Adsorption of Humic Substances on Goethite: Comparison between

Sep 12, 2006 - The relative change of the adsorption with pH is bigger for HA than for FA at relatively ... data and compare these to similar data of ...
0 downloads 0 Views 311KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 7494-7500

Adsorption of Humic Substances on Goethite: Comparison between Humic Acids and Fulvic Acids† L I P I N G W E N G , * ,‡ WILLEM H. VAN RIEMSDIJK,‡ LUUK K. KOOPAL,§ AND TJISSE HIEMSTRA‡ Department of Soil Quality, Wageningen University, P.O. Box 8005, 6700 EC, Wageningen, The Netherlands, and Laboratory of Physical Chemistry and Colloid Science, Wageningen University, P.O. Box 8038, 6700 EK, Wageningen, The Netherlands

The adsorption of humic acids (HA) to goethite (at pH 3-11) and the proton co-adsorption (at pH 4.0, 5.5, and 7.0) were measured, and the results were compared to those of fulvic acids (FA). Compared to FA, the adsorption of HA is stronger and more ionic strength dependent. The adsorption of both HA and FA decreases with increasing pH. The relative change of the adsorption with pH is bigger for HA than for FA at relatively low pH. At relatively high pH, it is the opposite. Protons are released at pH 4.0 and coadsorbed at pH 5.5 and 7.0 upon the adsorption of both HA and FA. The observed pH dependency of HA and FA adsorption is in agreement with the proton co-adsorption data. Model calculations show that the adsorbed FA particles are on average located in the Stern layer, whereas the adsorbed HA particles protrude beyond the Stern layer. The closer location to the surface of the adsorbed FA leads to stronger electrostatic interactions between the FA particles and the surface, which explains the larger amount of protons released at low pH and co-adsorbed at high pH with each mass unit of FA adsorbed than that with HA adsorbed. The model also reveals that for FA a meanfield (smeared-out) approximation is reasonable, but for HA a patchwise approach is more appropriate at relatively low loading.

Introduction Fulvic acids (FA) and humic acids (HA) are the most important reactive fractions of natural organic matter (NOM) in soils, sediments, surface water, and groundwater. Humic substances adsorb readily onto silicate clay and metal oxide minerals (1-14). The adsorption of humics reduces the solubility of natural organic matter and thus the mobility of the nutrients and pollutants bound to it. FA and HA are operationally defined fractions of NOM based on their solubility difference in acid and base. Both FA and HA are chemically heterogeneous materials and show a continuum of properties. In spite of their similarities, some general distinctions can still be made. Compared to FA, HA particles are larger in size (higher molar mass), contain more † This article is part of the Modeling Natural Organic Matter Focus Group. * Corresponding author e-mail: [email protected]. ‡ Department of Soil Quality. § Laboratory of Physical Chemistry and Colloid Science.

7494

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006

carbon, and are less soluble. In modeling the pH-charge behavior of FA and HA, Milne et al. (15) have concluded that compared to FA, HA has a lower total charge (per unit of mass) and a smaller proportion of low proton affinity (“carboxylic”) type of sites to the high affinity (“phenolic”) type of sites. Adsorption of humic substances to minerals is strongly influenced by pH. It is generally found that the adsorption decreases at increasing pH (1-14). Sometimes, an adsorption maximum was found around pH 4-6 (16). It has been shown that the pH dependency of the adsorption (of both small and big molecules) is directly linked to the mass balance of protons in the adsorption process (17-19). When protons are consumed (co-adsorbed) in the process, adsorption will decrease at increasing pH. On the contrary, when there is a net proton release, adsorption will increase with the increase of pH. It has been demonstrated that the proton coadsorption or release depends on the charge of the ions (particles) and its location in the electrostatic field near the surface (17, 19). The difference in size and charging behavior of FA and HA may therefore lead to different electrostatic interactions with the oxide surface and therefore to a different proton co-adsorption and pH dependency of the adsorption. In this paper, we will present new experimental data of HA adsorption to goethite together with proton co-adsorption data and compare these to similar data of FA (2, 20). With the interpretation of the data specific attention will be paid to (i) the pH dependency of FA and HA adsorption, (ii) the relation between pH dependency and proton co-adsorption, and (iii) the effects of the location of the adsorbed FA and HA near the oxide surface on the proton co-adsorption. For the last purpose, model calculations will be presented that can describe the proton co-adsorption.

Materials and Methods Goethite. The goethite preparation has been described by Hiemstra et al. (21). The BET-N2 surface area of the goethite material prepared is 94 m2/g. Before use, the goethite suspension was adjusted to pH 5.3 by adding nitric acid and continuously purged with clean, moist N2 gas to remove CO2. The surface charge has been determined with acid-base titrations in NaNO3 solution (9). The pristine point of zero charge (PZC) of this goethite material is pH 9.3. Goethite material used in the experiments of FA (2) has been prepared in a similar way, with a BET-N2 surface area of 94 m2/g and PZC of pH 9.2. The charge of goethite can be described with the CD-MUSIC model (see below) with the parameters as given in the Supporting Information. Within the experimental uncertainty, the charging behavior of the present goethite is equal to the charging properties of the goethite used previously in the FA experiments (2). Humic Acids and Fulvic Acids. HA were prepared from the B-horizon of a forest soil from The Netherlands (Tongbersven) following the IHSS (International Humic Substance Society) procedures (22, 23). Purified HA material (carbon content 54%) was freeze-dried and stored at room temperature. A stock solution was prepared from the freeze-dried material by dissolving the material first in base solution followed by dilution with the background electrolyte solution. The weight average molar mass of the HA is 13.2 kDa, which was measured using size exclusion chromatography (TSK 3000) in combination with TOC measurements (24). The charging behavior has been determined previously with acidbase titration in KNO3 solution (15). FA (Strichen) used by Filius et al. (2) were prepared from the Bs-horizon of a peat soil from Scotland (carbon content 43%, molar mass 0.683 10.1021/es060777d CCC: $33.50

 2006 American Chemical Society Published on Web 09/12/2006

kDa). The charging behavior of the FA has been determined with acid-base titration in NaCl solution (2). The pH-charge curves for HA are presented in the Supporting Information together with those of FA (2). The charge of FA can be described well with the NICA-Donnan model (see below) using the generic parameters (15). For HA, the model parameters that were optimized by fitting the pH-charge data of this HA to the NICA-Donnan model were used (15). All the model parameters can be found in Supporting Information. The site density shows that the total number of reactive groups on HA is less than that on FA when expressed in unit of mass. But when expressed per molecule or particle, the number of groups on one HA particle is much larger than that on one FA particle (see Supporting Information). HA Adsorption Experiment. Adsorption experiments of HA on goethite were carried out in a similar way to that of FA adsorption (2). Gastight 20 mL polyethylene vessels with screw caps were used. During the filling procedure, the vessels were flushed with N2 gas to minimize the influence of CO2. Certain amounts of stock solutions of goethite and NaNO3 were added to the vessels followed by addition of ultrapure water and certain amounts of HA stock solution. Three salt levels were studied: 0.002, 0.01, and 0.10 M NaNO3. The initial HA concentrations were 150, 300, and 450 mg/L. Acid or base solution (0.100 M NaOH and HNO3) was added to adjust the pH to values in the range of 3-11. The final concentration of goethite was 1.0 g/L. After shaking for 3 days at 20 °C, the suspensions were centrifuged at 18000g for 30 min. A subsample of the supernatant was acidified to pH 3-4, and the concentration of dissolved organic C in the supernatant was measured with a TOC analyzer. The final pH was measured in the supernatant. The HA adsorption experiments are faced with the difficulty that upon centrifugation also some non-adsorbed HA is removed. To calibrate this loss, control samples without goethite were used. The amount of HA removed by centrifugation is presented in the Supporting Information. The results show that above pH 3, the removed amount is almost constant for a given initial amount of HA and salt concentration (about 10% at 0.01M NaNO3, 20% at 0.10 M NaNO3). Below pH 3, some extra fraction of the HA can be removed, which indicates that coagulation happens at pH < 3. For the HA adsorption experiment, the amount of HA adsorbed at pH > 3 was calculated as the difference of HA in solution after the centrifugation between the control (without goethite) and the treatment. Proton Co-adsorption Experiment. The proton coadsorption upon HA adsorption on goethite was measured with pH-stat titrations using an automated titration set up (25) under N2 atmosphere at 20 °C. Suspensions (with 10 g/L goethite) in 0.01 M or 0.10 M NaNO3 were titrated at a constant pH of 4.0, 5.5, or 7.0 with a HA solution (5 g/L) having the same pH and salt concentration. After each HA addition, the pH was corrected to the initial value with acid or base (0.100 M HNO3 and NaOH). A reaction time of at least 20 min and a maximum drift criterion of 0.002 pH units per minute were used between each addition of HA. The total amount of added HA was sufficiently small as compared to the total surface area of the goethite to have virtually all HA adsorbed. The proton co-adsorption was calculated from the amount of acid and base added to maintain the pH.

Modeling Modeling approaches are used to calculate the charge of pure goethite, FA, and HA in solution and the proton mass balance in the FA-goethite and HA-goethite systems. The CD-MUSIC (Charge Distribution MUlti SIte Complexation) model (15, 26) in combination with the Basic Stern (BS) model is used to calculate the charge of pure goethite.

In the CD-MUSIC model, the basic charging of goethite is assumed to be caused by the protonation and deprotonation of the singly and triply coordinated surface oxygens. Indifferent electrolyte ions can form ion pairs with both types of sites. The charge carried by the deprotonated and protonated surface sites is located at the surface plane (0-plane), whereas the charge carried by the ion pairs is located at the head end of the diffuse layer (d-plane). (See Supporting Information for more details.) The NICA-Donnan model (15, 27-30) is used to calculate the charge of FA and HA in solution. In the NICA model, the basic charging of FA and HA is assumed to be caused by the protonation and deprotonation of the carboxylic and phenolic type of groups present on the material. The protonation reaction of the carboxylic and phenolic groups is calculated with the NICA equation. The electrostatic effects on the charge of humics in solution phase is described with the Donnan model. Electrolyte ions, such as K+ and Na+, are assumed to interact with humics via only electrostatic interactions, calculated with the Boltzmann law. (See Supporting Information for more details.) For the FA-goethite and HA-goethite systems, the Ligand and Charge Distribution (LCD) model concept (9, 20, 31) is used, which combines the CD-MUSIC and NICA model. Following the LCD concept, the NICA model that is used to describe the specific interaction of ions with humics in solution can also be applied to calculate the specific interaction of ions with adsorbed humics. In the context of this paper, we will only consider the chemical binding of protons to the carboxylic and phenolic groups of adsorbed humics. It is assumed that the NICA model parameters remain the same as for the FA and HA in the bulk solution. The electrostatic potential at goethite surface in the presence of adsorbed humics is calculated with the Basic Stern model, taking into account the charge carried by the adsorbed humics. The change of protonation and charge of goethite surface upon humics adsorption is calculated with the CDMUSIC model. The protonation of the adsorbed humics and goethite is calculated for a given amount of adsorbed FA or HA. More detailed description of the LCD model can be found elsewhere (9, 20, 31). The model calculations were done with the computer code ORCHESTRA (32) and the calculation was carried out numerically in an iterative way.

Results and Discussion Adsorption of HA and FA. The adsorption envelopes of HA measured with the batch adsorption experiments are depicted in Figure 1. The adsorption decreases gradually with increasing pH and the decrease is stronger at lower salt concentration. A striking observation is that the adsorption is still positive at pH values 1.5 pH units above the pristine PZC of goethite. This is an indication that, besides the electrostatic interactions, other mechanisms such as specific adsorption, hydrophobic interactions are involved. To clearly illustrate the effect of the HA concentration, the adsorption envelopes are transformed into adsorption isotherms for various pH values (see Figure 2). For comparison, we have also presented the FA adsorption isotherms on goethite at the same pH values and similar salt concentrations (2) (see Figure 2). Due to lack of data at low humics concentration, the initial slope of the adsorption isotherms is uncertain. The isotherms for both HA and FA show plateau values that are pH dependent, but at the same mass concentration in solution (mg/L), HA adsorbs much stronger (in the unit of mg/m2) than FA. A greater extent of HA adsorption than FA or a preference of higher molecular weight NOM in the adsorption has been reported previously (7, 1113, 33, 34). The adsorption levels for HA compare well with those of Saito et al. (5) for the purified Aldrich humic acid and goethite system. Literature values of the adsorption VOL. 40, NO. 24, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7495

FIGURE 1. Adsorption envelopes of HA measured at three salt levels and three HA loadings. Black symbols: 0.10 M NaNO3. Gray symbols: 0.01 M NaNO3. Open symbols: 0.002 M NaNO3. therein), which is in accordance with the range of the adsorption density measured for FA and HA in this work (Figure 2). It is illustrative to express the amount of humics not only in units of mass but also in moles as is done in Figure 2 with a second y-axis using the molar mass of FA or HA for recalculation. On a molar basis, the maximum amount of HA adsorbed is much less than FA. Next to the adsorbed amount also the salt dependency is quite different for both humics. The adsorption of HA is strongly dependent on the salt concentration and that of FA barely. A significant positive salt dependency of HA adsorption on goethite, hematite, and aluminum oxides has also been found previously (5, 11). The ionic strength effects on the adsorption of FA reported in the literature are not consistent. Schlautman and Morgan (11) have found that the adsorption of FA on aluminum oxides increases with increasing ionic strength. Vreysen and Maes (8) have found hardly any influence of the ionic strength on the adsorption of FA on bentonites. Among other interaction mechanisms, the salt ions can compete with the humics particles for the sites and charge neutralization of the surface, leading to a reduced adsorption. On the other hand, the salt ions can screen the charge and reduce the electrostatic repulsion between the adsorbed humics particles, leading to an enhanced adsorption (7). The net effect is a combination of the results from various interactions.

FIGURE 2. Adsorption isotherms of HA and FA. Symbols are data estimated from the envelopes. Lines are hand-drawn to guide the eye. Data of FA are from Filius et al. (2). Black lines and symbols: 0.10 M NaNO3. Gray line and symbols: 0.01 M (HA) or 0.015 M (FA) NaNO3. densities of natural organic matter on oxides at around pH 4 range from 0.3 to 4.3 mg/m2 (Saito et al. (5) and references 7496

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006

Another remarkable difference between the adsorption of FA and HA is the pH dependency. At low pH, the pH dependency of HA adsorption is stronger than that of FA adsorption. However, this trend is reversed at high pH (Figure 1 and Figure 2 in ref 2). At pH 10, still a considerable adsorption is present in case of HA in comparison to FA. More discussion about the pH dependency will be found below. Proton Co-adsorption. The measured net proton coadsorption upon the adsorption of HA to goethite is shown in Figure 3. For comparison, similar data for FA (20) are also given (Figure 3). The net proton co-adsorption depends on

FIGURE 3. Proton co-adsorption data. Black symbols: 0.10 M NaNO3. Gray symbols: 0.01 M NaNO3. Data of FA are from Filius et al. (2). the pH, the salt level, and the loading of the oxide with FA or HA. For both FA and HA, at pH 4.0, a net proton release is seen (negative proton co-adsorption) at relatively low loading. At pH 5.5 and pH 7.0, a net proton adsorption is observed. At pH 5.5 and pH 7.0, the total amount of protons co-adsorbed increases close to linearity with the amount of FA or HA adsorbed. At pH 4.0, however, the sign of the slope inverts at a certain amount of humics adsorbed, which means a variable ratio between proton co-adsorbed and FA or HA adsorbed. When humics are adsorbed to a positively charged surface, the particles will tend to release protons due to the less negative potential in the vicinity of the surface than in the electrostatic phase in the solution. However, the introduction of negative charge close to the surface will lead to more proton binding to the surface sites (19). The overall proton coadsorption is the sum of these two opposing processes. At pH 4.0 the goethite surface is highly (positively) charged, and humics are weakly (negatively) charged. The effects of the positive charge of oxides on the potential change for humics dominates; therefore, a net proton release is observed. At pH 5.5 and 7.0 the effects of negative charge of humics on the oxide surface dominate, which leads to a net proton adsorption. The results show that, compared to that of HA, with each milligram of FA adsorbed, there are more protons released at pH 4.0 and more protons co-adsorbed at pH 5.5 and pH 7.0. This indicates a stronger electrostatic interaction of FA with the surface than HA (see below). However, when expressed per mole HA or FA adsorbed, the amount of protons released or co-adsorbed is larger for HA than for FA due to a higher molar mass of HA. The proton co-adsorption data represent the sum of the change in goethite charge and that of the adsorbed FA or HA due to the interaction. Therefore the total charge carried by

FIGURE 4. Sum of the charge of HA-goethite or FA-goethite. Thin lines are calculated as the sum of the charge of goethite and HA or FA by assuming no interactions (i.e., no proton co-adsorption). Thick lines are calculations by taking into account the measured proton co-adsorption. Results for only 0.10 M NaNO3 are shown. the HA-goethite or FA-goethite complex can be calculated as the sum of the initial charge of goethite and that of HA or FA, corrected for the amount of protons co-adsorbed or released. The initial charge of goethite and HA or FA can be calculated using respectively the CD-MUSIC and the NICADonnan model. The calculated total charge of the humicsgoethite complex is given in Figure 4, which shows that for FA the charge reversal is limited (pH 4.0 and pH 5.5) or not reached (pH 7.0), whereas for HA strong overcompensation of the surface charge occurs at all three pH values. The proton co-adsorption has not been measured up to the adsorption plateau values, but when we extrapolate the curves for FA to these values we see that also for FA at pH 7.0 an overcompensation of the surface charge is likely. On the basis of these results, it may be concluded that the plateau values of the adsorption isotherms (see Figure 2) are due to repulsive electrostatic forces between the humics particles. pH Dependency and Proton Co-adsorption. The relation between the proton co-adsorption and pH dependency of FA and HA adsorption observed can be illustrated with the concept of thermodynamic consistency (18), which can be expressed as (17, 19)

(

)

∂ log Ci,t ∂ pH

Γi

) (Nnet)pH

(1)

where (∂ log Ci,t/∂ pH)Γi is the ratio of the change of the total concentration of the adsorbing molecule i (mol/L, logarithm) in the solution (∂ log Ci,t) to the change of pH (∂ pH) at a constant adsorbed amount of i (Γi). (Nnet)pH is the net molar ratio of proton co-adsorbed to that of molecule i adsorbed at a constant pH, which can be directly measured in a well VOL. 40, NO. 24, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7497

FIGURE 5. Model prediction of proton co-adsorption. Results are for pH 5.5, 0.10 M NaNO3. The loading used in the patchwise approach is 0.18 µmol/m2 for HA and 0.56 µmol/m2 for FA. Symbols are experimental data. Lines are model calculations. designed experiment (as done here in the proton coadsorption experiments). Equation 1 says, in terms of HA or FA adsorption, that at a certain amount of HA or FA adsorbed, the HA or FA adsorption will decrease with increasing pH when Nnet is positive (net proton co-adsorption), whereas the adsorption will increase with increasing pH if Nnet is negative (net proton release). The values of Nnet can be derived from the slopes of the curves in Figure 3, which show that at pH 5.5 or pH 7.0 the slope (Nnet) is positive for both FA and HA. According to eq 1, this means that around these pH values, the adsorption will decrease with increasing pH, which is in agreement with the observed adsorption data (Figure 1 and Figure 2 in ref 2). At pH 4.0, the slope (Nnet) is first negative and becomes positive above a certain loading. This means that around this pH the adsorption will increase with increasing pH if the loading of the surface with HA or FA is lower than the turning points in the curves at pH 4.0 in Figure 3, whereas the adsorption will decrease with increasing pH if the loading is higher than these turning points. In the presented adsorption data (Figure 1 and Figure 2 in ref 2), however, no decrease of adsorption at pH below 4 is observed due to either almost 100% adsorption or a loading larger than the turning points. From Figure 3, it can be derived that the amount of protons released at pH 4.0 or co-adsorbed at pH 5.5 and pH 7.0 with each mole of HA is larger than that of FA (i.e., the absolute value of Nnet (mol/mol) is larger for HA than for FA). According to eq 1, this implies a bigger shift of the adsorption isotherms with pH (i.e., a bigger (∂ log Ci,t/∂ pH)Γi). From the HA and FA adsorption data (Figure 1 and Figure 2 in ref 2), we can estimate this shift ((∂ log Ci,t/∂ pH)Γi). The results (not shown) confirm that, at low pH, this shift is bigger for HA than for FA. On the contrary, at high pH, the results derived from these figures indicate a smaller shift for HA than for FA, which means a smaller Nnet value for HA than for FA. On this basis, we can predict that, at high pH (>7), the slope of proton co-adsorption curves (as shown in Figure 3) will be larger for FA than for HA. 7498

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006

Location of HA and FA Adsorbed. Although there maybe chemical binding between humics and oxides, the proton co-adsorption is mainly dependent on the charge distribution of the adsorbed molecules and the electrostatic interactions between the adsorbed molecules and the surface (17, 19). Provided that the adsorbed amounts of humics are known, it is possible to calculate the proton co-adsorption, in which the protonation of the bare goethite and the protonation of the FA and HA in the solution phase are calculated using respectively the CD-MUSIC and the NICA-Donnan model, whereas the protonation of the oxide and the humics in the humics-oxide complex is calculated using the LCD model (see Modeling section). The net proton co-adsorption equals the difference between the total amount of protons adsorbed (on both goethite and humics) after and before the adsorption. For the humics covered surface, a realistic representation of the electrostatic field near the surface requires consideration of the local variation of charge and potential, and electrostatic interactions between humics and surfaces and between humics particles. This type of modeling would be rather complicated and conceptually and mathematically involved. For simplicity, the mean-field assumption is commonly used, in which it is assumed that the charge of the adsorbed molecules is smeared-out over the entire surface. Another simplified approach assumes that the charge differs locally whether or not an adsorbed humics particle is present on that part of the surface. This approach will be referred to as patchwise approach. Both approaches (smeared-out and patchwise) are applied in this study to calculate the proton co-adsorption with HA or FA adsorption. In the patchwise approach, the surface area of the oxide that is influenced by the charge of adsorbed humics follows from the effective electrostatic size of the humic particles. For the calculations given in this study, these values are estimated from the turning points of the proton co-adsorption curves at pH 4.0 (Figure 3), under the assumption that at this point

the loading of the surface with humics is high enough that the electrostatic interactions between adsorbed particles occur. With the above assumptions, the proton co-adsorption as a function of the adsorption of FA or HA is calculated for the experimental conditions used. Typical results obtained for both the smeared-out and the patchwise assumption for pH 5.5 are depicted in Figure 5. For FA, the quality of prediction is similar for the smeared-out and patchwise approaches. The smeared-out approach has the advantage that the bending of the curve can be represented. The calculations suggest that the average representative charge location for FA is between the 0-plane and d-plane. A reasonable description of the proton co-adsorption is found if 10% of the FA groups is located at the 0-plane while the rest is placed at d-plane. For HA, the smeared-out approach gives a poor description of the co-adsorption data at low loading. This changes when we use the patchwise approach. The calculation shows that a reasonable description is found if 50% of the HA is placed at the d-plane and the other part is in the diffuse layer. For the part of HA in the diffuse layer, we assume that the thickness of the HA in the diffuse layer is 1 nm. As discussed above, when expressed per milligram HA or FA adsorbed, the amount of proton released at pH 4.0 and co-adsorbed at pH 5.5 and pH 7.0 is larger for FA than for HA (Figure 3). The reason for this observation is due to partly a more negative charge (per mg) on FA than on HA. Another reason is that the average location of the reactive groups on FA is closer to the surface in comparison to that of HA. The electrostatic interactions between the adsorbed particles and the surface will be stronger when the particles are located closer to the oxide surface. The relatively small size of FA as compared to that of HA makes it easier for a FA particle to get into more close contact with the surface. The model calculations also indicate that for FA the mean-field or smeared-out approach can describe the proton co-adsorption data reasonably, while for HA the patchwise approach leads to a better description when the loading is low. The meanfield assumption is a better approximation for relatively small molecules and high adsorption density. This observation will be of direct importance for the development of physically realistic models for the description of humics adsorption to oxides.

Acknowledgments This research was partly funded by the EU project, FUNMIG (516514, F16W-2004). We would like to thank Andre Van Zomeren from ECN (Energy Research Center, The Netherlands) for measuring the molar mass of HA.

Supporting Information Available More detailed information about (i) the models and model parameters, (ii) pH-charging curves of goethite, HA and FA, (iii) removal of HA by centrifugation. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Davis, J. A. Adsorption of natural dissolved organic matter at the oxide/water interface. Geochim. Cosmochim. Acta 1982, 46, 2381-2393. (2) Filius, J. D.; Lumsdon, D. G.; Meeussen, J. C. L.; Hiemstra, T.; Van Riemsdijk, W. H. Adsorption of fulvic acid on goethite. Geochim. Cosmochim. Acta 2000, 64, 51-60. (3) Gu, B. H.; Schmitt, J.; Chen, Z. H.; Liang, L. Y.; McCarthy, J. F. Adsorption and desorption of natural organic-matter on iron oxidesmechanisms and models. Environ. Sci. Technol. 1994, 28, 38-46. (4) Parfitt, R. L.; Fraser, A. R.; Farmer, V. C. Adsorption on hydrous oxides. 3. Fulvic-acid and humic-acid on goethite, gibbsite and imogolite. J. Soil Sci. 1977, 28, 289-296.

(5) Saito, T.; Koopal, L. K.; Van Riemsdijk, W. H.; Nagasaki, S.; Tanaka, S. Adsorption of humic acid on goethite: Isotherms, charge adjustments, and potential profiles. Langmuir 2004, 20, 689-700. (6) Tipping, E. The adsorption of aquatic humic substances by iron oxides. Geochim. Cosmochim. Acta 1981, 45, 191-199. (7) Vermeer, A. W. P.; Van Riemsdijk, W. H.; Koopal, L. K. Adsorption of humic acid to mineral particles. 1. Specific and electrostatic interactions. Langmuir 1998, 14, 2810-2819. (8) Vreysen, S.; Maes, A. Adsorption mechanism of fulvic acid onto freeze dried poly(hydroxo aluminum) intercalated bentonites. Appl. Clay Sci. 2006, 32, 190-196. (9) Weng, L. P.; Koopal, L. K.; Hiemstra, T.; Meeussen, J. C. L.; Van Riemsdijk, W. H. Interactions of calcium and fulvic acid at the goethite-water interface. Geochim. Cosmochim. Acta 2005, 69, 325-339. (10) Schulthess, C. P.; Huang, C. P. Humic and fulvic-acid adsorption by silicon and aluminum-oxide surfaces on clay-minerals. Soil Sci. Soc. Am. J. 1991, 55, 34-42. (11) Schlautman, M. A.; Morgan, J. J. Adsorption of aquatic humic substances on colloidal-size aluminum-oxide particlessinfluence of solution chemistry. Geochim. Cosmochim. Acta 1994, 58, 4293-4303. (12) Hur, J.; Schlautman, M. A. Humic substance adsorptive fractionation by minerals and its subsequent effects on pyrene sorption isotherms. J. Environ. Sci. Health Part A 2006, 41, 343358. (13) Ko, I.; Kim, J. Y.; Kim, K. W. Adsorption properties of soil humic and fulvic acids by hematite. Chem. Speciation Bioavailability 2005, 17, 41-48. (14) Reiller, P.; Amekraz, B.; Moulin, C. Sorption of Aldrich humic acid onto hematite: insight into fractionation phenomena by electrospray ionization with quadrupole time-of-flight mass spectrometry. Environ. Sci. Technol. 2006, 40, 2235-2241. (15) Milne, C. J.; Kinniburgh, D. G.; Tipping, E. Generic NICADonnan model parameters for proton binding by humic substances. Environ. Sci. Technol. 2001, 35, 2049-2059. (16) Davis, J. A.; Gloor, R. Adsorption of dissolved organics in lake water by aluminum-oxide-effect of molecular-weight. Environ. Sci. Technol. 1981, 15, 1223-1229. (17) Hiemstra, T.; Van Riemsdijk, W. H. In Encyclopaedia of Surface Colloid Science; Hubbard, A. T., Ed.; Marcel Dekker: New York, 2002. (18) Perona, M. J.; Leckie, J. O. Proton stoichiometry for the adsorption of cations on oxide surfaces. J. Colloid Interface Sci. 1985, 106, 64-69. (19) Rietra, R.; Hiemstra, T.; Van Riemsdijk, W. H. Electrolyte anion affinity and its effect on oxyanion adsorption on goethite. J. Colloid Interface Sci. 2000, 229, 199-206. (20) Filius, J. D.; Meeussen, J. C. L.; Lumsdon, D. G.; Hiemstra, T.; Van Riemsdijk, W. H. Modeling the binding of fulvic acid by goethite: the speciation of adsorbed FA molecules. Geochim. Cosmochim. Acta 2003, 67, 1463-1474. (21) Hiemstra, T.; Van Riemsdijk, W. H.; Bolt, G. H. Multisite proton adsorption modeling at the solid-solution interface of (hydr)oxidessa new approach. 1. Model description and evaluation of intrinsic reaction constants. J. Colloid Interface Sci. 1989, 133, 91-104. (22) Aiken, G. R.; Thurman, E. M.; Malcolm, R. L.; Walton, H. F. Comparison of XAD macroporous resins for the concentration of fulvic acid from aqueous solution. Anal. Chim. Acta 1979, 51, 1799-1803. (23) Swift, R. S. In Methods of Soil Analysis. Part 3. Chemical Methods; Sparks, D. L., Page, A. L., Helmke, P. A., Loeppert, R. H., Soltanpour, P. N., Tabatabai, M. A., Johnston, C. T., Sumner, M. E., Eds.; Soil Science Society of America Inc.: Madison, WI, 1996. (24) Van Zomeren, A.; Comans, R. N. J. Contribution of natural organic matter to copper leaching from municipal solid waste incinerator bottom ash. Environ. Sci. Technol. 2004, 38, 39273932. (25) Kinniburgh, D. G.; Milne, C. J.; Venema, P. Design and construction of a personal-computer-based automatic titrator. Soil Sci. Soc. Am. J. 1995, 59, 417-422. (26) Hiemstra, T.; Van Riemsdijk, W. H. A surface structural approach to ion adsorption: the charge distribution (CD) model. J. Colloid Interface Sci. 1996, 179, 488-508. (27) De Wit, J. C. M.; Van Riemsdijk, W. H.; Nederlof, M. M.; Kinniburgh, D. G.; Koopal, L. K. Analysis of ion binding on humic substances and the determination of intrinsic affinity distributions. Anal. Chim. Acta 1990, 232, 189-207. VOL. 40, NO. 24, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7499

(28) Koopal, L. K.; Van Riemsdijk, W. H.; De Wit, J. C. M.; Benedetti, M. F. Analytical isotherm equations for multicomponent adsorption to heterogeneous surfaces. J. Colloid Interface Sci. 1994, 166, 51-60. (29) Kinniburgh, D. G.; Van Riemsdijk, W. H.; Koopal, L. K.; Borkovec, M.; Benedetti, M. F.; Avena, M. J. Ion binding to natural organic matter: competition, heterogeneity, stoichiometry and thermodynamic consistency. Colloids Surf. A 1999, 151, 147-166. (30) Koopal, L. K.; Saito, T.; Pinheiro, J. P.; Van Riemsdijk, W. H. Ion binding to natural organic matter: General considerations and the NICA-Donnan model. Colloids Surf. A 2005, 265, 40-54. (31) Weng, L. P.; Van Riemsdijk, W. H.; Koopal, L. K.; Hiemstra, T. Ligand and charge distribution (LCD) model for the description of fulvic acid adsorption to goethite. J. Colloid Interface Sci. (in press).

7500

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006

(32) Meeussen, J. C. L. ORCHESTRA: an object-oriented framework for implementing chemical equilibrium models. Environ. Sci. Technol. 2003, 37, 1175-1182. (33) Specht, C. H.; Kumke, M. U.; Frimmel, F. H. Characterization of NOM adsorption to clay minerals by size exclusion chromatography. Water Res. 2000, 34, 4063-4069. (34) Namjesnik-Dejanovic, K.; Maurice, P. A.; Aiken, G. R.; Cabaniss, S.; Chin, Y. P.; Pullin, M. J. Adsorption and fractionation of a muck fulvic acid on kaolinite and goethite at pH 3.7, 6, and 8. Soil Sci. 2000, 165, 545-559.

Received for review March 31, 2006. Revised manuscript received July 29, 2006. Accepted August 2, 2006. ES060777D