Iron Oxide Surface-Catalyzed Oxidation of Ferrous Iron by

In addition, other studies have shown that carbonate readily sorbs to iron oxide .... After a period of 1−2 h, 40 μL of 10 mg/L Fe(II) stock was in...
7 downloads 0 Views 113KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 512-519

Iron Oxide Surface-Catalyzed Oxidation of Ferrous Iron by Monochloramine: Implications of Oxide Type and Carbonate on Reactivity PETER J. VIKESLAND* AND RICHARD L. VALENTINE Department of Civil and Environmental Engineering, The University of Iowa, Iowa City, Iowa 52242

The maintenance of monochloramine residuals in drinking water distribution systems is one technique often used to minimize microbial outbreaks and thereby maintain the safety of the water. Reactions between oxidizable species and monochloramine can however lead to undesirable losses in the disinfectant residual. Previous work has illustrated that the Fe(II) present within distribution systems is one type of oxidizable species that can exert a monochloramine demand. This paper extends this prior work by examining the kinetics of the reactions between Fe(II) and monochloramine in the presence of a variety of iron oxide surfaces. The identity of the iron oxide plays a significant role in the rate of these reactions. Surface areanormalized initial rate coefficients (kinit) obtained in the presence of each oxide at pH ≈6.9 exhibit the following trend in catalytic activity: magnetite > goethite > hematite ≈ lepidocrocite > ferrihydrite. The differences in the activity of these oxides are hypothesized to result from variations in the amount of Fe(II) sorbed to each of the oxides and to dissimilarities in the surface site densities of the oxides. The implications of carbonate on Fe(II) sorption to iron oxides are also examined. Comparing Fe(II) sorption isotherms for goethite obtained under differential carbonate concentrations, it is apparent that as the carbonate concentration (CT,CO3) increased from 0 to 11.7 mM that the Fe(II) sorption edge (50% sorption) shifts from a pH of approximately 5.8 to a pH of 7.8. This shift is hypothesized to be the result of the formation of aqueous and surface carbonate-Fe(II) complexes and to competition between carbonate and Fe(II) for surface sites. The implications of these changes are then discussed in light of the variable oxide studies.

Introduction As treated water is transported through a distribution system, it interacts with materials present at the pipe-water interface. These materials, often referred to as pipe deposits, can have a tremendous adverse effect on overall water quality (1, 2). Deposits, which consist of a complex mixture of iron oxides and hydroxides (goethite, magnetite, lepidocrocite, ferrihy* Corresponding author present address: Department of Civil and Environmental Engineering, Virginia Polytechnic Institute, 407 Durham Hall, Blacksburg, VA 24061-0246. E-mail: [email protected]. 512

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 3, 2002

drite), metal carbonates (siderite, calcite, aragonite), ferrous iron, and natural organic matter (2-4), readily react with drinking water disinfectants (5-7). These reactions ultimately lead to disinfectant loss and to subsequent declines in water quality. In light of increasingly stringent disinfection guidelines and disinfectant dosages, an improved understanding of the interactions that can occur between deposit materials and drinking water disinfectants is necessary. Unfortunately, due to their heterogeneous nature, experiments involving actual deposit material do not provide quantitative information relating to the reaction mechanisms responsible for disinfectant loss. As such, experiments with actual deposits only serve to illustrate the importance of deposit-mediated disinfectant losses. In an effort to better understand the types of interactions that can occur between disinfectants and deposit material, we have attempted to characterize the reactions that occur between monochloramine, a commonly employed disinfectant (8), and ferrous iron. By starting with this relatively simple system, it is possible to gain insight into the complicated processes responsible for depositmediated disinfectant loss within distribution systems. Our previous work (9, 10) has shown that Fe(II) and monochloramine readily react via a direct interaction between molecular monochloramine and aqueous ferrous iron, Fe(II)soln. These reactions are autocatalytic, with the iron oxide product of the aqueous-phase reaction acting to accelerate the overall reaction rate by enabling the formation of highly reactive Fe(II) surface complexes (>FeOFe+ and >FeOFeOH, collectively referred to hereafter as Fe(II)surf). Electron spin resonance evidence indicates that these reactions produce the radical intermediate amidogen (•NH2). This radical can react with Fe(II), or it is scavenged through reactions with other species (e.g., carbonate, oxygen). These scavenging reactions result in a nonelementary stoichiometry for this system. Kinetic studies (10) that accounted for all of these processes led to the development of the following overall rate expression:

-

d[NH2Cl] 1 d[Fe(II)tot] )) dt Θ dt kNH2Cl,soln[OH-][Fe(II)soln]

(

+ kNH2Cl,surf1[>FeOFe+]

+ kNH2Cl,surf2[>FeOFeOH]

)

[NH2Cl] (1)

where Θ is the experimentally measured reaction stoichiometry and is included to account for the scavenging of the amidogen radical; kNH2Cl,soln (M-2 min-1) is the rate coefficient for the aqueous-phase reactions; kNH2Cl,surf1 (M-1 min-1) and kNH2Cl,surf2 (M-1 min-1) are rate coefficients for the surfacemediated reactions involving the surface complexes >FeOFe+ and >FeOFeOH. Experimentally determined values for the rate coefficients and Θ are tabulated in Table 1. Because the iron oxide precipitate catalyzes the reactions between Fe(II) and monochloramine, it can be inferred that other oxides, such as those found within distribution system deposits, could also act catalytically. Previous studies have illustrated the catalytic effect of a number of iron oxides on redox reactions involving Fe(II) and a wide variety of oxidants (11-18). These oxides are hypothesized to act catalytically because complexation of Fe(II) at the oxide surface increases electron density at the iron center, thereby making it a better reductant than Fe(II) in solution (17, 19). Prior work has illustrated that different oxides exhibit variable catalytic effects on Fe(II) oxidation (12, 15, 18). 10.1021/es010935v CCC: $22.00

 2002 American Chemical Society Published on Web 01/04/2002

TABLE 1. Model Parameters for Aqueous-Phase and Autocatalytic Reactions between Fe(II) and Monochloraminea kinetic equations and parameters kNH

2Cl,soln

NH2Cl + Fe(II)soln + OH- 98 products K R 2Cl,surf1 surf1

kNH

NH2Cl + >FeOFe+ 98 products K R 2Cl,surf2 surf2

kNH

NH2Cl + >FeOFeOH 98 products apparent stoichiometry (Θ)

value

ref

3.1 × 109 M-2 min-1

10

0.56 M-1 min-1

10

3.5 × 10-8 M-1 min-1

10

0.32 × pH - 0.92

9

a

As described elsewhere (10), the rate coefficients for the surface-mediated reactions are presented as mixed coefficients that reflect not only reaction kinetics but also sorption of Fe(II) (Ksorb1 and Ksorb2) to a limited number of surface sites (R).

Klausen et al. (12) observed that at a fixed pH of ≈7.0 the ability of the oxide to catalyze the reactions between surface Fe(II) and substituted nitrobenzenes decreased as the oxide was changed from magnetite to goethite and ultimately to lepidocrocite. Similarly, Buerge and Hug (15) observed that at pH 4.9 Fe(II) sorbed to goethite exhibited a slightly higher reactivity toward Cr(VI) reduction than Fe(II) sorbed to lepidocrocite. They attributed the differential reactivity of the oxides to be the result of differences in the total amount of Fe(II) sorbed to the oxide surface at any given pH (15). Because of these variations and due to the heterogeneity of the iron oxide and hydroxide phases present within distribution system deposits, it is important to assess the catalytic activities of a wide range of oxide species to determine how each alters the kinetics of the reactions between Fe(II) and monochloramine. In conjunction with our investigation of the effect of variations in the oxide type on the Fe(II)-NH2Cl reactions, we also examine the implications of carbonate concentration on Fe(II) complexation at the oxide surface. Although a considerable amount of research has examined Fe(II) complexation to a variety of iron oxide surfaces (13, 16, 20, 21), little if any work has examined the effects of carbonate on this process. This is unfortunate because drinking water distribution systems typically contain significant carbonate alkalinity, and this may affect Fe(II) sorption and reactivity. Previous work has shown that Fe(II) readily forms aqueous complexes with carbonate and that these complexes largely dictate the reactivity of Fe(II) toward both oxygen (22) and hydrogen peroxide (23). In addition, other studies have shown that carbonate readily sorbs to iron oxide surfaces (24-31). Taken together, these results suggest that sorbed carbonate could affect Fe(II) surface complexation and reactivity. Although the effect of carbonate on Fe(II) complexation at an iron oxide surface has not been previously investigated, recent work (24-26, 28-31) has illustrated that carbonate complexation at the oxide-water interface can alter the surface properties of the oxide and thereby affect cation adsorption. In the presence of carbonate, the iron oxide surface is coated by inner-sphere monodentate carbonatesurface complexes (29). The formation of these complexes decreases the isoelectric point (IEP) and the electrophoretic mobility of both goethite and amorphous iron oxides (25); these changes may directly affect surface complexation. In fact, increased carbonate concentrations have been found to enhance Pb(II) sorption onto goethite at pH values from 4.5 to 6.5 (28, 31). Along with the direct effects of carbonate sorption to the oxide surface, the formation of carbonate-cation solution complexes may also alter surface complexation. Work examining both Np(V) (26) and Pb(II) (28) sorption to goethite has shown that the formation of metal-carbonate solution complexes significantly alters metal sorption relative to carbonate-free samples. For Np(V) (26), the formation of the solution complexes NpO2CO3-, NpO2(CO3)23-, and NpO2(CO3)35- actually causes the amount of Np(V) sorbed to the

iron surface to decrease significantly as the pH is raised above 7.5. At pH values below 7.5, however, carbonate exhibited no significant effect on Np(V) surface complexation. As noted previously, higher carbonate concentrations enhanced Pb(II) sorption onto goethite at pH values from 4.5 to 6.5; however, they decreased Pb(II) sorption at pH values greater than 6.5 (28). This decrease was attributed to competition between the surface and the lead carbonate solution species. In light of these prior results and their potential implications on Fe(II) complexation at the oxide surface, we conducted a preliminary assessment of the effect of variable carbonate concentrations on Fe(II) complexation at the goethite surface. As we illustrate, changes in the carbonate concentration strongly affect Fe(II) surface complexation. This result suggests that a critical reevaluation of the importance of carbonate on the surface complexation of metals in both natural and engineered systems may be warranted.

Materials and Methods All experiments were conducted using deionized water produced by a Barnstead ULTRO pure water system. A Fisher Scientific model 50 pH meter coupled with an Orion Scientific combination reference/analytical electrode was used for all pH measurements. Activity corrections for all ionic species were calculated through the use of the Davies equation (32). Preparation of Iron Oxides. The iron oxides ferrihydrite (2-line), goethite, hematite, lepidocrocite, and magnetite were all prepared according to procedures described elsewhere (33). In short, ferrihydrite, goethite, and hematite were produced by precipitating the ferric salt Fe(NO3)3‚9H2O; lepidocrocite was produced by oxidizing an FeCl2‚4H2O solution; and magnetite was formed by anaerobically precipitating Fe(SO4)‚H2O. The resultant oxide slurries were dialyzed (6000-8000 MW cutoff; Spectra/Por Corp.) until the specific conductivity of the permeate solution was less than that of a 0.1 mM NaCl stock. To minimize alterations to the quasi-stable ferrihydrite precipitate, the dialysis procedure for ferrihydrite was limited to only 2 days. The oxides were isolated from the dialyzed slurries either by airdrying (goethite, lepidocrocite) or by freeze-drying (ferrihydrite, hematite, magnetite), and the dried materials were then stored at -5 °C. The crystalline identity of each oxide was confirmed by X-ray diffraction (XRD) analysis (Dr. George McCormick, The University of Iowa Geology Department). Multipoint N2-BET surface areas (Porous Materials, Inc., Ithaca, NY) for each oxide are tabulated in Table 2. Preparation of Solutions. Monochloramine stocks were produced according to previously published procedures (9, 34). Once prepared, the monochloramine stock was diluted using solutions of the same ionic strength and bicarbonate concentration. For all of the experiments involving monochloramine and Fe(II), a constant total carbonate concentration of 5.7 mM was employed. These dilute solutions were then poured into crimp-cap vials that were subsequently sealed. All experiments were conducted using deaerated VOL. 36, NO. 3, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

513

TABLE 2. Characteristics of Iron Oxides and βoxide Values BET surface area (m2 g-1) pHzpc

oxide

formula

ferrihydrite goethite hematite lepidocrocite magnetite

Fe5HO8‚4H2O R-FeOOH R-Fe2O3 γ-FeOOH Fe3O4

a

Ref 40.

b

Ref 42. c Ref 43.

d

246.0 42.6 38.5 97.7 14.5

βoxide (L M-1 m-2 min-1)

7.9a 15.0 ( 3.0 7.5b 340 ( 31.0 8.5c 48.0 ( 15.0 7.29d 52.0 ( 8.0 6.4d 501 ( 436

Ref 44.

conditions (dissolved oxygen (DO) < 1 mg/L). In control experiments, it was found that this level of DO did not lead to significant levels of iron oxygenation within the time scale of the Fe(II)-monochloramine reactions. Experimental Procedure. The experimental vials were prepared by equilibrating the dilute monochloramine solutions with a given iron oxide for a minimum of 1 h. The iron oxides were added to the monochloramine vials by injecting aliquots of a 10 g/L iron oxide stock. These stock solutions were buffered at the same pH as the monochloramine solutions. A concentrated iron oxide stock was employed because it was found that experimental reproducibility was greater with this technique than if the dried oxides were directly added to the vials. Comparisons between experimental vials containing a given oxide and control vials that did not contain any oxides indicated that over the time frame employed in these experiments none of the tested oxides exerted a measurable monochloramine demand (data not shown). The kinetic experiments were initiated by adding concentrated ferrous sulfate (10 µg/µL) to a vial containing monochloramine and a given iron oxide. The reaction temperature was set at 25 °C by keeping the reactor vials in a thermostated shaker table. Periodic samples were taken, and the Fe(II) and/or NH2Cl concentrations were determined. The dissolved and total Fe(II) concentration was measured using a modified Ferrozine method (35). In control experiments, the Ferrozine-Fe(II) complex was found to form instantaneously and to be unreactive toward monochloramine (36). It was therefore possible to quench the monochloramine-Fe(II) reactions by simply adding Ferrozine. Monochloramine was quantified via potentiometric titration with phenylarsine oxide (PAO). This method was used because of the interference of Fe(II)/Fe(III) on the standard DPD/FAS method (10). In general, rate expressions and kinetic coefficients for these reactions were obtained using measurements of Fe(II) concentration. Monochloramine concentrations were primarily measured to verify that the reactions were complete and that the reaction stoichiometry was constant. Fe(II) Sorption Isotherms. Sorption isotherms for Fe(II) onto goethite under variable carbonate concentrations were obtained using the following procedure: In an anaerobic glovebox, suspensions of goethite were prepared by mixing 0.25 g of goethite in a deaerated bicarbonate/CO2 solution for 1-2 h. The well-mixed suspension was then poured into 40-mL vials that were capped headspace free. The pH in these vials was adjusted by the addition of variable aliquots of 0.25 M NaOH or 0.25 M HCl. After a period of 1-2 h, 40 µL of 10 mg/L Fe(II) stock was injected to each vial. In the glovebox, the vials were mixed end-over-end for a period of 24-25 h after which the final solution pH, the dissolved Fe(II) concentration, and the total Fe(II) concentration were measured. For all vials, the total Fe(II) concentration was constant throughout these experiments, indicating that Fe(II) oxidation by trace quantities of oxygen was minimal. 514

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 3, 2002

FIGURE 1. Effect of goethite on iron oxidation rates. (A) Raw data and kinetic model fits. (B) Pseudo-first-order plot. Lines reflect initial rates only (ln [Fe(II)soln]/[Fe(II)]0 < -1.0). The deviation from pseudo-first-order behavior with time is attributed to the aqueousphase and the autocatalytic surface-catalyzed Fe(II)-NH2Cl reactions. [NH2Cl]0 ) 250 µM, [Fe(II)]0 ) 95 µM, pH 6.9 ( 0.05, CT,CO3 ) 5.7 mM, µ ) 0.1 M (NaClO4), DO < 1.0 mg/L. Θ ) 1.0 ( 0.1.

Results and Discussion The catalytic activities of ferrihydrite, goethite, hematite, lepidocrocite, and magnetite were evaluated by measuring ferrous iron oxidation rates in solutions containing variable quantities of a given oxide. For each oxide, it was determined that both the iron oxidation rate (Figure 1a) and the monochloramine reduction rate (Supporting Information Figure S1) increased as the oxide concentration was raised. During the initial phase of these reactions, the oxidation of Fe(II) followed a pseudo-first-order kinetic rate law (Figure 1b):

-

d[Fe(II)tot] ) kinit[Fe(II)soln] dt

(2)

where kinit (min-1) is the observed pseudo-first-order rate coefficient. Deviations from pseudo-first-order behavior were observed when the reactions were monitored over longer periods (Figure 1b). These deviations are attributed to the previously discussed autocatalytic effect of the ferric iron precipitate formed by the aqueous-phase reactions between Fe(II) and monochloramine. To minimize the influence of autocatalysis on the kinetic evaluation, only the initial reaction kinetics were considered when evaluating the catalytic activity of a given oxide. The overall reaction kinetics in the presence of a specific oxide are further evaluated in the following section.

-

(

d[Fe(II)tot] ) dt kNH2Cl,soln[OH-][Fe(II)soln] + kNH2Cl,surf1[>FeOFe+]

Θ + kNH2Cl,surf2[>FeOFeOH]

FIGURE 2. Catalytic effect of goethite on Fe(II) oxidation/NH2Cl reduction. The positive y-intercepts reflect the aqueous-phase and autocatalytic Fe(II)-monochloramine reactions and are offset from one another due to the differing monochloramine concentrations. [NH2Cl]0 ) 250 and 500 µM, [Fe(II)]0 ) 95 µM, pH 6.9 ( 0.05, CT,CO3 ) 5.7 mM, µ ) 0.1 M (NaClO4), DO < 1.0 mg/L.

For all of the oxides, pseudo-first-order iron oxidation kinetics were observed over a wide range of oxide concentrations, and a linear relationship exists between the kinit values and the surface area-normalized oxide concentration (in m2/L) (Figure 2). Except for the largest oxide loadings for goethite and hematite, the total Fe(II) concentration in these experiments was larger than the total number of surface sites. Under these conditions, as discussed by Amonette et al. (17), changes in the oxide concentration for a fixed total Fe(II) concentration lead to changes in the total concentration of sorbed Fe(II). As the oxide concentration increases, more Fe(II) can sorb to the iron surface, and thus additional reactive Fe(II)surf sites are created. The kinit values exhibited a first-order dependence on the initial monochloramine concentration, a result that is consistent with the first-order dependency previously observed for both the aqueous-phase and surface-mediated autocatalytic reactions (10). Because the reaction kinetics are a function of the monochloramine concentration, it is apparent that electron transfer between Fe(II) and monochloramine, and not the adsorption of Fe(II) to the oxide surface, is the rate-limiting process in this system. Were the reaction kinetics independent of the monochloramine concentration, it would have been a potential indication that the sorption of Fe(II) to the oxide surface was the primary mechanism for Fe(II) loss from solution. Modeling the Effect of Oxide Surfaces on Fe(II) Oxidation by Monochloramine. Our previous kinetic studies have illustrated that the reactions between Fe(II) and NH2Cl can be described by a model that accounts for the oxidation of dissolved ferrous iron and the oxidation of Fe(II) sorbed to the precipitated oxide product (10). In the presence of iron oxides other than the precipitate formed in situ, it is logical that analogous surface catalyzed reaction pathways exist: kNH

2Cl,oxide1

>FeoxideOFe+ + NH2Cl 98 products kNH

2Cl,oxide2

>FeoxideOFeOH + NH2Cl 98 products

(3)

+ kNH2Cl,oxide1[>FeoxideOFe ]

+ kNH2Cl,oxide2[>FeoxideOFeOH]

This expression implicitly neglects the effects of carbonate on Fe(II) speciation. This simplification is necessary due to the complicated role that carbonate plays in the monochloramine-Fe(II) reactions. Not only can carbonate affect Fe(II) speciation (22) but it can also accelerate monochloramine auto-decomposition (37, 38) and act as a radical scavenger (39). Because of these complicating effects, the incorporation of terms reflecting carbonate complexes into this expression was deemed scientifically unsound. Instead, we present the model in a general form and note that the rate coefficients and equilibrium constants should be considered “conditional” and not applicable to solution conditions that depart significantly from those described herein. For conditions where the pathways involving the added iron oxide dominate over the homogeneous or the autocatalytic pathways (e.g., for large oxide concentrations or for short reaction times before the autocatalytic oxide precipitate can form), this rate law simplifies to

-

d[Fe(II)tot] ) dt kNH2Cl,oxide1[>FeoxideOFe ] [NH2Cl] (6) Θ + kNH2Cl,oxide2[>FeoxideOFeOH]

(

)

This relationship is further simplified by assuming that the equilibria between ferrous iron in solution (Fe(II)soln) and the sorbed Fe(II) species develop sufficiently rapidly relative to the kinetics of the Fe(II)-NH2Cl redox reaction. Under this assumption, the Fe(II) surface complex concentrations can be written as [>FeoxideOFe+] ) Koxide1K-1 w [>FeoxideOH][OH ][Fe(II)soln] (7)

[>FeoxideOFeOH] )

- 2 Koxide2K-2 w [>FeoxideOH][OH ] [Fe(II)soln] (8)

where Koxide1 and Koxide2 are equilibrium coefficients that describe the formation of these two surface complexes (as noted above, these equilibrium coefficients are conditional in that they are functions of the carbonate concentration). Substituting into eq 6 leads to the following:

-

(

)

kNH2Cl,oxide1Koxide1K-1 d[Fe(II)tot] w [OH ] )Θ - 2 × dt + kNH2Cl,oxide2Koxide2K-2 w [OH ]

[NH2Cl][>FeoxideOH][Fe(II)soln] (9)

When the pH is fixed and monochloramine is present in significant excess, this expression simplifies to the pseudofirst-order expression given by eq 2, with

(4)

where >FeoxideOFe+ and >FeoxideOFeOH are the unprotonated and the protonated forms of the Fe(II) sorbed to the added oxide surface. Accordingly, eq 1 was extended to account for these additional reactions:

)

[NH2Cl] (5)

+

kinit ) Θ

(

kNH2Cl,oxide1Koxide1K-1 w [OH ] - 2 + kNH2Cl,oxide2Koxide2K-2 w [OH ]

)

×

[>FeoxideOH][NH2Cl]0 ) βoxide[>FeoxideOH][NH2Cl]0 (10)

VOL. 36, NO. 3, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

515

FIGURE 3. Comparison of the catalytic activity of ferrihydrite, goethite, hematite, lepidocrocite, and magnetite. The average y-intercept of 350 ( 40 min-1 M-1 reflects the aqueous-phase and autocatalytic Fe(II)-monochloramine reactions. [NH2Cl]0 ) 250500 µM, [Fe(II)]0 ) 90-95 µM, pH 6.9 ( 0.05, CT,CO3 ) 5.7 mM, µ ) 0.1 M (NaClO4), DO < 1.0 mg/L. where βoxide (L M-1 m-2 min-1) is a conditional overall rate coefficient for a given oxide under a given set of solution conditions. A preequilibrium assumption like that employed in the derivation of eqs 7 and 8 is valid so long as the kinetics of Fe(II) complexation at the iron surface are fast relative to the kinetics of electron transfer between the surface-bound Fe(II) and the aqueous oxidant (15). In a control experiment conducted using our reaction conditions but without monochloramine present, it was found that Fe(II) complexation at the iron oxide interface was characterized by a fast initial step (FeoxideOH] in eq 10, determined by normalizing the kinit values for each oxide to the initial NH2Cl concentration, were plotted versus the corresponding oxide concentrations to obtain values for βoxide at pH 6.9 in a 5.7 mM carbonate buffer (Figure 3; Table 2). These βoxide values, in conjunction with the previously determined rate coefficients for the aqueous-phase and autocatalytic reactions (10; Table 1), were then used to model Fe(II) oxidation in the presence of the different oxide surfaces. As illustrated in Figures 1a and 4, for reactions in the presence of goethite and lepidocrocite, respectively, the inclusion of terms for the aqueous-phase and the autocatalytic reactions enables us to model the entire progress of the reaction. When the terms for the solution and autocatalytic reactions were not 516

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 3, 2002

FIGURE 4. Effect of lepidocrocite on iron oxidation rates. [NH2Cl]0 ) 500 µM, [Fe(II)]0 ) 90 µM, pH 6.9 ( 0.05, CT,CO3 ) 5.7 mM, µ ) 0.1 M (NaClO4), DO < 1.0 mg/L. Θ ) 0.8 ( 0.1.

FIGURE 5. Evaluation of catalytic activity of goethite as a function of [OH-]. For comparative purposes, the normalized kinit values for the homogeneous reaction are also plotted (10). [NH2Cl]0 ) 250 µM (w/goethite) or 141-704 µM (w/o goethite), [Fe(II)]0 ) 90-95 µM, pH 5.80-8.05, CT,CO3 ) 5.7 mM, µ ) 0.1 M (NaClO4), DO < 1.0 mg/L. included the overall rate expression, the reaction rate was severely underestimated at the lowest oxide dosages (Supporting Information Figure S3). Additional model results for magnetite, hematite, and ferrihydrite may be found in Supporting Information Figures S4-S6. Further Evaluation of the Catalytic Effect of Goethite. An additional series of experiments were conducted using fixed concentrations of goethite () 1.70 m2/L), monochloramine () 500 µM), and ferrous iron () 95 µM) but variable solution pH. By varying the solution pH, it was possible to determine the pH dependence of the βoxide value for goethite. As the solution pH was raised from 5.80 to 7.71, the pseudofirst-order kinit values increased from 0.037 to 3.56 min-1 (data not shown), thereby indicating that the goethite surfacemediated reactions are a function of the solution pH. To determine the reaction order with respect to OH-, the collected kinit values were normalized to the monochloramine concentration (βoxide[>FeoxideOH] ) kinit/[NH2Cl]0), and the log of these values was plotted versus the log of the OHconcentration (Figure 5). Previous studies (10) have shown that changes in the apparent stoichiometry do not alter kinit significantly; therefore, it was unnecessary to normalize the kinit values by dividing through by Θ. The observed slope of one indicates that over the tested pH range that the surfacemediated reaction has an apparent first-order dependency on the hydroxide ion concentration. This first-order depen-

dence indicates that βoxide for goethite over this range of pH values can be expressed as βoxide ) Θ(kNH2Cl,oxide1Koxide1K-1 w [OH ])

(11)

Accordingly, for the reaction conditions tested here, this equation can be rearranged to arrive at

kNH2Cl,oxide1Koxide1 )

βoxide ΘK-1 w [OH ]

(12)

Using this equation and the collected data, we calculate a value for kNH2Cl,oxide1Koxide1 of 2.09 (( 0.15 at 95% CI) × 10-4 L m-2 min-1. As discussed in the following section, a value describing Fe(II) surface complexation to goethite in the presence of carbonate (Koxide1) is not currently available, we therefore report a mixed rate coefficient that is applicable over the pH range of 5.8-7.7. At either higher or lower pH values, additional reaction terms may be necessary. Interestingly, in contrast to the variable reaction stoichiometry (Θ) previously observed for the homogeneous and autocatalytic reactions (10), in the presence of goethite we calculated (based on measurements of monochloramine concentrations before and after the reactions with Fe(II) were complete) a constant stoichiometry of 1.23 (( 0.02 at the 95% confidence level) (mM Fe(II) oxidized:mM NH2Cl reduced) over the entire pH range of 5.80-7.71. Although competitive scavenging reactions involving the amidogen radical do occur when goethite is present (as evidenced by a stoichiometry less than 2:1), the kinetics of the Fe(II)NH2Cl reactions appear to be enhanced such that discerable differences in radical scavenging over this pH range are elimininated. As shown in Figure 5, the kinit/[NH2Cl]0 values for the aqueous-phase Fe(II)-NH2Cl reactions (10) are consistently 0.25-0.5 log unit lower than those obtained with goethite present. The higher rates in the presence of goethite must be sufficient to maintain a stoichiometry of 1.23 even though the aqueous-phase stoichiometry varied from 1.07 to 1.54 over this range. Further work to clarify the reasons behind this observed effect is required to better examine this interesting issue. Effect of Carbonate Concentration on Fe(II) Complexation by Goethite. Previous studies have examined the surface complexation of Fe(II) at a variety of iron oxide surfaces (13, 16, 20, 21). From these studies, equilibrium constants for the formation of the Fe(II) surface complexes >FeOFe+ and >FeOFeOH have been determined. Unfortunately, these equilibrium constants were obtained under conditions where carbonate was excluded from the system; therefore, their applicability in waters containing measurable levels of carbonate is unknown. Fe(II) surface complexation in the presence of carbonate was investigated by measuring sorption isotherms for Fe(II) on goethite for four different total carbonate concentrations: no carbonate and CT,CO3 ) 2.7, 5.7, and 11.7 mM (Figure 7). As the total carbonate concentration increased from 0 to 11.7 mM, the sorption edge for Fe(II) shifted from around pH 6 (CT,CO3 ) 0 mM) to pH values ranging from 7 to 8 (CT,CO3 ) 2.7-11.7 mM). In accord with the previously discussed Np(V) (26) and Pb(II) (28) studies, this shift presumably occurs because the carbonate-Fe(II) complexes do not bind at the oxide surface as readily as the aquo or hydroxo complexes of Fe(II). To better understand this behavior, the collected sorption isotherms were evaluated using the triple-layer model (40) as contained in the chemical speciation program FITEQL 3.2 (41). For this preliminary evaluation, equilibrium constants describing solution and surface complexation were obtained from the literature (Table 3).

FIGURE 6. Adsorption of Fe(II) () 185 µM) onto goethite () 10.65 m2/L) as a function of the carbonate concentration. µ ) 0.1 M (NaClO4).

FIGURE 7. Fe(II) adsorption isotherms for ferrihydrite (16), goethite (16), hematite (16), lepidocrocite (20), and magnetite (12). For all experiments, the total number of surface sites was in excess of the total Fe(II) concentration. In the absence of carbonate the Fe(II) sorption isotherm was well fit using the literature values for the complexes >FeOFe+ and >FeOFeOH of Liger et al. (16; Table 3 and Figure 6). Site densities smaller than 10 sites/nm2 were unable to account for the total sorption of Fe(II) to the surface; therefore, a site density of 13 sites/nm2 as obtained elsewhere was employed (42). Our ability to fit the data obtained in the absence of carbonate suggests that the goethite used in this study exhibits similar sorptive capabilities as that studied elsewhere. In the presence of carbonate, however, the incorporation of terms describing the formation of aqueous Fe(II)-carbonate (22) and carbonate-goethite (30) complexes into the model (Table 3) was insufficient to account for the observed effects. This result suggests that surface complexes in addition to those listed in Table 3 (e.g., ternary complexes involving both Fe(II) and carbonate) may need to be considered to fully model this system. Although our attempts to model the effect of carbonate on Fe(II) surface complexation were unsuccessful, the results shown in Figure 6 indicate that for goethite at pH ≈6.9 in a 5.7 mM carbonate buffer that only 5-6% of the available of Fe(II) can complex at the surface. This value is considerably less than the 90-100% that would be predicted based on the sorption isotherm obtained in the absence of carbonate. Translating these results to the other oxides examined in this study, we anticipate that considerably less Fe(II) was VOL. 36, NO. 3, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

517

TABLE 3. Equilibria Used in Triple-Layer Surface Complexation Modela aqueous complexes OH-

H2O a H+ + Fe2+ + H2O a FeOH+ + H+ Fe2+ + 2H2O a Fe(OH)2 + 2H+ H2CO3 a H+ + HCO3HCO3- a H+ + CO32Fe2+ + HCO3- a FeHCO3+ Fe2+ + CO32- a FeCO30 Fe2+ + 2CO32- a Fe(CO3)2Fe2+ + CO32- + OH- a Fe(CO3)(OH)surface complexes

log β

ref

-13.775 -9.7 -20.6 -6.301 -10.23 1.40 5.50 7.26 9.82

16 16 16 22b 22b 22b 22b 22b 22b

log β

ref

>FeOH + H+ a >FeOH2+ 7.47 16 >FeOH a >FeO- + H+ -9.51 16 >FeOH + Fe2+ a >FeOFe+ + H+ 0.11 16 >FeOH + Fe2+ + H2O a >FeOFe+ + 2H+ -7.64 16 >FeOH + CO32- + H+ a >FeO-0.2COO-0.8 + H2O 12.76 30 >FeOH + CO32- + 2H+ a >FeOCOOH + H2O 18.29 30 a Goethite ) 10.65 m2/L, inner layer capacitance ) 1.0 F/m2, outer layer capacitance ) 0.2 F/m2, 13 sites/nm2, µ ) 0.1 M (NaClO4). b Interpolated to an ionic strength of 0.1 M.

sorbed to each of the oxides at pH 6.9 than is suggested by the sorption isotherms shown in Figure 7. These isotherms, obtained from the literature, show that Fe(II) sorption as a function of solution pH varies widely between each of the oxides. Because each of the oxides could exhibit different affinities for carbonate sorption (26), the net effect of carbonate on each of these Fe(II) sorption isotherms is unknown and requires additional evaluation. Comparing the Catalytic Activity of Iron Oxides. Comparing the calculated βoxide values to one another indicates that at pH ≈6.9 in a 5.7 mM carbonate buffer that the catalytic activities of the oxides can be ranked in the following order: magnetite > goethite > lepidocrocite ≈ hematite > ferrihydrite. This differential reactivity is presumably the result of variations in the total amount of Fe(II) sorbed to each oxide surface (15) and differences in the site densities of each oxide (17). Unfortunately, because the effects of carbonate on Fe(II) complexation by each of the oxides have yet to be fully evaluated, it is impossible to say which of these effects plays the largest role in this system. The results presented in this work have illustrated that various iron oxide surfaces exhibit different levels of catalytic activity toward the Fe(II)-NH2Cl reactions. This observation implies that the composition of the deposits found in distribution systems may dictate the importance of surface catalyzed monochloramine loss and iron oxidation. Accordingly, in distribution systems characterized by deposits coated with magnetite, surface-catalyzed reactions may be of greater importance than in distribution systems coated with ferrihydrite precipitates. The trends observed here may also apply to a variety of systems (e.g., in reduced natural environments or in cast iron permeable reactive barriers) where Fe(II) has the potential to sorb onto iron oxide surfaces. These results also suggest that the presence of carbonate can dramatically affect metal ion complexation at the iron oxide surface and that these effects may play a dramatic role in many natural and engineered systems. Additional studies further characterizing the relationship between carbonate concentration and metal ion surface complexation are required to better understand many of the complexities inherent in this system. Nevertheless, the results presented here indicate that the role of carbonate needs to be better addressed in many environmental systems. 518

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 3, 2002

Acknowledgments We thank Michael Shoup for his help with the experimental procedures and for helping synthesize the iron oxides. The insightful comments of three anonymous reviewers greatly augmented the clarity and presentation of this work. This material presented herein was supported by the Abel Wolman Doctoral Fellowship of the American Water Works Association (AWWA) and by a grant from the American Water Works Association Research Foundation (AWWARF).

Supporting Information Available Supplemental data and additional model simulations. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Rossman, L. A.; Brown, R. A.; Singer, P. C.; Nuckols, J. R. Water Res. 2001, 35, 3483-3489. (2) McNeill, L. S.; Edwards, M. J. Am. Water Works Assoc. 2001, 93, 88-100. (3) Benjamin, M. M.; Sontheimer, H.; Leroy, P. In Internal Corrosion of Water Distribution Systems, 2nd ed.; AWWARF and DVGW: Denver, CO, 1996; pp 29-70. (4) Sarin, P.; Snoeyink, V. L.; Bebee, J.; Kriven, W. M.; Clement, J. A. Water Res. 2001, 35, 2961-2969. (5) Tuovinen, O. H.; Mair, D. M.; Banovic, J. Environ. Technol. Lett. 1984, 5, 97-108. (6) Tuovinen, O. H.; Button, K. S.; Vuorinen, A.; Carlson, L.; Mair, D. M.; Yut, L. A. J. Am. Water Works Assoc. 1980, 72, 626-635. (7) Hackett, M. A. Master’s Thesis, University of Iowa, Iowa City, IA, 1997. (8) Norton, C. D.; LeChevallier, M. W. J. Am. Water Works Assoc. 1997, 89, 66-77. (9) Vikesland, P. J.; Valentine, R. L. Environ. Sci. Technol. 2000, 34, 83-90. (10) Vikesland, P. J.; Valentine, R. L. Environ. Sci. Technol. (in press). (11) Sørensen, J.; Thorling, L. Geochim. Cosmochim. Acta 1991, 55, 1289-1294. (12) Klausen, J.; Tro¨ber, S. P.; Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci. Technol. 1995, 29, 2396-2404. (13) Charlet, L.; Liger, E.; Gerasimo, P. J. Environ. Eng. (N.Y.) 1998, 124, 25-30. (14) Haderlein, S. B.; Pecher, K. In Mineral-Water Interfacial Reactions; Sparks, D. L., Grundl, T. J., Eds.; American Chemical Society: Washington, DC, 1998; pp 342-357. (15) Buerge, I. J.; Hug, S. J. Environ. Sci. Technol. 1999, 33, 42854291. (16) Liger, E.; Charlet, L.; Van Cappellen, P. Geochim. Cosmochim. Acta 1999, 63, 2939-2955. (17) Amonette, J. E.; Workman, D. J.; Kennedy, D. W.; Fruchter, J. S.; Gorby, Y. A. Environ. Sci. Technol. 2000, 34, 4606-4613. (18) Strathmann, T. J. Ph.D. Dissertation, The Johns Hopkins University, Baltimore, MD, 2001. (19) Luther, G. W., III. In Aquatic Chemical Kinetics; Stumm, W., Ed.; Wiley-Interscience: New York, 1990; pp 173-198. (20) Zhang, Y.; Charlet, L.; Schindler, P. W. Colloids Surf. 1992, 63, 259-268. (21) Coughlin, B. R.; Stone, A. T. Environ. Sci. Technol. 1995, 29, 2445-2455. (22) King, D. W. Environ. Sci. Technol. 1998, 32, 2997-3003. (23) King, D. W.; Farlow, R. Mar. Chem. 2000, 70, 201-209. (24) van Geen, A.; Robertson, A. P.; Leckie, J. O. Geochim. Cosmochim. Acta 1994, 58, 2073-2086. (25) Su, C.; Suarez, D. L. Clays Clay Miner. 1997, 45, 814-825. (26) Kohler, M.; Honeyman, B. D.; Leckie, J. O. Radiochim. Acta 1999, 85, 33-48. (27) Bargar, J. R.; Reitmeyer, R.; Davis, J. A. Environ. Sci. Technol. 1999, 33, 2481-2484. (28) Ostergren, J. D.; Trainor, T. P.; Bargar, J. R.; Brown, G. E., Jr.; Parks, G. A. J. Colloid Interface Sci. 2000, 225, 466-482. (29) Villalobos, M.; Leckie, J. O. Geochim. Cosmochim. Acta 2000, 64, 3787-3802. (30) Villalobos, M.; Leckie, J. O. J. Colloid Interface Sci. 2001, 235, 15-32. (31) Villalobos, M.; Trotz, M. A.; Leckie, J. O. Environ. Sci. Technol. 2001, 35, 3849-3856.

(32) Stumm, W.; Morgan, J. J., Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd ed.; John Wiley and Sons: New York, 1996. (33) Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory: Preparation and Characterization; VCH: Weinheim, Germany, 1991. (34) Vikesland, P. J.; Ozekin, K.; Valentine, R. L. Environ. Sci. Technol. 1998, 32, 1409-1416. (35) Roden, E. E.; Zachara, J. M. Environ. Sci. Technol. 1996, 30, 1618-1628. (36) Vikesland, P. J. Ph.D. Dissertation, The University of Iowa, Iowa City, IA, 1998. (37) Valentine, R. L.; Jafvert, C. T. Environ. Sci. Technol. 1988, 22, 691-696. (38) Vikesland, P. J.; Ozekin, K.; Valentine, R. L. Water Res. 2001, 35, 1766-1776. (39) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1988, 17, 1027-1284.

(40) Davis, J. A.; Leckie, J. O. J. Colloid Interface Sci. 1978, 67, 90107. (41) Herbelin, A. L.; Westall, J. C. FITEQL. A Computer Program for Determination of Chemical Equilibrium Constants from Experimental Data; Department of Chemistry, Oregon State University: Corvallis, OR, 1996. (42) Davies, S. H. R.; Morgan, J. J. J. Colloid Interface Sci. 1989, 129, 63-77. (43) Liang, L.; Morgan, J. J. Aquat. Sci. 1990, 52, 32-55. (44) Cornell, R. M.; Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses; VCH: Weinheim, Germany, 1996.

Received for review May 2, 2001. Revised manuscript received August 28, 2001. Accepted September 27, 2001. ES010935V

VOL. 36, NO. 3, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

519