Kinetics of Haloacetic Acid Reactions with Fe(0) - American Chemical

acid were mass transfer limited, with corrected mass transfer coefficients of 3.7-3.9 × 10-4 m/s. The reactions of trichloroacetic acid (TCAA), dichl...
0 downloads 0 Views 146KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 6881-6889

Kinetics of Haloacetic Acid Reactions with Fe(0) LI ZHANG, WILLIAM A. ARNOLD,* AND RAYMOND M. HOZALSKI* Department of Civil Engineering, University of Minnesota, Minneapolis, Minnesota 55455-0116

Detailed kinetic studies of the reactions of haloacetic acids (HAAs) with Fe(0) were performed in longitudinally mixed batch reactors. The reactions of tribromoacetic acid (TBAA), bromodichloroacetic acid, and chlorodibromoacetic acid were mass transfer limited, with corrected mass transfer coefficients of 3.7-3.9 × 10-4 m/s. The reactions of trichloroacetic acid (TCAA), dichloroacetic acid (DCAA), chloroacetic acid (CAA), and bromoacetic acid (BAA) were reaction limited. Bromochloroacetic acid (BCAA) and dibromoacetic acid (DBAA) were partially reaction limited. For the reaction limited species and partially reaction limited species, intra- and interspecies competition effects were observed. A Langmuir-Hinshelwood-HougenWatson kinetic model incorporating a mass transfer term was adopted to account for these effects. The lumped kinetic parameters for the HAAs ranged from 0.04 to 248 µM min-1 for an iron loading of 0.3 g of Fe/125 mL and followed the trend DBAA > BCAA > TCAA > BAA > DCAA. The adsorption parameters ranged from 0.0007 to 0.0065 µM-1. The effect of dissolved oxygen (DO) on the reaction of TBAA or BAA with Fe(0) was also investigated. No significant effect of DO on the reaction rate of TBAA, which is a mass transfer limited species, was observed. A lag phase, however, was observed for the reaction of BAA, which is a reaction limited species, until the DO was depleted. Simulations were performed to investigate the potential significance of the reactions of HAAs with Fe(0) in water distribution systems.

Introduction Haloacetic acids (HAAs) are an important class of disinfection byproducts (DBPs) formed during the chlorination of water and wastewater. HAAs occur in surface waters due to wastewater discharges and deposition of HAAs formed by photodegradation of chlorinated solvents in the atmosphere (1, 2). The main HAAs of concern in drinking water are the nine chlorinated and/or brominated species that include chloroacetic acid (CAA), bromoacetic acid (BAA), dichloroacetic acid (DCAA), bromochloroacetic acid (BCAA), dibromoacetic acid (DBAA), trichloroacetic acid (TCAA), bromodichloroacetic acid (BDCAA), chlorodibromoacetic acid (CDBAA), and tribromoacetic acid (TBAA). CAA, DCAA, and TCAA are the major HAAs detected in the environment with reported concentrations for the sum of these three compounds of up to 1.3 µg/L for surface water, 78 µg/L for drinking water, and 640 µg/L for wastewater (1, 3). * Address correspondence to either author. (W.A.A.) Phone: (612)625-8582; fax: (612)626-7750; e-mail: [email protected]. (R.M.H.) Phone: (612)626-9650; fax: (612)626-7750; e-mail: hozal001@ umn.edu. 10.1021/es049267e CCC: $27.50 Published on Web 11/05/2004

 2004 American Chemical Society

Many HAAs are known or suspected carcinogens (4), and their concentrations in drinking water are currently regulated by the U.S. Environmental Protection Agency. Under the Stage 1 Disinfectants/Disinfection Byproducts Rule, the sum of five HAAs (HAA5 ) CAA + DCAA + TCAA + BAA + DBAA) in drinking water must comply with the maximum contaminant level of 60 µg/L (5). To understand the fate of HAAs in water distribution systems and to develop possible treatment processes to remove HAAs from drinking water or wastewater, understanding the reactivity of HAAs is essential. Fe(0) is a powerful reductant that readily reduces halogenated aliphatic compounds, nitroaromatic or halogenated aromatic compounds, and inorganic compounds (6-30). Previously, we reported the rates and pathways of the degradation of four trihalogenated HAAs (TCAA, TBAA, BDCAA, and CDBAA) by Fe(0) (18). Pseudo-first-order degradation rates for these compounds ranged from 0.08 to 10.6 h-1, and the compounds were degraded via sequential hydrogenolysis (18). The pH was not controlled in the former experiments, and the degradation rates of the reaction intermediates were not investigated independently. In addition, pseudo-first-order kinetics may not be sufficient to interpret surface mediated reactions. Thus, detailed kinetic studies of the reactions of the nine chloro- and bromo-HAAs with Fe(0) under controlled pH conditions were performed, and the results are reported herein. Intra- and interspecies competition effects were systematically investigated. The potential role of mass transfer was also studied, for mass transfer limitations are often important in experimental reactors and in engineered systems such as permeable reactive barriers used to treat contaminants in groundwater (31-33). Of the approximately 370 000 mi of water mains in place in the United States, 22% of the pipes are composed of unlined cast iron or ductile iron (34). Thus, reduction of HAAs by Fe(0) may play a role in determining the fate of HAAs in water distribution systems. It is also important to consider the potential inhibitory effects of competing oxidants, such as chlorine and oxygen, that are ubiquitous in U.S. water distribution systems (35). Limited experiments have been conducted to investigate the effects of dissolved oxygen on Fe(0)-mediated reactions. The degradation rates of TCE and carbon tetrachloride in a reactor with a palladized-iron cathode and platinum mesh anode did not change with increasing dissolved oxygen concentrations from 0 to 9 mg/L (14). Conversely, oxygen inhibited the reduction of TCE by Fe(0) in the gas phase (36), and the transformation rate of 1,2-dibromo-3-chloropropane by Fe(0) decreased in the presence of oxygen (37). Thus, experiments were performed to investigate the effect of oxygen on the rates of HAA degradation by Fe(0).

Kinetic Models To predict the persistence of HAAs in systems where Fe(0) is present, information concerning contaminant reaction kinetics is required. Pseudo-first-order kinetic models are often used to quantify heterogeneous reaction rates (6, 7, 10, 14, 16, 19, 21). Some studies, however, have demonstrated that pseudo-first-order kinetic models may not be appropriate for Fe(0) surface-mediated reactions (11-13, 25-31). A mixed-order model (eq 1) was developed by Wu ¨ st et al. (11) to account for declining rate constants as the initial concentration of trichloroethylene (TCE) increased:

[TCE]k0 d[TCE] )dt [TCE] + K1/2 VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

(1) 9

6881

where k0 is the zero-order rate constant and K1/2 corresponds to the aqueous TCE concentration at one-half of the maximum transformation rate. Mixed-order kinetics were also used to fit experimental data for 2-chloroacetophenone and 2,4,6-trinitrotoluene (TNT) degradation (29). Intra- and interspecies competitive effects were observed during the reduction of chlorinated ethylenes and chlorinated acetylenes by Fe(0) (12, 13, 31). The single-site Langmuir-Hinshelwood-Hougen-Watson (LHHW) kinetic model (eqs 2 and 3) was used to account for the competition for reactive sites (12, 13, 31): Nj

dCi

(

∑k

s

ij

St)KiCi

)-

j)1

dt

) -kobsCi

Nm

1+

(2)

∑K

m Cm

m)1 Nj

dCj

∑

( )

Np

kijsSt)KiCi

j)1

dt

( -

∑k

Nm

1+

St)KjCj (for all j)

Nm

∑K

m Cm

m)1

s jp

p)1

1+

∑K

(3)

mCm

6882

9

dCA ) -kma(CA - CAs) dt

m)1

where (kijsSt) is the lumped kinetic parameter (µM min-1), kijs is the kinetic constant for a surface-limited reaction, St is the abundance of sites per liter (linearly related to surface area per liter), and K is the adsorption parameter (µM-1). Equation 2 represents the reaction of a parent species i, and eq 3 is for the appearance and disappearance of all its possible daughter products j. The denominators in eqs 2 and 3 account for all possible intra- and interspecies competitive effects. The mixed-order model (eq 1) does not include interspecies competition. A nonequilibrium model accounting for adsorption, desorption, and dechlorination was also developed to predict TCE and cis-dichloroethylene degradation by Fe(0) (25). The model predictions were in good agreement with experimental and modeling results of Arnold and Roberts (12, 13). Bandstra and Tratnyek (26) developed a two-site LHHW model to provide a better fit of the experimental data for the reduction of TNT by Fe(0). An exponential decay term was added to a model based on Michaelis-Menten (MM)/LHHW kinetics to account for the effect of aging Fe(0) particles during the reaction with TCE (27). The MM/LHHW model is essentially a single-site LHHW kinetic model without interspecies competition terms (27). Janda et al. (30) tested kinetic models for the reactions of TCE, tetrachloroethylene, and trichloromethane with Fe(0). The results showed that first-order kinetics were not suitable for modeling the experimental results. Power law models, LHHW analogy models, and general models of heterogeneous reactions provided a better fit (30). The single-site LHHW kinetic model, however, should provide an adequate fit in most cases (26). For the surface-mediated reactions of interest in this research, a HAA molecule must be transported to the Fe(0) surface prior to reaction. Thus, mass transfer may play a role in the observed kinetics. To identify mass transfer limitations and consequently identify different kinetic scenarios under our experimental conditions (longitudinally rotated bottles), the mass transfer coefficient (km) needs to be known. An empirical expression developed by Harriott (38) and used for zero-valent metal systems (32) is used for the estimation of km:

Sh )

where Sh is the Sherwood number, Re is the Reynolds number, Sc is the Schmidt number, dp is the particle diameter (m), and Dw is the diffusion coefficient (m2/s). Further details and calculations are provided in the Supporting Information. Three scenarios are possible depending on the relative rate of mass transfer of the HAA to the Fe(0) surface and the reaction rate: (i) the mass transfer limited case, where mass transfer of HAA to the surface is slower than the reactions at the surface, and the aqueous concentration near the iron surface of the parent compound is zero; (ii) the reaction limited case, where reaction of HAA on the surface is slower than the mass transfer of HAA to the surface, and the aqueous concentration near the iron surface of the parent compound is equal to the bulk concentration; (iii) the mixed control or partial reaction limited case, where the mass transfer rate of HAA to the surface is comparable to the reaction rate of HAA at the surface. In this last case, an expression including both mass transfer and kinetic terms is required. The simple “resistances in series” model [(1/koverall) ) (1/kmass transfer) + (1/kreaction)] cannot be used with LHHW kinetics, for the reaction term is not first order. The necessary expression is found by first specifying the near-surface concentration of species A as CAs. The mass transfer rate is then:

kmdp ) 2 + 0.6Re1/2Sc1/3 Dw

(4)

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004

(5)

where a is the geometric surface area of the metal per volume of solution (assuming spherical particles) and CA is the bulk, aqueous concentration. The expression describing the surface reaction is

dCA (ksSt)KACAs )dt 1 + KACAs

(6)

The right-hand sides of eqs 5 and 6 are set equal to solve for CAs in terms of rate coefficients, KA, and CA. The expression for CAs is then substituted into eq 6 to give the desired expression (see Supporting Information):

dCA dt

)

[(

-(ksSt)KA CA -

[(

2 + KA CA -

) x( ) x(

ksSt 1 + KA kma

ksSt 1 + KA kma

ksSt 1 - CA + k ma K A

ksSt 1 + - CA kma KA

)

)

2

]

CA KA

+4

2

]

CA KA (7)

+4

Kinetic modeling was performed for reaction limited species and partially reaction limited species. For mass transfer controlled reactions, kobs is simply equal to kma, and no kinetic modeling was necessary. Given the general success of the LHHW model for Fe(0)-mediated reactions, this kinetic model (eqs 2 and 3) was used for the reaction limited species. Equation 7 was applied for species exhibiting mixed control.

Experimental Section Chemicals. The following chemicals were used as received: CAA (99+%, Aldrich), BAA (99+%, Aldrich), DCAA (99+%, Aldrich), BCAA (99.9%, Supelco), DBAA (97%, Aldrich), TCAA (99%, Aldrich), BDCAA (99.9%, Supelco), CDBAA (99.9%, Supelco), TBAA (99+%, Aldrich), sodium phosphate dibasic (98+%, EM Science), sodium phosphate monobasic (99.4%, Fisher Scientific), hexadimethrine bromide (HDMBr) (100%, Aldrich), sodium acetate trihydrate (100%, Fisher Scientific), sodium bromide (100%, Fisher Scientific), sodium chloride

(100%, Mallinckrodt Baker Inc.), tetradecyltrimethylammonium bromide (TTABr) (99%, Sigma), 3-morpholinopropanesulfonic acid (MOPS) (99.5%, Sigma), sodium hydroxide (97%, Aldrich). MOPS buffer (50 mM adjusted to pH 7.5 with 0.5 M NaOH) was deoxygenated (argon-sparged) before adding to the batch reactors except for experiments investigating oxygen competition. Argon used for deoxygenation was purified using an in-line molecular sieve and oxygen traps. Metal Preparation. Electrolytic iron (100 mesh, Fisher Scientific) was selected for the study because of its high purity and relatively uniform particle size. The iron particles were acid-washed using the procedures described in refs 9 and 12. The cleaned iron was stored in an anaerobic chamber (containing 7% H2 and 93% N2) and used within 24 h. The surface area measured via BET was 0.061 m2/g, and the external surface area assuming spherical particles was calculated to be 0.0051 m2/g. Experimental Systems. All batch experiments, except those with CAA, were carried out in 125 mL serum bottles. All batch reactors, except those for the oxygen competition experiments, were prepared in an anaerobic chamber. The bottles were filled with an aqueous HAA solution buffered at pH 7.5 with 50 mM deoxygenated MOPS. The bottles were wrapped in aluminum foil to prevent photodecomposition reactions, and acid-washed electrolytic iron was then added to the bottles. The bottles were topped with Teflon-faced butyl rubber septa (Fisher Scientific), sealed with aluminum crimp caps, and loaded onto a rotator (Glas-Col Laboratory). The bottles were mixed about their longitudinal axes at 45 rpm except when the effect of mixing speed was investigated. At predetermined intervals, the bottles were removed from the rotator and placed over a magnet to settle the iron particles. Approximately 0.4 mL of sample was withdrawn from the reactors while simultaneously injecting 0.4 mL of 50 mM pH 7.5 deoxygenated MOPS buffer. Each sample was filtered through a 0.2 µm pore size syringe-mounted filter (Gelman Acrodisc) into a polyethylene autosampler vial for analysis by capillary electrophoresis (CE). Preliminary experiments were conducted to investigate the effect of iron loading and the effect of mixing rate. In the iron loading experiments with 40 µM BDCAA as the parent compound, the rotational speed was fixed at 39 rpm, and iron loadings varied from 0.1 to 0.5 g of iron/125 mL. In the mixing rate experiments, TBAA or BDCAA was chosen as the parent compound because fast reacting species are more susceptible to mass transfer limitations. Again the initial concentration was 40 µM. The mixing rate was varied from 15 to 60 rpm for a fixed iron loading of 0.3 g/125 mL. To study intraspecies competition effects, the initial concentration of the target HAA was varied from 15 to 816 µM. Five or six experiments at different initial concentrations were conducted simultaneously for each HAA except CAA. To study interspecies competition effects, two species were simultaneously introduced into the reactor, and the impact of the second species on the reaction rate of the first species was determined. The target species was TCAA or BAA and the competitor was CAA for TCAA and was TCAA, DCAA, CAA, or acetate for BAA. The initial concentration of the target species was in the range of 80-100 µM, and the initial concentration of the competitor was 18-25 times higher than that of the target species. Thus, it can be assumed that the competitor concentration is constant, even when the competitor is produced by the reaction of the target species, which simplifies the kinetic modeling. Because the reaction rate was expected to be slow as compared to the other HAAs, intraspecies competition was not investigated for CAA, and only one experiment was performed (C0 ) 280 µM) at an iron loading of 0.5 g/38 mL.

In the oxygen competition experiments, the iron was added to the serum bottles in the anaerobic chamber. The bottles were then removed, filled with a 40 µM TBAA or 100 µM BAA solution that had been equilibrated with air, and capped. Bottles prepared with deoxygenated solutions served as positive controls. The batch bottles were periodically removed from the rotator and sampled for analysis of TBAA or BAA. For measurement of dissolved oxygen (DO), sets of serum bottles identical to those used for monitoring HAA concentration were prepared. One of these bottles was sacrificed for measurement of DO at each sampling event. All batch experiments were performed at room temperature (21 ( 1 °C). The pH of the solutions at the end of incubations were measured using a pH meter (Accumet portable AP62, Fisher Scientific) and ranged from 7.6 to 8.0. Analytical Methods. HAAs, acetate, chloride, and bromide were analyzed by CE (model HP3D CE, Hewlett-Packard). The analytical method was similar to that described previously (18) except that an extended light path capillary (75 m i.d. × 56 cm effective length, Agilent) was used for increased sensitivity. Because BAA and acetate comigrated under these conditions, whenever BAA was present in a sample, a longer standard capillary (75 m i.d. × 104 cm effective length) together with a background electrolyte comprised of 25 mM phosphate and 1 mM tetradecyltrimethylammonium hydroxide (TTAOH) at pH 5.7 was used for the separation. The TTAOH was prepared by passing a TTABr aqueous solution through an anion exchange cartridge (Maxi-clean ICOH plus, Alltech Associates). Dissolved oxygen was measured using an oxygen microelectrode (MI730, Microelectrode Inc.). The microelectrode was calibrated with a two-point calibration curve (deoxygenated water and air-saturated water) at room temperature. Kinetic Modeling. Kinetic modeling was performed using Scientist for Windows (v. 2.01, Micromath Research, St. Louis, MO). The entire suite of intraspecies competition experiments for HAAs, which encountered reaction limitations (except CAA), were fit simultaneously to generate the adsorption and lumped kinetic parameters for the parent compound using either eq 2 for reaction limited species or eq 7 for the mixed-control species. The adsorption and lumped kinetic parameters for an intermediate were set to those values determined in the intraspecies experiments where it was used as the parent compound. For each data set at a different initial concentration, the concentrations of parent compound and daughter products were normalized by the measured parent compound concentration at t ) 0 (12, 13, 31). Additionally, overall, initial pseudo-first-order rate constants (kobs) were determined via linear regression of plots of ln(concentration) versus time. As the reaction of CAA with Fe(0) was not performed at different initial concentrations, the adsorption constant for CAA was obtained by fitting the interspecies competition data. The adsorption parameter for acetate was generated by fitting the acetate and BAA interspecies competition data. The kinetic and adsorption parameters for BAA, however, were needed to fit the BAA-acetate interspecies competition data. Those parameters were obtained by fitting the BAA intraspecies experimental data using a model without the acetate competition term. Then, the kinetic and adsorption parameters for BAA were regenerated using a model with acetate competition and the adsorption parameter for acetate. The same approach was used to obtain kinetic and adsorption parameters for DCAA and TCAA and the adsorption parameter for CAA. Because the reaction of CAA with Fe(0) is very slow (estimated t1/2 ) 199 d scaled to an iron loading of 0.3 g/125 mL), degradation of CAA by Fe(0) was assumed to be negligible in the DCAA intraspecies competition experiments where the reactions were run for up to 17 d. VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6883

TABLE 1. Observed Rate Constant (kobs) for the Reactions of HAAs with Fe(0) at Different Initial HAA Concentrations and Calculated and Corrected Mass Transfer Coefficients compd TBAA

CDBAA

BDCAA

DBAA

BCAA

TCAA

DCAA

BAA

CAAb a

C0 (µM)

kobs (min-1)

15 40 99 205 397 20 42 102 203 406 15 43 100 213 427 45 77 112 224 430 44 78 115 220 426 16 42 105 216 425 45 110 202 405 56 114 208 434 627 816 279

0.34 ( 0.32 ( 0.03 0.33 ( 0.02 0.34 ( 0.04 0.31 ( 0.02 0.29 ( 0.07 0.30 ( 0.03 0.26 ( 0.01 0.28 ( 0.01 0.26 ( 0.02 0.27 ( 0.05 0.26 ( 0.05 0.25 ( 0.02 0.28 ( 0.02 0.29 ( 0.02 0.14 ( 0.01 0.12 ( 0.01 0.11 ( 0.01 (8.43 ( 1.60) × 10-2 (8.80 ( 1.10) × 10-2 (7.57 ( 1.06) × 10-2 (8.50 ( 0.49) × 10-2 (7.53 ( 0.62) × 10-2 (6.45 ( 0.63) × 10-2 (6.15 ( 0.38) × 10-2 (1.14 ( 0.33) × 10-2 (9.51 ( 1.80) × 10-3 (8.66 ( 0.47) × 10-3 (8.37 ( 0.33) × 10-3 (6.30 ( 0.45) × 10-3 (3.06 ( 0.50) × 10-4 (1.36 ( 0.09) × 10-4 (1.25 ( 0.06) × 10-4 (1.27 ( 0.04) × 10-4 (4.47 ( 1.08) × 10-3 (3.27 ( 1.15) × 10-3 (2.73 ( 0.63) × 10-3 (1.99 ( 0.56) × 10-3 (2.02 ( 0.46) × 10-3 (1.52 ( 0.43) × 10-3 (1.36 ( 0.11) × 10-5

calculated km (m/s)

0.08a

Errors represent 95% confidence limits.

b

calculated kma (min-1)

corrected kma (min-1)

10-4

0.113

0.274

mass transfer limited

1.56 × 10-4

0.116

0.281

mass transfer limited

1.61 × 10-4

0.119

0.289

mass transfer limited

1.67 × 10-4

0.123

0.299

partially reaction limited

1.71 × 10-4

0.127

0.309

partially reaction limited

1.80 × 10-4

0.134

0.326

reaction limited

1.86 × 10-4

0.138

0.335

reaction limited

1.89 × 10-4

0.140

0.340

reaction limited

2.00 × 10-4

0.148

0.360

reaction limited

1.52 ×

Note different iron loading as described in text.

Results and Discussion Reaction Pathways. As described previously (18), HAAs undergo sequential hydrogenolysis in the presence of Fe(0) with bromine preferentially removed over chlorine. CAA was the final degradation product of HAAs containing chlorine for experiments of up to 4 d. Nevertheless, the reduction of CAA by Fe(0) was observed with the concomitant production of acetate in a long-term experiment (4 months; Supporting Information). Consequently, all HAAs reacted with Fe(0) via sequential hydrogenolysis, and acetate was the terminal reaction product. Effect of Metal Loading and Mixing Rate. The kobs for the reaction of BDCAA with Fe(0) was linear for iron loadings ranging from 0 to 0.4 g/125 mL (Supporting Information). Thus, an iron loading of 0.3 g/125 mL was selected for all subsequent experiments. There was little effect of mixing rate on kobs for the reactions of TBAA or BDCAA with Fe(0) for mixing rates from 36 to 60 rpm (Supporting Information). At mixing rates less than 36 rpm, however, an effect on kobs was seen. A mixing rate of 45 rpm was selected for all subsequent experiments. The metal loading and mixing rate studies were used to verify maximum efficiency of the mixing system. Reactions may still be mass transfer limited under 6884

9

kinetic domain

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004

these conditions, however, because mass transfer is limited by the particle settling velocity, and this rate may still be slower than the reaction rate for rapidly reacting species. Detailed Kinetics. The calculated km values for each HAA are listed in Table 1. Results from the intraspecies competition experiments are also summarized in Table 1 and Figure 1. Reactions of TBAA, CDBAA, and BDCAA with Fe(0) were mass transfer limited as their calculated kma values (assuming spherical particles, where a is the ratio of geometric surface area of the particles, to the volume of solution) were less than their kobs values at each initial concentration (Table 1). The calculated kma values include a correction factor (multiplied by 1.5) for a recirculating liquid (38). Because the calculated kma values for the mass transfer limited cases were only about one-third of the kobs values, the kma values for each HAA were multiplied by an additional correction factor. This correction factor was obtained by computing the ratio of kobs/calculated kma for each mass transfer limited reaction and then taking the arithmetic mean of the values. The additional correction factor likely reflects uncertainty in assumptions necessary to calculate kma. Smooth, spherical iron particles were assumed in the calculation of a and in calculation of the terminal settling velocity (Supporting

FIGURE 1. Effect of initial substrate concentration on observed rate constants for the reduction of TBAA (2), DBAA(b), and DCAA (9) by Fe(0). Error bars represent 95% confidence intervals. Experiments were conducted in 125 mL of 50 mM deoxygenated MOPS (pH 7.5) with 0.3 g of iron and rotated at 45 rpm. The dashed lines are presented for visualization only.

TABLE 2. kobs for the Target Species in the Interspecies Competition Experiments and in the Control Experiments

a

experiment

target species

C0 of target species (µM)

competitor

C0 of competitor (µM)

kobs (min-1)

1 2 3 4 5 (control) 6 7 (control)

BAA BAA BAA BAA BAA TCAA TCAA

81 80 81 80 75 108 100

TCAA DCAA CAA Acetate none CAA none

2040 1765 1879 1696 naa 1914 NA

(2.3 ( 0.2) × 10-3 (2.4 ( 0.2) × 10-3 (2.1 ( 0.5) × 10-3 (1.5 ( 0.3) × 10-3 (2.9 ( 0.4) × 10-3 (2.9 ( 0.4) × 10-3 (8.7 ( 0.5) × 10-3

Not applicable.

Information). Shape and roughness effects on these values may explain the factor of ∼3 required to make the predictions match the experimental results. The corrected kma value for the other HAAs was compared to the kobs values so that the rate controlling process for each reaction could be identified. If kobs was much less than the corrected kma, the system was deemed reaction limited. If kobs ≈ corrected kma, the system was deemed mass transfer limited. If kobs/corrected kma was in the range of 0.2-1, the system was considered partially reaction limited (Table 1). The corrected kma values were then used for the mass transfer term in the kinetic model (eq 7) for partially reaction limited species (i.e., DBAA and BCAA). The kobs for the reactions of TBAA, CDBAA, and BDCAA were independent of initial HAA concentration (Table 1). This result is consistent with mass transfer limitation. In this case, kobs should be the product of the mass transfer coefficient (km) and the metal surface area per liter of the solution (a). As stated above, a correction factor was necessary to achieve this equality. For TCAA, DCAA, and BAA, kobs decreased with increasing initial concentration (Table 1). For example, kobs decreased by a factor of 2.4 for DCAA with increasing initial concentration from 45 to 405 µM (Figure 1 and Table 1). By comparing the corrected kma values with the kobs values, the reactions of TCAA, DCAA, and BAA were found to be reaction limited. The decreasing trend in kobs with increasing initial concentration of each species suggests intraspecies competition for reactive sites on the Fe(0) particles. For BCAA and DBAA, kobs also decreased with increasing initial concentration (Figure 1 and Table 1) although the changes in kobs were smaller than those for TCAA, DCAA,

and BAA. This observed decrease in kobs also provided evidence for intraspecies competition. Comparing kobs to the kma values, the reactions of BCAA and DBAA were controlled by a mixture of reaction and mass transfer (Table 1). The percentage of reaction limitation [(1 - kobs/corrected kma) × 100%] of BCAA reactions increased from 76% to 80% with increasing initial concentration of BCAA from 44 to 426 µM. In the case of DBAA, the percentage increased from 55% to 70% as initial concentration of DBAA increased from 45 to 430 µM (Supporting Information). This implies that as the initial concentration of BCAA or DBAA increases, the reaction of BCAA or DBAA with Fe(0) becomes the more important rate-limiting step due to an increasing intraspecies competitive effect. The observed decreases in kobs with increasing initial concentration need not be attributed to intraspecies competition, although this is the explanation that has been most commonly invoked (11-13, 25-31). The effect could result from any two rate processes (one fast and one slow) that are weighted by initial concentrations. Even if this is the case, the LHHW model is still a useful empirical fit of observed behavior. Thus, without additional evidence for alternative controlling processes, we use the LHHW model. Interspecies Competition. No significant change in kobs (at the 95% confidence level) for BAA was observed when the competitor was TCAA, DCAA, or CAA as compared to the control experiment with BAA alone (Table 2). Nevertheless, kobs for BAA decreased when the competitor was acetate (Table 2). The kobs for TCAA also decreased (at the 95% confidence level) in the presence of CAA as compared to the control experiment with TCAA alone (Table 2). VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6885

TABLE 3. Model-Derived Kinetic and Adsorption Parameters for HAAs and Acetate

parent 1 2 3 4 5 6 7 8

DCAA DCAA TCAA TCAA BAA BAA DBAA BCAA

primary product CAA CAA DCAA DCAA acetate acetate BAA CAA

competitor 1 2 a

acetate CAA

From Intraspecies Competition Experiments kSSt Ki (µM min-1)a (µM-1)b nc 0.0405 ( 0.00732 0.0504 ( 0.0111 15.9 ( 11.0 15.1 ( 11.5 2.27 ( 0.53 1.89 ( 0.35 248 ( 554 132 ( 218

10-3

(6.49 ( 2.06) × (4.85 ( 1.65) × 10-3 (7.84 ( 5.71) × 10-4 (8.77 ( 7.06) × 10-4 (1.91 ( 0.63) × 10-3 (2.30 ( 0.66) × 10-3 (6.97 ( 16.0) × 10-4 (8.10 ( 14.1) × 10-4

5 5 5 5 6 6 5 5

note without CAA competition term with CAA competition term kDCAAfCAA, K DCAA from 2 kDCAAfCAA, KDCAA from 1 with acetate competition term without acetate competition term kBAAfacetate, KBAA from 5

From Interspecies Competition Experiments target species Kcompetitor (µM-1)

Kacetate ) (1.02 ( 0.29) × 10-3 KCAA ) (8.13 ( 1.52) × 10-4

BAA TCAA

nc 1 1

ksSt values are for 0.3 g of Fe/125 mL buffer solutions. b Ki is for the parent HAA. c n is the number of experiments.

If the target species and competitors react at the same sites, kobs for the target species (i.e., BAA or TCAA) can be expressed as

kobs )

(ksSt)targetKtarget 1 + KtargetCtarget + KcompetitorCcompetitor

(8)

A 2-10-fold decrease in kobs for BAA or TCAA was expected under the experimental conditions based on the K values from the intraspecies competition results. The anticipated decrease in kobs was only observed in the TCAA-CAA and BAA-acetate experiments. This revealed that TCAA and CAA compete for reactive sites. The inhibition by acetate indicates that species that do not undergo chemical transformation are still capable of binding to reactive sites. On the basis of these results, interspecies competition effects were included in the kinetic modeling where appropriate. The results also suggest that species containing bromine and those containing only chlorine may react at different sites on the Fe(0) surface. Because chlorinated species compete with each other, we speculate that brominated species would also compete with each other for reactive sites. Due to the complicating effects of mass transfer limitation for the brominated species, this could not be tested in our experimental system. The lack of interspecies competition exerted on BAA by the chlorinated HAAs may also indicate that another kinetic paradigm (as mentioned above) is necessary for the brominated species or that differences in physical/chemical properties of the species containing bromine versus those only containing chlorine may somehow influence the competition. Again, without additional data necessary to evaluate such scenarios, we have chosen to interpret our data in the context of the LHHW model. Model Fitting. Model derived kinetic and adsorption parameters for HAAs that encountered full or partial reaction limitation are shown in Table 3. The kinetic parameters for TCAA, DCAA, or BAA were smaller than those for DBAA or BCAA. This is consistent with TCAA, DCAA, and BAA being reaction limited and DBAA and BCAA being partially reaction limited. The adsorption parameters for the HAAs were in the range of 0.0007-0.0065 µM-1. Whether or not CAA and acetate interspecies competition terms are included, the parameters for DCAA and BAA do not change significantly at the 95% confidence level (Table 3). This was possibly because the concentrations of CAA and acetate generated in the reactions were relatively small. The experimental data and model fits for the reduction of DCAA and DBAA by Fe(0) at high and low initial 6886

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004

FIGURE 2. Reduction of DCAA (b) by 0.3 g of Fe(0) at initial concentrations of (a) 45 and (b) 405 µM. Major products: CAA (9) and Cl- (2). CAA competition term is incorporated in the model. The experiments were conducted in 125 mL of 50 mM deoxygenated MOPS buffer (pH 7.5) and rotated at 45 rpm. Lines represent model fits based on the parameters presented in Table 3 (-, DCAA; - -, CAA and Cl- ). Chloride was always below the detection limit (100 µM) for C0 ) 45 µM. concentrations are shown in Figures 2 and 3, respectively. In general, the models fit the experimental data reasonably well. This suggests that an LHHW model incorporating intraspecies and interspecies competition is appropriate for fitting the experimental data, and the use of corrected mass transfer coefficients is reasonable. Oxygen Competition Experiments. TBAA was degraded in 10 min in the presence of oxygen (initial DO of 300 µM), and no significant change in kobs for TBAA was observed as

its presence delayed the degradation of BAA, which is a reaction limited species. This provides evidence that oxygen, which is mass transfer limited, was able to out compete BAA for reactive sites. Nevertheless, competition for reactive sites was not important for the TBAA reaction because the ratelimiting step for both species was the transport to the surface, not the reaction. Thus, the presence of dissolved oxygen should not affect the observed kinetics for mass transfer limited species. Environmental Significance. Reduction of HAAs by Fe(0) may play a role in the fate of HAAs in water distribution systems comprised of iron (unlined cast or ductile) pipe. Furthermore, the degradation of HAAs by Fe(0) suggests a potential process for removing HAAs from water and wastewater. A water distribution system calculation is shown below and that for a packed bed system is provided in the Supporting Information. A steady-state plug flow reactor equation was used to compute the concentration of HAAs along the water distribution system:

(

C ) C0 exp -

FIGURE 3. Reduction of DBAA (b) by 0.3 g of Fe(0) at initial concentrations of (a) 45 and (b) 430 µM. Major products: BAA (9) and Br- (2). The experiments were conducted in 125 mL of 50 mM deoxygenated MOPS buffer (pH 7.5) and rotated at 45 rpm. Lines represent model fits based on the parameters presented in Table 3 (-, DBAA; - -, BAA; ‚‚‚, Br-). Note that Br- produced exceeds that of BAA because the BAA reacts to form acetate (not shown).

FIGURE 4. Reduction of BAA (b) by 0.3 g of Fe(0) in the presence of oxygen (9). ([O2]0 ) 300 µM). Major products: acetate (results not shown) and Br- (2). The experiment was conducted in 125 mL of 50 mM MOPS buffer (pH 7.5) and rotated at 45 rpm. The dashed lines are presented for visualization only. compared to the control experiment without oxygen (results not shown). Oxygen was depleted in 15 min. A lag phase was observed for BAA in the first 10 min when oxygen was present (Figure 4). After the lag phase, BAA was degraded at a rate [kobs ) (6.47 ( 2.23) × 10-3 min-1] that was not statistically different at the 95% confidence level from that for the control experiment without oxygen for C0 ) 114 µM (Table 1). Although the presence of oxygen did not affect the degradation of TBAA, which is a mass transfer limited species,

)

kobsx u

(9)

where C is the concentration of HAA along the water distribution system, C0 is the initial concentration of the HAA of interest, u is the flow velocity, and x is the distance along the water distribution system. Initial concentrations of 7.5, 30, 40, 18, 18, and 1.6 µg/L were assumed for CAA, DCAA, TCAA, BCAA, DBAA, and BAA, respectively. These are the maximum reported HAA concentrations in finished drinking waters (3). A mass transfer correlation for a smooth pipe (39, 40) was used to calculate mass transfer coefficients for HAAs in the water distribution system:

Sh )

kmd ) 0.023Re0.83Sc0.333 D

(10)

where d is pipe diameter (6 in. or 15.2 cm), and u ) 0.1 ft/s (0.03 m/s). Mass transfer coefficients of HAAs were in the range of 1.5 × 10-6 to 1.8 × 10-6 m/s. The reaction rate coefficients for CAA, DCAA, TCAA, BCAA, DBAA, and BAA were calculated using eq 2. The (ksSt) and K values were obtained from the batch experiments and (ksSt) values for the HAAs were scaled up according to the ratio of available iron surface area to solution volume [12.2 m-1 for pipe surface 100% covered with Fe(0)] and assuming similar reactivity for different sources of iron. Note that the observed concentrations of individual HAAs in distribution systems are less than 1 µM. At these concentrations, the intra- and interspecies competitive effects of the HAAs are minimal (i.e., ∑KiCi , 1), and the reaction rate coefficient for an HAA can be estimated as (ksSt)Ki, which is a pseudo-first-order rate constant, scaled to the appropriate iron surface area to solution volume ratio. Competition for sites with other oxidants (oxygen, free chlorine) may be important for reaction limited species but are ignored in these calculations. Using the resistances in series model for first-order processes, the overall first-order rate constant kobs was then used in eq 9. Under these conditions, TBAA, BDCAA, CDBAA, BCAA, DBAA, and TCAA are mass transfer limited; DCAA and BAA are partially mass transfer limited; and CAA is reaction limited. In water distribution systems, the DO concentration is typically in the range of 7-9 mg/L (35) except in some areas (e.g., dead zones), which may be anoxic. On the basis of our batch experiments containing TBAA or BAA and oxygen, we speculate that reduction of HAAs by Fe(0) may be important for the fate of mass transfer limited species but not important VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6887

for the fate of partially or fully reaction limited species in water distribution systems. Taking TCAA as a representative mass transfer limited species and assuming no production of TCAA from the reactions of natural organic matter with residual chlorine, TCAA would be completely degraded to DCAA within 3 mi (4.8 km) in a new distribution system (100% Fe(0), kobs,TCAA ) 0.0023 min-1). For a water distribution system where 10% of the pipe surface is Fe(0) (kobs,TCAA ) 0.00023 min-1), TCAA would be completely degraded within 18 mi (29 km). Similar results are obtained for the other mass transfer limited species. If the percentage of Fe(0) on the pipe surface is far less than 10%, as might be the case in older, corroded pipes, the degradation of HAAs by Fe(0) may not be important. On the basis of our experimental results investigating oxygen competition effects, we hypothesize that oxygen and other mass transfer limited oxidants would block the reactions of DCAA, BAA, and CAA, leading to little degradation of these species. Pipe surfaces in water distribution systems are often covered by corrosion solids, and HAAs might also react with Fe(II) or mixed Fe(II)/Fe(III) corrosion products (e.g., green rust, magnetite) or Fe2+ sorbed on corrosion solids. Thus, reactions of HAAs with these minerals may also play a role in the fate of HAAs in water distribution systems.

Acknowledgments The authors thank the National Science Foundation (Grant BES-0332085) and the American Water Works Association Research Foundation (Project 2644) for financial support. The authors also thank the University of Minnesota for providing a fellowship to L.Z. We also thank the anonymous reviewers for their insightful comments on the kinetic model and possible interpretations.

Supporting Information Available Procedure for the calculation of km, derivation of the mixedcontrol kinetic expression, a plot of CAA degradation by Fe(0), figures illustrating the effects of Fe(0) loading and mixing speed, a figure showing the dependence of the fraction of reaction control as a function of initial concentration for DBAA and BCAA, additional plots of experimental data with LHHW model fits, and an Fe(0) packed bed design calculation. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Berg, M.; Muller, S. R.; Muhlemann, J.; Wiedmer, A.; Schwarzenbach, R. P. Concentrations and mass fluxes of chloroacetic acids and trifluoroacetic acid in rain and natural waters in Switzerland. Environ. Sci. Technol. 2000, 34 (13), 2675-2683. (2) Frank, H.; Scholl, H.; Renschen, D.; Rether, B.; Laouedj, A.; Norokorpi, Y. Haloacetic acids, phytotoxic secondary air pollutants. Environ. Sci. Pollut. Res. Int. 1994, 1 (1), 4-14. (3) Krasner, S. W.; Pastor, S.; Chinn, R.; Sclimenti, M. J.; Weinberg, H. S.; Richardson, S. D.; Thruston, A. D., Jr. The occurrence of a new generation of DBPs (beyond the ICR). Proc.-Water Qual. Technol. Conf. 2001, 1592-1615. (4) National Academy of Science Safe Drinking Water Committee. Drinking Water and Health; National Academy Press: Washington, DC, 1980; Vol. 6. (5) U.S. EPA. National Primary Drinking Water Regulations: Disinfectants and Disinfection Byproducts; Final Rule. Fed. Regist. 1998, 63, 69389-69476. (6) Gillham, R. W.; O’Hannesin, S. F. Enhanced degradation of halogenated aliphatics by zero-valent iron. Ground Water 1994, 32 (6), 958-967. (7) Burris, D. R.; Campbell, T. J.; Manoranjan, V. S. Sorption of trichloroethylene and tetrachloroethylene in a batch reactive metallic iron-water system. Environ. Sci. Technol. 1995, 29 (11), 2850-2855. 6888

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004

(8) O’Hannesin, S. F.; Gillham, R. W. Long-term performance of an in situ “iron wall” for remediation of VOCs. Ground Water 1998, 36 (1), 164-170. (9) Fennelly, J. P.; Roberts, A. L. Reaction of 1,1,1-trichloroethane with zero-valent metals and bimetallic reductants. Environ. Sci. Technol. 1998, 32 (13), 1980-1988. (10) Su, C.; Puls, R. W. Kinetics of trichloroethene reduction by zerovalent iron and tin: pretreatment effect, apparent activation energy, and intermediate products. Environ. Sci. Technol. 1999, 33 (1), 163-168. (11) Wu ¨ st, W. F.; Koeber, R.; Schlicker, O.; Dahmke, A. Combined zero- and first-order kinetic model of the degradation of TCE and cis-DCE with commercial iron. Environ. Sci. Technol. 1999, 33 (23), 4304-4309. (12) Arnold, W. A.; Roberts, A. L. Pathways and kinetics of chlorinated ethylene and chlorinated acetylene reaction with Fe(0) particles. Environ. Sci. Technol. 2000, 34 (9), 1794-1805. (13) Arnold, W. A.; Roberts, A. L. Inter- and intraspecies competitive effects in reactions of chlorinated ethylenes with zero-valent iron in column reactors. Environ. Eng. Sci. 2000, 17 (5), 291302. (14) Li, T.; Farrell, J. Reductive dechlorination of trichloroethene and carbon tetrachloride using iron and palladized-iron cathodes. Environ. Sci. Technol. 2000, 34 (1), 173-179. (15) Farrell, J.; Melitas, N.; Kason, M.; Li, T. Electrochemical and column investigation of iron-mediated reductive dechlorination of trichloroethylene and perchloroethylene. Environ. Sci. Technol. 2000, 34 (12), 2549-2556. (16) Farrell, J.; Kason, M.; Melitas, N.; Li, T. Investigation of the longterm performance of zero-valent iron for reductive dechlorination of trichloroethylene. Environ. Sci. Technol. 2000, 34 (3), 514-521. (17) Kim, Y.-H.; Carraway, E. R. Dechlorination of pentachlorophenol by zero-valent iron and modified zero-valent irons. Environ. Sci. Technol. 2000, 34 (10), 2014-2017. (18) Hozalski, R. M.; Zhang, L.; Arnold, W. A. Reduction of haloacetic acids by Fe0: Implications for treatment and fate. Environ. Sci. Technol. 2001, 35 (11), 2258-2263. (19) Butler, E. C.; Hayes, K. F. Factors influencing rates and products in the transformation of trichloroethylene by iron sulfide and iron metal. Environ. Sci. Technol. 2001, 35 (19), 3884-3891. (20) Mantha, R.; Taylor, K. E.; Biswas, N.; Bewtra, J. K. A continuous system for Fe0 reduction of nitrobenzene in synthetic wastewater. Environ. Sci. Technol. 2001, 35 (15), 3231-3236. (21) Alowitz, M. J.; Scherer, M. M. Kinetics of nitrate, nitrite, and Cr(VI) reduction by iron metal. Environ. Sci. Technol. 2002, 36 (3), 299-306. (22) Oh, S.-Y.; Cha, D. K.; Chiu, P. C. Graphite-mediated reduction of 2,4-dinitrotoluene with elemental iron. Environ. Sci. Technol. 2002, 36 (10), 2178-2184. (23) Arnold, W. A.; Winget, P.; Cramer, C. J. Reductive dechlorination of 1,1,2,2-tetrachloroethane. Environ. Sci. Technol. 2002, 36 (16), 3536-3541. (24) Westerhoff, P.; James, J. Nitrate removal in zero-valent iron packed columns. Water Res. 2003, 37 (8), 1818-1830. (25) Scha¨fer, D.; Ko¨ber, R.; Dahmke, A. Competing TCE and cis-DCE degradation kinetics by zero-valent ironsexperimental results and numerical simulation. J. Contam. Hydrol. 2003, 65 (3-4), 183-202. (26) Bandstra, J. Z.; Tratnyek, P. G. Applicability of single-site rate equations for reactions on inhomogeneous surfaces. Ind. Eng. Chem. Res. 2004, 43 (7), 1615-1622. (27) Venkatapathy, R.; Bessingpas, D. G.; Canonica, S.; Perlinger, J. A. Kinetics models for trichloroethylene transformation by zerovalent iron. Appl. Catal. B: Environ. 2002, 37 (2), 139-159. (28) Devlin, J. F.; March, C. Investigating the kinetic limitations of granular iron over a large range of 4-chloronitrobenzene concentrations. Prepr. Pap. ACS Natl. Meet., Am. Chem. Soc., Div. Environ. Chem. 2003, 43, 585-589. (29) Miehr, R.; Tratnyek, P. G.; Bandstra, J. Z.; Scherer, M. M.; Alowitz, M. J.; Bylaska, E. J. Diversity of contaminant reduction reactions by zerovalent iron: role of the reductate. Environ. Sci. Technol. 2004, 38 (1), 139-147. (30) Janda, V.; Vasek, P.; Bizova, J.; Belohlav, Z., Kinetic models for volatile chlorinated hydrocarbons removal by zero-valent iron. Chemosphere. 2004, 54, 917-925. (31) Arnold, W. A. Ph.D. Thesis, The Johns Hopkins University: Baltimore, MD, 1999. (32) Arnold, W. A.; Ball, W. P.; Roberts, A. L. Polychlorinated ethane reaction with zero-valent zinc: pathways and rate control. J. Contam. Hydrol. 1999, 40 (2), 183-200.

(33) Scherer, M. M.; Johnson, K. M.; Westall, J. C.; Tratnyek, P. G. Mass transport effects on the kinetics of nitrobenzene reduction by iron metal. Environ. Sci. Technol. 2001, 35 (13), 2804-2811. (34) AWWA and AWWARF. Water:\Stats 1996 Survey Database; 1998. (35) Zhang, M.; Semmens, M.; Schuler, D.; Hozalski, R. M. Biostability and microbiological quality in a chloraminated distribution system. J. Am. Water Works Assoc. 2002, 94 (9), 112-122. (36) Uludag-Demirer, S.; Bowers, A. R. Effects of surface oxidation and oxygen on the removal of trichloroethylene from the gas phase using elemental iron. Water, Air, Soil Pollut. 2003, 142 (1-4), 229-242. (37) Siantar, D. P.; Schreier, C. G.; Chou, S.-S.; Reinhard, M. Treatment of 1,2-dibromo-3-chloropropane and nitrate-contaminated

water with zero-valent iron or hydrogen/palladium catalysts. Water Res. 1996, 30 (10), 2315-2322. (38) Harriott, P. Mass transfer to particles: Part 1. Suspended in agitated tanks. AIChE J. 1962, 8, 93-102. (39) Rossman, L. A.; Clark, R. M.; Grayman, W. M. Modeling chlorine residuals in drinking-water distribution systems. J. Environ. Eng. 1994, 120 (4), 803-820. (40) Edwards, D. K.; Denny, V. E.; Mills, A. F., Transport Processes; McGraw-Hill: New York, 1976.

Received for review May 17, 2004. Revised manuscript received August 20, 2004. Accepted September 9, 2004. ES049267E

VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6889