Effect of Strongly Competing Background Compounds on the Kinetics

Feb 21, 2008 - Division of Environmental Science and Engineering, Faculty of Engineering, National University ... University of Illinois at Urbana-Cha...
0 downloads 0 Views 1011KB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 2606–2611

Effect of Strongly Competing Background Compounds on the Kinetics of Trace Organic Contaminant Desorption from Activated Carbon P R I S C I L L A C . T O , †,‡ B E N I T O J . M A R I Ñ A S , * ,‡ V E R N O N L . S N O E Y I N K , †,‡ A N D WUN JERN NG§ Division of Environmental Science and Engineering, Faculty of Engineering, National University of Singapore, Singapore 117576, Singapore, Department of Civil and Environmental Engineering and Center of Advanced Materials for the Purification of Water with Systems, University of Illinois at Urbana-Champaign, 205 N. Mathews Avenue, Urbana, Illinois 61801, and Nanyang Environment and Water Research Institute (NEWRI), 50 Nanyang Avenue, Block N1-B3b-29, Singapore 639798, Singapore

Received October 16, 2007. Revised manuscript received January 2, 2008. Accepted January 3, 2008.

Strongly competing (SC) compounds, naturally found in any drinking water source, are known to decrease the adsorption capacity of activated carbon for trace contaminants. While the effect of these substances on the capacity and adsorption kinetics of trace contaminants is fairly well studied, relatively little is known about their impact on desorption kinetics. The purpose of this study was to investigate the relationship between SC matter and trace compound desorption kinetics. A surrogate SC compound, 1,4-dichlorobenzene (p-DCB), was used to displace the preadsorbed target trace contaminant, atrazine, from powdered activated carbon (PAC). The initial concentrations of p-DCB and atrazine were varied to achieve different degrees of competition to atrazine. Atrazine’s desorption diffusion coefficient was found to increase with increasing adsorbed concentration of the SC matter, expressed as an equivalent background compound (EBC). The EBC was modeled with atrazinelike adsorption properties, thus representing the portion of p-DCB that competed to occupy atrazine adsorption sites. The increase in atrazine diffusion rate can be explained by a shift from surface diffusion to diffusion through the carbon’s pores as the availability of surface sites decreased due to the EBC’s competition. The observed desorption kinetic relationship was consistent with the effect of SC competition on adsorption kinetics; further, the effect was consistent for three different types of SC matter. These findings highlight that the impact of SC matter on activated carbon applications could be either detrimental (displacing adsorbed trace contaminants and enhancing their rate of release) or beneficial (offsetting pore constriction effects by enhancing their rate of uptake). * Corresponding author phone: +1-217-333-6961; fax: +1-217333-6968; e-mail: [email protected]. † National University of Singapore. ‡ University of Illinois at Urbana-Champaign. § NEWRI. 2606

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008

Introduction The phenomenon of trace compound desorption from activated carbon has not been well-studied, even though it is probably a common occurrence in adsorption systems in which the adsorbent residence time is much longer than the hydraulic residence time. The continuous supply of new competing compounds in the water matrix can cause continual displacement of previously adsorbed trace contaminants, resulting in shorter adsorbent life. Desorption of compounds such as endocrine disruptors, pharmaceuticals, and other such chemicals affecting the human nervous system could have potentially adverse health effects, especially when the adsorption system effluent is not monitored for their presence. The findings of past adsorption/desorption studies relating to the effect of competing solutes on the kinetics of a trace contaminant suggest that the complex mixture of solutes in natural organic matter (NOM) has the potential to either increase or decrease internal diffusion rate. Carter et al. (1) hypothesized that preadsorbed NOM caused pore blockage, and likely hindered a target contaminant’s diffusion into carbon pores. Studies by Pelekani and Snoeyink (2) and Li et al. (3) have described the ability of NOM to hinder internal diffusion and thus slow the uptake rate of a trace contaminant. Pelekani and Snoeyink (2) used activated carbon fibers with several different narrowly defined pore size ranges to demonstrate that pore constriction and blockage by preadsorbed Congo Red decreased atrazine adsorption rate. Li et al. (3) divided NOM into two fractions: a strongly competing (SC) portion of similar size to atrazine that competes for atrazine adsorption sites, and a pore-blocking (PB) fraction of larger molecules that hinders the uptake rate of atrazine. As the adsorbed concentration of PB NOM increased, the surface diffusion coefficient of atrazine adsorption decreased (4). The three-component (trace compound, SC NOM, and PB NOM) model developed by Li et al. (3) provided good predictions for adsorption kinetics in bench-scale continuous flow PAC-ultrafiltration tests over a period of 15 h. On the other hand, some studies point to the ability of competing solutes to accelerate the diffusion rate of a target compound. In the batch kinetic tests of Li et al. (5), the diffusion coefficient of atrazine adsorption in the presence of p-DCB (147 Da) was larger than that of atrazine adsorption alone. This molecule, of similar size to atrazine, was not expected to affect atrazine adsorption kinetics. As a result, Li et al. (5) concluded that these results could have been due to slight errors in parametrization. Several soil studies have attempted to explain the increased desorption kinetics of a target contaminant observed in the presence of competing solutes. White and Pignatello (6) found that the desorption rate of the target solute, phenanthrene, increased with increasing concentrations of the competing solute, pyrene. According to the dualmode model of soil organic matter, the authors hypothesized that soil consisted of two domains: a hole-filling domain (which might be likened to adsorption sites on the activated carbon pore surface) and a diffusion domain (which might be likened to the pores themselves). They explained that it was the filling of holes by pyrene that prevented phenanthrene from adsorbing, and this increased the diffusion rate of phenanthrene through the diffusion domain. Following that study, Zhao et al. (7) created a mathematical dual-mode diffusion model and were able to achieve good predictions of the phenanthrene desorption results of White and Pignatello (6). They found that the diffusion coefficient of 10.1021/es702609r CCC: $40.75

 2008 American Chemical Society

Published on Web 02/21/2008

phenanthrene desorption was concentration dependent, such that as the concentration of one solute increased, more holes were filled, and thus there were less available holes to retard the diffusion of a second compound. The similarity of the phenomena observed in soil studies to those involving activated carbon applications is uncertain, since soil and activated carbon differ in composition and structure. Another factor to consider concerning bisolute kinetics is adsorbability. The work of Braida et al. (8) demonstrated that a compound’s apparent diffusivity increased as its affinity for the soil surface decreased. Sontheimer et al. (9) initially assumed that in simultaneous adsorption, solutes do not affect each other in mass transfer. Under this assumption, they compared adsorption predictions made using singlesolute diffusion coefficients to experimental data for the simultaneous adsorption of p-nitrophenol and phenol. p-Nitrophenol adsorption kinetics could be well predicted, but the kinetics of phenol was faster than the surface diffusion model predicted. They hypothesized that phenol, the weaker adsorbing compound, could be more mobile because competitive interactions drove it to occupy sites with lower bonding forces. In other words, p-nitrophenol preferentially adsorbed to surface sites, almost as if it were the only solute. Phenol had faster kinetics as it continued to diffuse toward the center of the particle to adsorption sites where pnitrophenol was not yet present. Besides solubility or functionality, other causes of weaker adsorbability could be greater site competition from other compounds, or decreasing affinity for the surface when approaching higher concentrations. Qi et al. (10) demonstrated that the intraparticle diffusivity of a single-solute such as atrazine in activated carbon was not a function of its own concentration on the adsorbent at aqueous concentrations at which this herbicide would be encountered in most natural waters. However, other studies such as those by Sudo et al. (11) and Suzuki and Fujii (12) have found that the intraparticle diffusivity of single compounds tested at higher aqueous concentrations could increase with their increased surface loading on activated carbon thus supporting that a relatively high surface coverage of the adsorbent could affect the solute intraparticle diffusivity. The purpose of this study was to investigate the effect of strongly competing background organic solutes on the rate of trace contaminant desorption from activated carbon, and to examine whether the effect also extends to kinetics of adsorption. Elucidating this effect is essential to understanding the overall adsorption/desorption phenomena and incorporating considerations of desorption into improving and optimizing the design of adsorption systems.

Materials and Methods Water. Milli-Q ultrapure water (Millipore, Billerica, Massachusetts) was adjusted to pH 7 with 1 mM phosphate buffer. Adsorbates. Radio-labeled atrazine (Syngenta Crop Protection AG, Basel, Switzerland) with a specific radioactivity of 52.25 µCi/mg was used as the target trace contaminant. The atrazine was received in acetone, which was evaporated using a direct stream of nitrogen gas. The solid atrazine residue was dissolved in ultrapure water. The resulting stock solution was subsequently analyzed by size exclusion chromatography (HPLC-SEC) to confirm that acetone was only present at trace levels. The atrazine manufacturer reported that the initial radiochemical purity was 98.6% and chemical purity was 99.4%. Chemical purity was measured periodically using HPLC (area distribution). After two months, chemical purity was 99.2%. After seven and fourteen months, purity was consistently 98.0%. Two comparative kinetic tests were conducted using a second batch of atrazine (Sigma, St. Louis, Missouri) with specific radioactivity of 43.12 µCi/mg. Received in solid form, the stock solution was prepared by dissolving

the atrazine in ultrapure water. The reported initial radiochemical purity was 98.2%. After all experiments conducted with this batch of atrazine were completed, the measured chemical purity was 98.8%. The adsorbate used to represent the SC component of NOM was p-DCB (Sigma, St. Louis, Missouri). Li et al. (5) showed that p-DCB directly competes for the carbon surface but does not hinder atrazine adsorption rate by pore blockage, thus enabling investigation of the SC effect on atrazine kinetics, independent of the PB influence. A p-DCB stock solution of 10 000 mg/L was prepared in methanol and stored at 4 °C. Additionally, NOM from Suwannee River, Georgia (International Humic Substances Society, St. Paul, Minnesota), isolated by reverse osmosis, was used to demonstrate competition from natural organic compounds. Received in dry form, the NOM was dissolved in ultrapure water. The resulting solution was filtered with a 0.45 µm nylon filter to remove suspended solids. Sodium azide at 0.01% w/w was added to prevent biodegradation, and the stock solution was stored at 4 °C. The stock concentration measured 164.4 mg/L as organic carbon. Subsequent solutions of NOM were prepared by diluting the stock concentration to the desired NOM concentration, and maintaining the 0.01% sodium azide concentration. Adsorbent. A commercial PAC, Norit SAUF (NORIT France, S.a.r.l., Le Blanc Mesnil Cedex, France) was dried at 105 °C overnight, cooled in a desiccator, then weighed. The properties of this carbon can be found elsewhere (5). Analytical Methods. All collected samples were filtered through a 0.45 µm nylon syringe filter to remove the PAC. For atrazine analysis, samples of 2.5 mL were combined with 18 mL of liquid scintillation cocktail (Ecoscint, National Diagnostics, Inc. Atlanta, Georgia). The samples were then analyzed with a liquid scintillation counter (Beckman LS6500, Fullerton, California) to measure radioactivity. For experiments with p-DCB, duplicate samples of 1.6 mL were collected in 2 mL GC autosampler vials. Immediately, 0.4 mL of hexane was added and the vial was capped. Each sample was analyzed in triplicate using GC-MS (Shimadzu GC-2010 and GCMS-QP2010, Tokyo, Japan). For the experiments involving NOM, 16 mL samples were collected, and NOM concentration was measured by nonpurgeable organic carbon (NPOC) combustion oxidation analysis (Shimadzu TOC-Vcsh, Tokyo, Japan) and expressed in mg/L of NPOC. Displaced Desorption Kinetic Tests. The steps in this series of experiments were to preload and equilibrate the PAC with atrazine, and then to add p-DCB and measure the desorption rate of atrazine. Experimental tests with variable initial p-DCB concentration were performed with six batch reactors, each with 5 mg/L of PAC and 100 µg/L of atrazine in buffered ultrapure water. The bottles were placed on an orbital shaker for continuous mixing. On day 7, a sample was taken to check for adsorption equilibrium. On day 8, p-DCB was introduced to each of the six bottles at initial concentrations of 0.5, 1.0, 1.5, 2.0, 2.5, and 3.0 mg/L. Under continuous stirring, atrazine displacement was measured for the next four hours. The bottle was sealed except for a valved sampling port and a valved Tedlar bag with nitrogen gas to replenish sampled volume. Bottle headspace was also minimized to prevent volatilization of p-DCB. In experimental tests with variable initial atrazine concentration, 5 mg/L of PAC was first contacted with 50 and 150 µg/L of atrazine. After mixing for 7 days, the solutions were checked for adsorption equilibrium. On day 8, 2 mg/L of p-DCB was introduced into each bottle and with continuous stirring, atrazine displacement was measured for four hours. In separate but similar tests, 4 mg/L and 6 mg/L PAC were preloaded with solutions containing 84.6 µg/L and 104 VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2607

µg/L atrazine, respectively, and mixed continuously on an orbital shaker. Atrazine equilibrium was checked on day 11. On day 12, p-DCB was spiked to 2 mg/L, and subsequent atrazine displacement was measured for four hours. Samples were taken in all batch displaced desorption kinetic tests to measure p-DCB concentration throughout the 4-h desorption period. Nondisplaced Desorption Kinetic Tests. One batch test was conducted in which 30 mg/L of PAC was contacted with 3.56 mg/L of atrazine in buffered ultrapure water and mixed continuously on an orbital shaker. Atrazine equilibrium was checked on day 11. On day 12, the solution was diluted to a PAC concentration of 6 mg/L using ultrapure water. (At the new volume, the equivalent initial atrazine concentration was 711 µg/L.) In this manner, the concentration gradient to drive desorption was created by decreasing the atrazine concentration in the bulk solution. Atrazine desorption was measured until a steady concentration was reached. A high initial atrazine concentration was selected to yield a high adsorbed concentration at equilibrium, ensuring that subsequent desorption would be sufficient for measurement. Adsorption Kinetic Tests. To measure the kinetics of atrazine adsorption alone, a solution with 102 µg/L of atrazine was contacted with 10 mg/L of PAC in a batch reactor. In two other tests conducted with a competing solute, 10 mg/L of PAC was simultaneously contacted with 101 µg/L of atrazine and 2 mg/L of p-DCB, and 4 mg/L of PAC was simultaneously contacted with 100 µg/L of atrazine and 3.8 mg/L of NOM. In all three tests, atrazine concentration was measured over four hours. Batch Isotherm Tests. Bottle-point isotherm tests were not conducted for all initial conditions, but only for 100 µg/L of atrazine with 0, 0.5, and 2 mg/L of p-DCB or 3.8 mg/L of NOM. Solute concentrations were measured after 7 days, and the concentrations after 14 days confirmed that equilibrium was attained by day 7.

FIGURE 1. Experimental and HSDM-fitted displaced desorption of (a) atrazine (C0 ) 100 µg/L) by p-DCB at various initial concentrations (C0 ) 0.5-3.0 mg/L) and (b) atrazine at three initial concentrations (C0 ) 50-150 µg/L) by p-DCB (C0 ) 2 mg/ L), with corresponding surface diffusion coefficients. Center set of data in (b) is repeated from (a) for comparison.

Results and Discussion Experimental results and fitted atrazine diffusion coefficients for the displaced desorption tests with various p-DCB and atrazine concentrations are displayed in Figure 1. Figure 2a shows similar results from the displaced desorption experiments with different PAC doses. The surface diffusion coefficient, Ds, of atrazine for each test was determined by fitting the corresponding data with the homogeneous surface diffusion model (HSDM). The HSDM was solved under a pseudosingle-solute assumption, in which the p-DCB competition causing the carbon’s equilibrium adsorption capacity for atrazine to decrease was described as a reduction of atrazine’s Freundlich isotherm adsorption capacity parameter, K. The reduced equilibrium parameter was determined from the equilibrated kinetic tests. The HSDM assumes a concentration-independent Ds, uniform adsorption of the preadsorbed compound throughout the particle, instantaneous equilibrium at the external surface of the particle, negligible external diffusion resistance, and physical adsorption without reaction. A nondisplaced desorption test was conducted to find atrazine’s desorption diffusion coefficient under no competition from background organic matter. In the absence of p-DCB, atrazine desorbed at a slower rate, reaching equilibrium only after 500 min (Figure 2b). It was confirmed that equilibrium was indeed reached by comparing the atrazine concentration after 3000 min to the single-solute isotherm data. The Ds of desorption without p-DCB competition was about 1 order of magnitude smaller than the Ds of desorption with p-DCB competition. Before each kinetic test, adsorption of the preloaded atrazine was confirmed to have reached equilibrium based 2608

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008

FIGURE 2. Experimental and HSDM-fitted (a) displaced desorption of atrazine (C0 ) 84.6 µg/L with PAC dose Cc ) 4 mg/L; C0 ) 104 µg/L with Cc ) 6 mg/L) by p-DCB (C0 ) 2.0 mg/ L) and (b) nondisplaced desorption of atrazine by expanding system volume from an initial PAC dose of Cc ) 30 to Cc ) 6 mg/L, with corresponding surface diffusion coefficients. on comparison with single-solute isotherm data. For two of the desorption kinetic tests in Figure 1a (C0,p-DCB ) 0.5 mg/L and 2.0 mg/L), Figure 3 plots the adsorbed and solution concentrations of atrazine two hours after p-DCB was added to desorb atrazine. The two-hour kinetic test data were in

C0,i ) Ceq,i + qiCc with i ) 1, 2

(2)

Figure 4 plots the relationship between the apparent diffusion coefficients for atrazine desorption and the surface concentration of the EBC, qEBC. The linear fit in Figure 4 takes the following form: Ds ) AqEBC + Ds,0 A ) 1.0 × 10-12 (cm2/min)(µg/mg PAC)-1 Ds,0 ) 4.8 × 10-11 cm2/min

FIGURE 3. Atrazine concentrations two hours after p-DCB was added to desorb atrazine in the kinetic tests, compared against seven-day single-solute and competitive atrazine isotherms with fresh PAC (C0,Atrazine ) 101 µg/L with C0,p-DCB ) 0 mg/L; C0, Atrazine ) 101 µg/L with C0,p-DCB ) 0.5 mg/L; C0,Atrazine ) 114 µg/ L with C0,p-DCB ) 2.0 mg/L). close agreement with independently determined seven-day isotherm data from tests with the same initial conditions. Thus, the kinetic tests for other concentrations of p-DCB were also assumed to have reached equilibrium when the tests were terminated. Increasing Rate of Desorption with Increasing SC Competition. The experimental conditions along with the fitted Ds for the desorption kinetic tests are listed in Table 1. In general, the Ds of atrazine desorption increased with increasing p-DCB initial concentration or decreasing atrazine initial concentration. As p-DCB adsorbed, it displaced adsorbed atrazine, and also decreased the surface sites available for the surface diffusion of atrazine. With restricted surface availability, it is likely that atrazine shifted from slower transport along the surface to faster transport through the pore liquid, thereby increasing its diffusion coefficient. The p-DCB was modeled as an equivalent background compound (EBC) so that the kinetic tests performed with different sources of SC matter could be later compared. The EBC method is designed to simplify a mixture of background compounds, like those in NOM, and model them as a single competing compound (13). This method allows one to bypass the complex individual properties of the competing solute(s), considering only the effect of the competing molecule on decreasing the target compound’s equilibrium adsorption capacity. The ideal adsorbed solution theory (14, 15) for bisolute competitive adsorption equilibrium could then be applied to determine the equilibrium liquid-phase and adsorbed concentrations of the EBC (eq 1). Subscripts 1 and 2 represent atrazine and EBC, respectively. Because only a portion of p-DCB can be assumed to be competing directly with atrazine for the same adsorption sites, the EBC could be simply modeled as a compound with the same equilibrium parameters as those of atrazine, in keeping with the approach proposed by Ding et al. (16). Thus, the EBC was assigned the same Freundlich K and 1/n parameters as atrazine. Using these parameters for p-DCB instead of the Freundlich parameters from the p-DCB isotherm disregards the amount of p-DCB that adsorbs in pores too small for atrazine to enter, taking into account only the amount that directly competes with atrazine. Equations 1 and 2 were then solved for the initial (C0,EBC) and equilibrium concentrations (Ceq,EBC and qEBC) of the EBC that would yield the equilibrium liquid and adsorbed atrazine concentrations known from each kinetic experiment. The qEBC values are listed in Table 1.

( (

q1 q 1 + q2 q2 Ceq,2 ) q 1 + q2

Ceq,1 )

)[ )[

n1q1 + n2q2 n1K1 n1q1 + n2q2 n2K2

] ]

n1

n2

(1)

(3)

where Ds,0 is the diffusion coefficient of the trace compound under no competition (or qEBC ) 0) and A represents the rate at which Ds increases with increasing qEBC. The parameter qEBC expresses the surface coverage of SC matter competing directly for atrazine adsorption sites. The increase in atrazine Ds is consistent with a shift from surface diffusion to pore diffusion. When more adsorption sites were occupied by p-DCB, specifically those sites that atrazine interacted with during diffusion, diffusion through the carbon’s pores could more readily take place. Furthermore, the change in qEBC reflects the relative strength of SC competition. If the competitiveness of SC matter increases, whether due to greater abundance or surface affinity, qEBC should increase accordingly. Notice that the most outlying point in Figure 4 corresponds to the data set at the lowest initial atrazine concentration from Figure 1b (Ds ) 6.8 × 10-10 cm2/min) for which the kinetic portion of the curve is not well characterized. Furthermore, the curve (not shown) corresponding to the lower Ds value obtained from applying eq 3 provided an equally satisfactory fitting of this data set. By expressing Ds as a function of the equilibrium parameter, qEBC, it was implicity assumed that the SC matter reached equilibrium before atrazine. Figure 5 illustrates the rapid adsorption of p-DCB during the atrazine displaced desorption tests (Figure 1). Whereas atrazine generally required more than 60 min to reach equilibrium, p-DCB neared equilibrium within 10 min. Indeed, the atrazine data showed that atrazine desorption began immediately upon p-DCB addition at time zero. Furthermore, the fitted Ds values for p-DCB adsorption were larger than those for atrazine desorption by a factor of about 3-10. Therefore, in the case of p-DCB, it was acceptable to use the equilibrium qEBC to represent the surface coverage affecting atrazine diffusion. Comparison of Adsorption Kinetics with Desorption Kinetics. The results from atrazine adsorption without competition and with simultaneous p-DCB competition are displayed in Figure 6. The data points displayed in Figure 4 for these adsorption experiments generally agree well with desorption kinetic results. The diffusion coefficients from the atrazine adsorption kinetic tests of Li et al. (5, 17) are also plotted in Figure 4. In that study, the atrazine diffusion coefficient for simultaneous adsorption with polystyrene sulfonate (PSS; nominal peak molecular weight ) 1800 Da) or p-DCB showed the same increasing trend with increasing qEBC. (Li et al. (5) showed that simultaneously adsorbed PSS did not hinder atrazine adsorption kinetics as preadsorbed PSS did, so the pore-blocking (PB) effect of PSS molecules could be neglected.) Those authors observed that the Ds values of atrazine under competitive adsorption were higher than for atrazine adsorption alone, but attributed the observation to experimental errors. However, the present findings indicate that their results were consistent with the faster diffusion rate resulting from surface coverage by SC matter. Thus, desorption and adsorption kinetics demonstrated the same dependence on the SC surface competition. Atrazine kinetics during simultaneous adsorption with Suwannee River NOM also fit well with the relationship shown in Figure 4. This agreement reinforces that (1) similar to PSS adsorption, the PB fraction of NOM did not hinder atrazine VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2609

TABLE 1. Summary of Experimental Conditions for Desorption and Adsorption Kinetic Tests with Corresponding Surface Diffusion Coefficients atrazine C0, µg/L

PAC Cc, mg/L

SC compound

SC C0, mg/L

99.8 99.8 100 100 100 100 50.1 151 84.6 104 711 102 101 100 100 96.9 99.9

5 5 5 5 5 5 5 5 4 6 6 10 10 4 not given 4 5

p-DCB p-DCB p-DCB p-DCB p-DCB p-DCB p-DCB p-DCB p-DCB p-DCB

0.5 1.0 1.5 2.0 2.5 3.0 2.0 2.0 2.0 2.0

p-DCB NOM

2.0 3.8

PSS p-DCB

10.0 2.0

a

qEBC after equilibrium, µg EBC/mg PAC 70.8 131 185 262 309 379 220 263 249 145 0 0 125 83.8 0 10.6 145

desorption Ds, cm2/min 1.8 × 10-10 1.7 × 10-10 2.4 × 10-10 4.3 × 10-10 3.9 × 10-10 4.0 × 10-10 6.8 × 10-10 2.9 × 10-10 2.2 × 10-10 1.4 × 10-10 2.7 × 11-10

adsorption Ds, cm2/min

4.0 × 11-10 3.2 × 10-10 1.9 × 10-10 6.3 × 11-10a 6.6 × 11-10a 1.8 × 10-10a

Experiments from previous studies (see refs 5 and 17) performed with the same activated carbon.

FIGURE 6. Atrazine simultaneous adsorption with NOM competition (b), p-DCB competition (1), and no competition (9) with corresponding surface diffusion coefficients.

FIGURE 4. Surface diffusion coefficient of atrazine desorption and adsorption versus adsorbed concentration of the SC compound, with the source of SC competition shown in parenthesis. * Data taken from previous studies (refs 5 and 17).

FIGURE 5. Adsorption of p-DCB during batch displacement of atrazine, with corresponding surface diffusion coefficients. kinetics during simultaneous adsorption and (2) the increased atrazine kinetics under NOM competition gave evidence that, similar to the rapid uptake of p-DCB, the SC fraction of NOM diffused with equal or faster kinetics compared to that of atrazine. Thus, the occurrence of increased atrazine adsorption or desorption kinetics were consistent with three different sources of SC molecules (p-DCB, PSS, and NOM). 2610

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008

Although the faster uptake of trace contaminants during competitive adsorption would be beneficial, the quicker release of displaced trace contaminants would be detrimental to activated carbon performance. This is less significant in batch systems, where the reduction of atrazine capacity is limited by the initial amount of SC matter present, so likewise, the increase in trace contaminant kinetics is limited. However, in continuous flow systems where the activated carbon residence time is much greater than the hydraulic residence time, the kinetic effect of SC matter demands more attention because SC matter may be continually introduced in the influent. As the adsorption of SC matter increases, it not only displaces more target contaminants, but the desorption rate of the trace contaminants will also increase. Ultimately, this has the consequence of shortening adsorbent usage life. Effect of Total Surface Concentration of Background Competing Matter on Kinetics. Figure 4 showed that Ds depended specifically on qEBC, that is, the coverage by a fraction of molecules competing directly with atrazine. Does the surface coverage of the remaining fraction of background matter contribute to increasing atrazine kinetics? For example, the total adsorbed mass concentrations at equilibrium in the different kinetic tests were approximately 260 mg PSS/g PAC (calculated from equilibrium isotherm parameters given in ref 5), 200 mg NOM/g PAC (assuming NOM carbon composition of 48.8% as reported by the supplier), and 80 mg p-DCB/g PAC (C0 ) 0.5 mg/L). The corresponding Ds values for atrazine were 6.6 × 10-11, 1.9 × 10-10, and 1.8 × 10-10 cm2/min, respectively. The highest mass adsorbed was that of PSS but the corresponding Ds for atrazine was closest

to the Ds under no competition. Moreover, although a greater mass of NOM adsorbed compared to that of p-DCB, the value of Ds for both tests was about the same. Several arguments can be made to explain why the total surface concentration did not correlate closely with the increase in atrazine kinetics. First, larger molecular weight molecules, such as in PSS and NOM, are expected to diffuse more slowly than atrazine (9, 18), so larger molecules would not be present on the surface to affect the adsorption rate of atrazine. The reported Ds of the PB NOM fraction was larger than the Ds of atrazine, but the actual Ds of the PB NOM fraction was probably smaller if a realistic, shorter diffusion distance was considered in the modeling (16). Second, it is generally assumed that the SC NOM fraction and micropollutants like atrazine adsorb primarily in micropores, and that larger molecules can only access and adsorb in wider pores (19). The wider pores could still have contained many unoccupied sites for atrazine to diffuse unimpeded. Third, large molecules of PSS and NOM are higher in mass, but compared to p-DCB, they may occupy less surface adsorption sites per unit of mass. Mass concentration may not be the best parameter for comparing surface coverage of different sized molecules. Fourth, even though larger molecules occupy surface sites, their effect on decreasing atrazine kinetics by constricting pores may be a more dominant opposing effect (2, 5, 20, 21). Further research is needed, particularly on the effect of pore-blocking background matter on desorption kinetics, to better comprehend the full impact desorption may have on activated carbon treatment processes.

Acknowledgments We thank the National Science Foundation Graduate Research Fellowship program, the WaterCAMPWS, a Science and Technology Center of Advanced Materials for the Purification of Water with Systems under NSF agreement number CTS-0120978, the National University of Singapore (NUS) for funding this research, and NUS for generously providing the use of laboratory facilities. The radio-labeled atrazine was donated by Syngenta Crop Production AG (Basel, Switzerland).

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

Literature Cited (1) Carter, M. C.; Weber, W. J., Jr.; Olmstead, K. P. Effects of background dissolved organic matter on TCE adsorption by> GAC. J. Am. Water Works Assoc. 1992, 84 (8), 81–91. (2) Pelekani, C.; Snoeyink, V. L. A kinetic and equilibrium study of competitive adsorption between atrazine and Congo red dye on activated carbon: the importance of pore size distribution. Carbon 2001, 39 (1), 25–37. (3) Li, Q.; Mariñas, B. J.; Snoeyink, V. L.; Campos, C. Threecomponent competitive adsorption model for flow-through PAC systems. 1. Model development and verification with a PAC/ membrane system. Environ. Sci. Technol. 2003, 37, 2997–3004. (4) Crittenden, J. C.; Weber, W. J., Jr. Predictive model for design of fixed-bed adsorbers: Parameter estimation and model

(19)

(20)

(21)

development. J. Environ. Eng. Div. (Am. Soc. Civ. Eng.) 1978, 104, 185–197. Li, Q.; Snoeyink, V. L.; Mariñas, B. J.; Campos, C. Elucidating competitive adsorption mechanisms of atrazine and NOM using model compounds. Water Res. 2003, 37, 773–784. White, J. C.; Pignatello, J. J. Influence of bisolute competition on the desorption kinetics of polycyclic aromatic hydrocarbons in soil. Environ. Sci. Technol. 1999, 33, 4292–4298. Zhao, D.; Pignatello, J. J.; White, J. C.; Braida, W.; Fernandino, F. Dual-mode modeling of competitive and concentration dependent sorption and desorption kinetics of polycyclic aromatic hydrocarbons in soils. Water Resour. Res. 2001, 37, 2205–2212. Braida, W. J.; White, J. C.; Ferrandino, F. J.; Pignatello, J. J. Effect of solute concentration on sorption of polyaromatic hydrocarbons in soil: Uptake rates. Environ. Sci. Technol. 2001, 35, 2765–2772. Sontheimer, H.; Crittenden, J. C.; Summers, S. Activated Carbon for Water Treatment, 2nd ed.; DVGW-Forschungsstelle: Karlsruhe, Germany, 1988; pp. 258–348.. Qi, S.; Adham, S. S.; Snoeyink, V. L.; Lykins, B. W., Jr. Prediction and verification of atrazine adsorption by PAC. J. Environ. Eng. 1994, 120 (1), 202–218. Sudo, Y.; Misic, D. M.; Suzuki, M. Concentration dependence of effective surface diffusion coefficients in aqueous phase adsorption on activated carbon. Chem. Eng. Sci. 1978, 33, 1287– 1290. Suzuki, M.; Fujii, T. Concentration dependence of surface diffusion coefficient of propionic acid in activated carbon particles. AIChE J. 1982, 28, 380–385. Najm, I. N.; Snoeyink, V. L.; Richard, Y. Effect of initial concentration of a SOC in natural water on its adsorption by activated carbon. J. Am. Water Works Assoc. 1991, 83 (8), 57–63. Radke, C. J.; Prausnitz, J. M. Thermodynamics of multi-solute adsorption from dilute liquid solutions. AIChE J. 1972, 18, 761– 768. Crittenden, J. C.; Luft, P.; Hand, D. W. Prediction of multicomponent adsorption equilibria in background mixtures of unknown composition. Water Res. 1985, 19, 1537–1548. Ding, L.; Mariñas, B. J.; Schideman, L. C.; Snoeyink, V. L.; Li, Q. Competitive effects of natural organic matter: Parameterization and verification of the three-component adsorption model COMPSORB. Environ. Sci. Technol. 2006, 40, 350–356. Li, Q. Competitive adsorption of trace organic compounds by PAC in membrane filtration systems. Ph.D. Thesis, University of Illinois at Urbana-Champaign, Urbana, IL, 2002. Li, Q.; Mariñas, B. J.; Snoeyink, V. L.; Campos, C. Pore blockage effects on atrazine adsorption in a powdered activated carbon/ membrane system II: Model verification and application. J. Environ. Eng. 2004, 130, 1253–1262. Li, L.; Quinlivan, P. A.; Knappe, D. R. U. Effects of activated carbon surface chemistry and pore structure on the adsorption of organic contaminants from aqueous solution. Carbon 2002, 40, 2085–2100. Newcombe, G.; Morrison, J.; Hepplewhite, C. Simultaneous adsorption of MIB and NOM onto activated carbon. I. Characterisation of the system and NOM adsorption. Carbon 2002, 20, 2135–2146. Newcombe, G.; Morrison, J.; Hepplewhite, C.; Knappe, D. R. U. Simultaneous adsorption of MIB and NOM onto activated carbon. II. Competitive effects. Carbon 2002, 40, 2147–2156.

ES702609R

VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2611