surfactants on the removal of the contaminants in liquid-liquid extraction systems. In addition, liquid-liquid extraction is compared to air stripping for surfactant-contaminant separation. Finally, conclusions are drawn as to the impact of surfactants on liquid-liquid extraction processes, and the significance of these impacts on the optimization of surfactant-enhanced subsurface remediation.

The effects of emission concentrations of particulate matters and sulfates on plume opacity are investigated by in situ measurements. The studies are conducted for three processes of two coal-fired plants and one oil-fired that are all equipped with electrostatic precipitators. Flue-gas sampling and analysis include the concentrations of particles and total water soluble sulfates, particle size distribution, and flue-gas composition; while in-stack and out-of-stack opacities are determined by a transmissometer and certified smoke inspectors, respectively. Experimental results show that plume opacity outside the stack linearly correlates well with the in-stack opacity. The mixing of hot flue gas with cold ambient air would result in the condensation of hygroscopic sulfuric acid aerosols and an increase about 1.6% out of typical 15–25% measured opacity. An empirical equation similar to the Beer-Lambert-Bouger form is derived for predicting the plume opacity in terms of the stack diameter and the concentrations of particles and total water soluble sulfates. Good comparisons are achieved between predictions by the empirical equation and other available field data.

A mass balance model was developed and applied to the Sacramento River in northern California during the July 1991 Sacramento River metam-sodium spill. The transport and reactions of metam-sodium, a soil fumigant, and the volatile and toxic methyl isothiocyanate (MITC) were simulated during the two-and-a-half days of movement along a 68-km stretch of river. Results from modeling were compared with field data for MITC, which is the only product measured downriver after the spill. Agreement between the simulated and measured values of MITC concentrations were found at Doney Creek (65.9 km downstream). Results illuminated the complexities and unique characteristics associated with the multiple kinetic processes of the chemical plume in the river. In particular, the photolysis of metam-sodium followed zero-order kinetics for high concentrations and first-order kinetics for low concentrations, a unique phenomenon consistent with the finding reported in a laboratory study. Concentrations of metam-sodium for transition from zeroto first-order, obtained by calibration and model sensitivity analyses, were in the same range as those in the reported laboratory results.

Simple budget models and correlations have been useful in providing order-of-magnitude estimates of eutrophication variables and toxic organic chemicals in lakes and reservoirs. A similar approach is used here to develop correlations to support management of phosphorus-limited, drinking-water supplies.

Stream reaeration has been studied extensively for several decades, yet little is known about the effect of barge traffic in rivers and canals. To address this lack of knowledge, a laboratory stream model was constructed that consists of a circular duct where model barges were towed at different traffic velocities and frequencies under various water conditions. Dissolved oxygen levels were monitored after initial deoxygenation. Excluding the effect of wind and flow, the reaeration coefficient (overall gas transfer coefficient) in the laboratory system increased as barge speed and traffic frequency (number of barge passages per hour) increased. The reaeration coefficient due to moving barges is inversely related to water depth. Finally, the reaeration coefficient due to moving barges is proportional to barge size, channel size, load, and draft.