Environ. Sci. Technol. 1997, 31, 704-709
Effects of Pore Water Velocity on the Transport of Arsenate JEFFREY E. DARLAND Department of Environmental Science and Engineering, Oregon Graduate Institute of Science and Technology, Portland, Oregon 97291-1000 WILLIAM P. INSKEEP* Department of Plant, Soil, and Environmental Sciences, Montana State University, Bozeman, Montana 59717
The potential for AsO4 to exhibit sorption-related nonequilibrium during transport has not been previously determined. Consequently, the objectives of this study were to determine the applicability of transport models based on the advection-dispersion equation (ADE) using various equilibrium and kinetic sorption expressions for describing AsO4 mobility. Saturated column transport experiments were performed using an applied AsO4 pulse at pore water velocities (PWVs) of 0.2, 1.0, 10, and 90 cm h-1 through a sand where the principal reactive phase was amorphous or poorly crystalline iron oxides. Observed AsO4 breakthrough curves (BTCs) demonstrated sorption-related nonequilibrium as evidenced by a leftward shift in observed BTCs and an increase in observed effluent recovery with increasing PWV. The use of independently derived equilibrium and kinetic sorption parameters in the ADE failed to describe observed AsO4 BTCs at all PWVs. Apparent sorption rate coefficients obtained by fitting observed BTCs to an nth-order kinetic model increased with increasing PWV. The time scales of the fitted rate coefficients are significantly longer than previously determined rate coefficients describing the chemical step of arsenate sorption by iron oxide minerals, suggesting that slower diffusional processes control the rate of arsenate sorption-desorption during transport.
Introduction Most metal-arsenate solid phases are too soluble (e.g., Ca, Fe, Mn, and Al) to exist in contaminated soils, and aqueous concentrations of AsO4 are often controlled by adsorptiondesorption reactions onto mineral surfaces. Moreover, AsO4 is generally preferentially sorbed to metal oxide surfaces compared to layer silicates (1-5). Arsenate adsorption by iron hydroxides can be categorized as a ligand exchange mechanism, and it has been shown by extended X-ray absorption fine structure spectroscopy (EXAFS) that AsO4 forms a combination of mono- and bidentate complexes with surface Fe sites (6). For AsO4 adsorption onto R-FeOOH, surface complexation models (e.g., constant capacitance model) have been used to successfully describe equilibrium sorption conditions over the pH range of 3.5-10 (7, 8). The rate of the chemical step describing the surface complexation of AsO4 onto iron oxides and clay minerals is quite rapid, generally on the order of milliseconds (8-10). Studies have shown, however, that even in well-mixed batch systems the approach to equilibrium can be slow, on the order of hours, for AsO4 adsorption onto ferrihydrite (5). These results indicate that mass transfer limitations such as intra* Corresponding author e-mail address:
[email protected].
704
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 3, 1997
particle and film diffusion may play a significant role in the kinetics of AsO4 sorption onto soil mineral phases (5). The extent to which these kinetic processes may affect the transport of AsO4 through soils and/or aquifer materials is unclear. Sorption processes influenced by slow mass transfer rates may be subject to chemical (sorption related) nonequilibrium under transport conditions (11). Under these situations, it is characteristic to observe changes in the shape (e.g., tailing) of solute breakthrough curves as a function of pore water velocity (11-14). Moreover, transport-derived rate constants representing the sorption process may vary with pore water velocity (11, 14, 15), indicating that the diffusional processes controlling sorption rates are sensitive to the mixing environment (16). While there has been substantial research on batch equilibrium sorption reactions of AsO4 onto geological minerals and natural soils, there have been few studies investigating the role of kinetic limitations on AsO4 sorption by soils. Furthermore, there is a lack of detailed data on the effects of kinetic processes on AsO4 transport through porous media. Consequently, the objectives of this study were to (i) determine the applicability of two equilibrium adsorption models (linear and Freundlich) and two simple kinetic adsorption models (first-order and nth-order reversible) in describing AsO4 transport using the advection-dispersion equation (ADE) and (ii) determine the effect of pore water velocity on apparent AsO4 sorption rate coefficients.
Transport Model Formulations The advection-dispersion equation describes the onedimensional steady-state transport of adsorbing solutes through soils:
δs δc δc δ2c +θ ) θD 2 - v δt δt δx δx
F
(1)
where F represents soil bulk density (M L-3), s represents sorbed concentration (M M-1), t represents time (T), θ represents volumetric water content (L3 L-3), c represents solute concentration (M L-3), D represents the hydrodynamic dispersion coefficient (L2 T-1), v represents average Darcy velocity (L T-1), and x represents distance (L). The left side of this mass-balance equation represents the distribution of solute in the sorbed (δs/δt) and liquid (δc/δt) phases, while the right side represents mass flux of solute due to dispersion and advection. Fundamentally, adsorption can be modeled as either an equilibrium or a kinetic process, depending on whether the sorbed phase is locally at equilibrium with the resident solution phase concentration (17). Common equilibrium and kinetic expressions (Table 1) used to describe the change in sorbed phase concentration (δs/δt) in the ADE include both linear and Freundlich (nth-order) adsorption isotherms, where KD, KF, KD′, and KF′ are distribution coefficients (L3 M-1), kf is a forward reaction rate constant (T-1), kr is a reverse reaction rate constant (T-1), and n is a dimensionless Freundlich reaction order. It is often assumed that the resident solution phase concentration is locally at equilibrium with the sorbed phase concentration (defined as the local equilibrium assumption, LEA). Solutions to the ADE using the LEA commonly employ the linear isotherm (e.g., ref 8) (Table 1, case A) and the Freundlich isotherm (19-21) (Table 1, case B). There have been many published reports, however, which indicate that nonequilibrium conditions during transport preclude the use of the LEA with either linear or Freundlich isotherms (13, 20, 22, 23).
S0013-936X(96)00247-7 CCC: $14.00
1997 American Chemical Society
TABLE 1. Sorbed Phase Accumulation Terms Used in the Basic Advection-Dispersion Equation and Their Corresponding Values at Equilibrium model
∂c ∂s ) KD ∂t ∂t
linear equilibrium
Case B
freundlich equilibrium ∂s ) KFc(1/n) ∂c ∂t n ∂t
Case C
first-order reversible
a
nth-order reversible
∂s θ ) k c - krs ∂t F f
s ) KDc s ) KFc1/n s ) KD′c θ kf KD′ ) F kr
1/n ∂s θ 1/n ) k c - krs s ) KF′c ∂t F f θ kf KF′ ) F kr
All symbols defined in list of symbols.
Transport models based on the ADE using kinetic expressions to describe adsorption-desorption processes have also been previously published (20, 21, 24, 25). The simplest kinetic expression is a first-order reversible reaction (Table 1, case C), which at equilibrium corresponds to a linear isotherm (20, 25). Kinetic expressions describing nth-order reversible adsorption (which at equilibrium reduces to the Freundlich isotherm) have also been used in the ADE (Table 1, case D) (e.g., refs 20, 21, and 24). Few studies have been directed toward the applicability of these models to the transport of AsO4, which is an important contaminant in soils, aquifers, and surface waters. Moreover, the extent to which batch-derived equilibrium and/or kinetic parameters may be useful for predicting AsO4 transport is unclear.
Materials and Methods Reagents and Analysis. Sand of mixed mineralogy (quartz, orthoclase, plagioclase feldspars) was obtained from the Unimin Corp., Emmett, ID (Granusil Grade 50), sieved to a particle size range of 0.25-0.50 mm, acid-washed with 0.05 M HCl (3 times), and then rinsed with double deionized water (DD-H2O) until the conductivity of the rinse stabilized at
