| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Plant Science, U-67, Univ. of Connecticut, Storrs, CT 06269
Corresponding author (c.schulthess@.uconn.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
2:1 at high pH values (>6), but they decreased rapidly at low pH values. The stoichiometry is 2:1 across a broader pH range when corrections are made for the change in Cl adsorption intensity by Al oxide in the presence of Se anions. Generalizing, if a weakly adsorbing anion is affected by another strongly adsorbing anion, then its impact on the balance of protons adsorbed (or balance of surface charge) must also be monitored. Although the Cl anion is very weakly adsorbing, the matrix needs to be considered as a ternary system consisting of H, Cl, and Se ions. A binary approach, one that views only H and Se ions, will misrepresent the H/Se adsorption ratios, particularly at low pH values.
Abbreviations: DRIFT, diffuse reflectance infrared Fourier transformed • EXAFS, extended x-ray absorption fine structure • FTIR, Fourier transformed infrared • ZPSE, zero point of salt effect • ZPT, zero point of titration
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
 | (1) |
 | (2) |
Adsorption of oxyanions by metal oxides or soil constituents is often accompanied by the coadsorption of protons. Some microscopic evidence for proton coadsorption on Al oxide was presented by Wijnja and Schulthess (1999)(2000a), who observed diffuse reflectance infrared Fourier transformed (DRIFT) spectral bands in the presence of adsorbed CO3 or SO4 attributable to extra protonated surface groups. It is important to note, however, that in addition to a microscopic identification of the resulting molecular surface-adsorbed structures, the stoichiometry of the reactants involved must be known in order to fully deduce the mechanism of adsorption. Microscopic data do not confirm the number of protons consumed in the surface reaction(s) (as illustrated, for example, in Eq. [1] and [2]).
Adsorption modeling is not a reliable alternative for quantifying the number of protons coadsorbed. The stoichiometry of the reaction cannot be properly inferred from the goodness-of-fit of mathematical models of the assumed adsorption mechanisms. For instance, Zhang and Sparks (1990) used only one selenate outer-sphere complex with a proton stoichiometry of unity on goethite, while another outer-sphere complex with a proton stoichiometry of two was also included by Hayes et al. (1988) on goethite. Ghosh et al. (1994) also assumed two outer-sphere reaction stoichiometries (2:1 and 1:1) for H/selenate on hydrous alumina. Similar diverse ratios are found in the literature for the H/selenite inner-sphere adsorption stoichiometries. Zhang and Sparks (1990) assumed both ratios (2:1 and 1:1) for H/selenite adsorption by goethite. Hayes et al. (1988) best optimized their modeling predictions when assuming two separate reactions with a 1:1 ratio for each. Hayes et al. (1988) also modeled the selenite adsorption on goethite as an outer-sphere complex, obtaining a close fit with a high background ionic strength and a H/selenite reaction stoichiometry of 2:1; the predictions were poor for the low ionic strength data. Ghosh et al. (1994) assumed a 0:1 ratio (or no proton coadsorption) in their predictions of selenite adsorption by a hydrous alumina. Hiemstra and van Riemsdijk (1999) suggested a 1:1 H/selenite adsorption ratio by goethite when using an alternate surface speciation for the Fe oxide, namely SOH0.5+2 and SOH0.5-.
Mathematical adsorption modeling has integrated together enough adjustment parameters that it cannot definitively prove that a particular adsorption stoichiometry has been identified. Regardless of what is assumed, a reasonably close fit with the adsorption data is usually found. Clearly, definitive macroscopic proton release and uptake data (or coadsorption data in the presence of adsorbing anions) are needed in defining the adsorption reactions by surface adsorption models (Dzombak and Morel, 1990).
To determine proton adsorption stoichiometries, a common technique used is the pH-stat procedure, which was used for H/Cd on goethite (Venema et al., 1996), OH/P on goethite (Hiemstra and van Riemsdijk, 1996), H/CO3 on goethite (Wijnja and Schulthess, 2001), and OH/SO4 on Al oxide and Ti oxide (He et al., 1996). A problem with the pH-stat method is that the coadsorption of protons with ions is often confounded by the inadvertent measurement of secondary reactions that also consume protons. Another method is the backtitration procedure developed by Schulthess and Sparks (1986), which was used for H/Cl on Al oxide (Schulthess and Sparks, 1987), and H/CO3 on Al oxide (Schulthess et al., 1998) and Ti oxide (Schulthess and Belek, 1998). The backtitration procedure subtracts the liquid-phase proton-consuming reactions from the analysis of the whole suspension and corrects for all non-adsorption-related reactions that have also released or consumed protons. Duquette and Hendershot (1993) also described a modified version of this backtitration technique. Regardless of the method used, it is essential that the impact of all the other ions in the system studied be minimal, or at least constant. For example, it is virtually impossible to avoid the presence of chlorides in pH-dependent experiments when Cl is part of the acid used (namely, HCl) or the background electrolyte used (namely, NaCl). However, it is generally assumed that the background electrolyte adsorbs so weakly that it will not compete with the adsorption patterns of other anions. When measuring the proton/anion adsorption stoichiometry of a strongly adsorbing anion (relative to the adsorbing strength of Cl), it is generally assumed that the Cl ions in the matrix will not impart any significant problem to the measurements.
The objective of this article is to show that using the proton adsorption data to determine a specific proton/anion coadsorption stoichiometry (such as those illustrated in Eq. [1] and [2]) can lead to erroneous conclusions when the reactivity of other compounds present in the mixture is changed. This will be illustrated with measurements of the adsorption of protons, selenate, and selenite on an Al oxide in the presence of background Cl- anions. As noted above, background Cl concentrations are generally assumed to adsorb weakly and, therefore, are typically not monitored. We hypothesize, however, that weakly adsorbing background anions (such as Cl) are affected by the presence of other stronger adsorbing anions (such as SeO3 and SeO4) and, consequently, interfere with the measurement of the proton/anion adsorption stoichiometry (namely, the proton/selenate and proton/selenite adsorption stoichiometries). A numerical correction procedure will be introduced that adjusts the number of total protons consumed by all surface reactions to the number of protons consumed by the Se surface adsorption reaction only.
Our hypothesis will be tested under two scenarios: one that has a strongly adsorbing anion (selenite) affecting the adsorption of very few Cl anions (these are present only when HCl is added), and another that has a weaker adsorbing anion (selenate) affecting the adsorption of a higher concentration of Cl anions (these are present as added HCl and surface contaminant). A low concentration of Cl anions may theoretically cause no influence to the adsorption patterns or stoichiometry of a strongly adsorbing anion (selenite). We will show this to be false. Conversely, a higher aqueous (and adsorbed) concentration of Cl anions may theoretically go uninfluenced by the presence of a weaker anion (selenate). We will show this to be false, too. Since these conditions cover the two possible extremes for these elements and starting material chosen, the results of all other combinations with these elements and starting material will, by implication, be elucidated as well.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
-Al2O3 (Aluminum Oxide-C, Degussa Corp., Teterboro, NJ) was used as an adsorbent in this study. The oxide had a Brunauer-Emmett-Teller surface area of 100 ± 15 m2 g-1, and the average primary particle size (dry) was 20 nm (supplied by the manufacturer). The Al oxide surface is modified when suspended in water (Furrer and Stumm, 1986), changing into an Al-hydroxide (bayerite) upon aging (Dyer et al., 1993; Wijnja and Schulthess, 1999). Bayerite is a soil constituent and, therefore, representative of a naturally occurring oxide. The Al oxide stock suspensions used for this study were prepared three different ways to illustrate the impact of washing procedures and low residual salt concentrations (or impurities) on the subsequent proton stoichiometry calculations. The Al stock suspensions were CO2 free, aged for 30 d, and in equilibrium with the aqueous phase. The three Al stock suspensions varied in residual Cl concentration values from high to low to none detected.
The first stock suspension (Lot 1) was prepared by adding Al oxide to deionized water. The slurry was centrifuged for 20 min at 18000 g, resuspended in fresh deionized water (first wash), and stirred for a minimum of 2 d. The Al oxide slurries were water-washed this way three times. Finally, the suspension was purged with pure air (CO2-free air, Connecticut Airgas, West Hartford, CT) for 7 d, and aged for 30 d prior to its first use. The supernatant conductivity was 52.3 µS cm-1, pH was 4.6, and the Al oxide concentration was 88.7 ± 0.75 g L-1 (gravimetric analysis with four replications). This Al oxide stock solution contained autochthonous Cl based on Cl-desorption data of the suspensions at high pH values (discussed below).
A second Al oxide stock suspension (Lot 2) was prepared as follows. The Al oxide was washed three times in 0.17 M NaOH (final supernatant pH = 8.8) to desorb the anion impurities (mainly Cl). The slurry was then washed five times with deionized water to obtain a constant pH and conductivity value in the supernatant. The oxide slurry was centrifuged for 15 min at 24 000 g between washes. The supernatant conductivity was 6.9 µS cm-1, pH was 6.7, and the Al oxide concentration was 85.9 ± 0.76 g L-1 (gravimetric analysis with four replications). It was also purged and aged as described above. A small amount of Cl was detected in the supernatant of this Al oxide stock solution at high pH values.
A third Al oxide stock suspension (Lot 3) was prepared as follows. The Al oxide was washed four times in 0.2 M NaOH (final supernatant pH = 10.2) to desorb the anion impurities (mainly Cl). The slurry was then washed nine times with deionized water to obtain a constant pH and conductivity value in the supernatant. The oxide slurry was centrifuged for 15 min at 24000 g between washes. The supernatant conductivity was 7.2 µS cm-1, pH was 7.1, and the Al oxide concentration was 48.4 ± 0.14 g L-1 (gravimetric analysis with four replications). It was also purged and aged as described above. No Cl was detected in this Al oxide stock solution at high pH values (pH 9.4).
Adsorption Measurements
Proton adsorption isotherms were measured using the backtitration technique developed by Schulthess and Sparks (1986). Titration samples were prepared in 50-mL (nominal) centrifuge tubes (Nalgene Oak Ridge polypropylene copolymer tubes, Fisher Scientific, Pittsburgh, PA), consisting of an aliquot of the Al oxide stock suspension (final weight of oxide present was 0.4435 g per tube when using Lots 1 and 3, or 0.4295 g per tube when using Lot 2), a specific volume of either 0.2 M HCl or 0.2 M NaOH to vary the pH (from pH 2 to 11), a specific volume of NaCl solution used to vary the ionic strength (0, 0.01, or 0.1 M final NaCl concentrations), and deionized water to achieve a total volume of 35 mL per tube. The headspace present in the capped tube was 6.5 mL. These proton adsorption experiments were repeated with the addition of 0.35 mL of 0.1 M solutions of Na2SeO4 or Na2SeO3 (final concentration was 0.001 M Se).
The proton adsorption isotherm samples were capped and mixed with gentle rocking and rotation of the centrifuge tubes for 19 to 21 h at 20°C on a hematology mixer (Fisher Scientific) in a temperature-controlled incubator, centrifuged at 20°C for 20 min at 24000 g, and then decanted into a 60-mL polyethylene bottle. A 0.25-mL supernatant aliquot for aqueous selenate, selenite, and chloride measurements was withdrawn. The remaining supernatant was capped and immediately weighed to determine the amount of supernatant recovered. The pH of this liquid was then measured and titrated using a Titrino 716 autotitrator (Metrohm, Herisau, Switzerland). An acidic supernatant was titrated with 0.02 M NaOH, while an alkaline supernatant was titrated with 0.02 M HCl. Atmospheric exposure was minimized through all phases of the experiment by keeping the containers tightly covered.
The amount of protons adsorbed or desorbed was calculated according to the mass balance equation described by Schulthess and Sparks (1986). Briefly, the amount adsorbed is equal to the amount of protons added (based on the known aliquots added to the mixture) minus the amount of protons remaining (based on the measured titratable acidity or alkalinity of the supernatant solution). The titration end point is identified based on a theoretical equivalence point analysis of the Al–H2O–Se system, as was described for similar systems elsewhere (Schulthess et al., 1998). The proton condition that needs to be solved for this matrix is
 |
= 2
+ 
for the samples with selenite, and = 
for the samples with selenate. The acidity constants used to identify the theoretical end points were: pK1 = 2.35 and pK2 = 7.94 for selenite, pK2 = 1.70 for selenate (Smith and Martell, 1976). The Al oxide samples did not have any inorganic C impurities present because of the extensive purging of the slurries with pure air (CO2-free air). All samples without Se and all samples with selenate had an end point at pH 6.85. The end point of samples with selenite varied from pH 6.85 (when the aqueous Se concentration was much lower than the aqueous Al concentration) to pH 8.35 (when the aqueous Se concentration was somewhat higher than the aqueous Al concentration). The amount of aqueous Al concentration present is easily extrapolated from the Al-induced buffering around pH 4.5 or 8.5 of the titration curves. The aqueous concentrations of selenate, selenite, and chloride were measured using a Dionex-300 IC/HPLC (Dionex, Sunnyvale, CA) with a Dionex IonPac AS4A SC column connected to a conductivity detector (CDM-3). The standard reference solutions were prepared using deionized water and analytical grade reagents. All tubes and bottles were acid washed, followed by an alkaline wash (pH 10), and rinsed with deionized water, which was necessary to eliminate the low-level background Cl contamination in our sample containers. All NaOH solutions were prepared CO2-free from saturated NaOH stock solutions, where the CO2-3 impurities are removed as precipitated solids. All the prepared NaOH and HCl solutions used in this study were kept CO2 free in bottles fitted with a CO2(g) trap (Thomas Scientific's Ascarite II, Swedesboro, NJ) on the cap.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Figure 1 shows that autochthonous chloride is present on the Al oxide (Lot 1), and the desorption of this surface impurity is pH dependent. As noted above, this oxide was washed with water only. Below pH 4.60, Cl anions are added to the matrix due to the addition of acid HCl. At pH 4.09, added aqueous Cl anions are adsorbed on the Al oxide surface and are represented in Fig. 1 as positive values (negative values represent desorption). As the pH increases, chloride is desorbed from the Al oxide. At pH 4.60 with no acid or base added in the samples, 0.23 µmol m-2 Cl anions are desorbed from the Al oxide. With the addition of NaOH, a maximum desorption of 0.70 µmol m-2 Cl is reached at around pH 6.8. Small amounts of Cl were detected in the second Al oxide slurry (Lot 2) at high pH in the presence of selenite (discussed further below). When using the third Al oxide slurry (Lot 3), no Cl was detected at high pH values (pH 9.4) in the absence of added selenate (data not shown). Noting the difference in their pretreatments (0, 3, or 4 x NaOH washing), one readily notes the sensitivity of the surface purity to the pretreatment choice and the difficulties present in obtaining a perfectly "clean" oxide surface.
|
With Lot 2, a very small amount of autochthonous Cl anions was present on the Al oxide, as noted by the small amount of Cl desorption at pH 9.1 in the presence of selenite (Fig. 1). A somewhat similar quantity of Cl desorption is expected in the absence of selenite at these high pH values. That is, no more Cl desorption is expected in the absence of SeO3 than that observed in the presence of SeO3 at high pH values. This is illustrated in Fig. 1 by the dotted line connecting the data.
Significant amounts of HCl are needed to lower the pH of the samples with selenite present (due to selenite's pK2 value of 7.94; Smith and Martell, 1976), and this results in the high concentration of total Cl present, which in turn yields an apparent increase in Cl adsorption in the presence of selenite (Fig. 1). This illusion was not present in the data for Lot 1, where the Cl concentrations on the Al oxide were already very high and the addition of selenate did not require the addition of excess HCl (selenate's pK2 value is 1.70; Smith and Martell, 1976). At low pH values, where differences in the aqueous HCl concentrations between the samples with and without selenite are small (Lot 2), the competitive effect of SeO3 adsorption on the Cl anions is noticeable (i.e., Cl adsorption is reduced in the presence of SeO3). In Fig. 1, a large open circle is placed over the symbols or on the lines to highlight the pH value where no acid or base additions were made to adjust the pH for each of the four series studied (i.e., large open circles identify the zero point of titration, ZPT).
The data shown in Fig. 1 confirm that the adsorbed Cl anion concentrations are not constant and may potentially interfere with subsequent calculations on the stoichiometry of proton/Se adsorption reactions. This will be discussed further in the various Stoichiometry sections below.
Adsorption of Selenate and Selenite
Chloride affects the adsorption of selenate (SeO4) anions. Figure 2 shows the adsorption of selenate on an Al oxide (Lot 1) in the presence and absence of various background NaCl solutions. Total removal of aqueous selenate corresponds to 0.7892 ± 0.0073 µmol m-2. At low ionic strength (0 or 0.01 M NaCl), there is a high removal (>95%) of selenate below pH 5.8. Increasing the concentration of the NaCl background electrolyte to 0.01 M causes a small decrease in selenate adsorption between pH 5.0 and 6.7. In the presence of 0.1 M NaCl, selenate adsorption is substantially reduced at all pH values. When the ionic strength is 0.01 or 0.1 M, there is a knee (or sharp bend) in the shape of the curves at pH 6.4. The initial adsorption edge (taken at 50% of the sharp rise in the data) is almost the same in the three different ionic strength conditions, which is pH 6.6 for 0 and 0.01 M NaCl and pH 6.5 for 0.1 M NaCl.
|
Figure 3 shows the adsorption of selenite (SeO3) on an Al oxide in the presence and absence of various background NaCl solutions. Here, the total removal of aqueous selenite corresponds to 0.8149 ± 0.0079 µmol m-2. The adsorption of selenite is nearly 100% below pH 7 at low ionic strength (0 and 0.01 M NaCl). As with selenate, selenite is not fully removed at low pH values in the presence of 0.1 M NaCl. From Fig. 1 and 3, it seems that chloride decreases the adsorption of selenite, while the addition of selenite decreases the adsorption of chloride, as would be expected in a competitive adsorption scenario. The adsorption edge for selenite is at pH 8.6. The higher adsorption edge for selenite (8.6) than for selenate (7.0) implies that selenite is held more strongly than selenate by the Al oxide surface. This is based on the fact that adsorption is pH dependent and that hydroxyl anions are involved in the adsorption mechanism. In other words, selenite can tolerate more competition from hydroxyl (OH-) anions than selenate.
|
Backtitration results for the Al oxide suspension (Lot 1) are shown for three salt concentrations (0, 0.01, and 0.1 M NaCl) in Fig. 4A . The lines drawn merely highlight the patterns observed for each data set. The zero point of salt effect (ZPSE) is the pH value where the proton adsorption curves at the different salt concentrations intersect. The ZPSE for this Al oxide is at pH 7.17 and has a horizontal shift of -0.73 µmol m-2. Figure 5A illustrates the proton adsorption by an Al oxide suspension that was washed with NaOH to remove the autochthonous Cl impurities (Lot 2). Upon washing the oxide suspension, the ZPSE has increased to a pH of 7.90, and the horizontal shift is closer to the zero position (-0.10 µmol m-2). The procedure was repeated for a NaOH-washed Al oxide suspension (Lot 3) (data not shown), where a smaller horizontal shift of around -0.08 µmol m-2 was also observed. Schulthess and Sparks (1987) observed a similar effect and attributed the horizontal shift to the removal of surface impurities. The removal of Cl anions from the Al oxide stock suspension (Lot 2 or 3 vs. Lot 1) reduces the magnitude of the horizontal shift of the ZPSE.
|
|
In the presence of selenite, the shapes of these adsorption curves are pushed to the right (toward more proton adsorption) (Fig. 5B). The new ZPSE value is poorly identified, but the patterns merge above pH 6.90 with a horizontal shift of 1.60 µmol m-2. The lines drawn merely highlight the patterns observed for each data set; the lines drawn also obey an assumed rule: do not cross another line unless its at the ZPSE. As with selenate, the adsorption of selenite (based on Fig. 3) is accompanied by an increase in H+ coadsorption by the Al oxide.
Proton/Chloride Adsorption Stoichiometry
A very important milestone toward understanding the impact of Cl anions on the coadsorption stoichiometry of protons/anions (such as selenate or selenite) by oxide surfaces is an understanding of the coadsorption stoichiometry of protons/chlorides. Schulthess and Sparks (1987) noted that the adsorption of each Cl anion on an Al oxide (using initial NaCl concentrations of 0 and 0.001 M from pH 2 to 8) is accompanied by the adsorption of a proton. This is again observed in Fig. 6
, where the adsorption of protons and the adsorption of Cl anions are following nearly identical patterns. The divergence between pH 4 and 5 on the water-washed Al oxide (Lot 1) data set is not observed when using the NaOH-washed Al oxide samples; it was also not observed by Schulthess and Sparks (1987) in their HCl/NaOH-washed Al oxide sample.
|
Proton/Selenium Adsorption Stoichiometries: Traditional Titration Approach
At this point, the ratio of proton adsorption to Se adsorption would normally be very straight forward. Briefly, the ratio (or coadsorption reaction stoichiometry) is merely the change of the amount of protons adsorbed divided by the amount of Se adsorbed at any given pH value
 | (3) |
H = change in proton adsorption due to the addition of selenate or selenite, and Se = amount of selenate or selenite adsorption at the specified pH value. The results of Eq. [3] on the selenate and selenite data are shown in Fig. 7 . The proton/selenate adsorption stoichiometry averages 2:1 above pH 6.5. There is much scatter in the selenate data above pH 7, where the amount of selenate adsorption is low (Fig. 2). Below pH 6.5, the protons/selenate ratios have a rapid decrease in values. The proton/selenite adsorption stoichiometry averages 2.2:1 above pH 5. Below pH 5, the protons/selenite ratios also have a rapid decrease in values.
|
The method described by Eq. [3] was used by Schulthess and Belek (1998), who observed a 3:1 H/CO3 (or 2:1 when expressed as H/HCO3) coadsorption stoichiometry on a Ti oxide, and by Schulthess et al. (1998), who observed a nearly 2:1 H/CO3 coadsorption stoichiometry on an Al oxide. Schulthess and Belek (1998) suggested that the cation desorption, which they also observed, was responsible for the unusually high stoichiometry of 3:1 H/CO3 on the Ti oxide. In other words, this method works well if there is no change in the adsorption patterns of all the other ions in solution, including those constituting the background salt solution. Schulthess and McCarthy (1990) did not notice any change in the Cl adsorption behavior by an Al oxide in the presence of weakly adsorbing anions, such as carbonate or acetate. Conversely, however, both selenate and selenite affect the adsorption of Cl, as was discussed above with Fig. 1, 2, and 3. Consequently, it is necessary to include the impact of Cl on the proton adsorption measurements in order to obtain the sought after proton/selenate or proton/selenite adsorption stoichiometries, which is the objective of the Modified Titration Approach section below.
Proton/Selenium Adsorption Stoichiometries: Traditional pH-Stat Approach
Perhaps the most common method of obtaining proton/anion adsorption stoichiometries is the pH-stat method, and it is therefore briefly included here for comparative purposes. In this method the amount of acid needed to maintain a specific pH value is measured after two solutions are mixed (one the adsorbent slurry, the other the adsorbate solution), both at the same specific pH value. The amount of acid added corresponds to the change in proton adsorption. The stoichiometry calculation is completed following the measurement of the amount of anion adsorption by the solid phase at equilibrium. There are a few problems with this method:
The results of the pH-stat method using selenite and Al oxide (Lot 2) are shown in Fig. 7. Due to selenite speciation, the actual measured values were one unit lower (e.g., 0.79:1 rather than 1.79:1 at pH 4.98) when expressed as H/HSeO3 rather than H/SeO3. The closeness of the values from the pH-stat method to the other values suggests that the pH-stat method is a reasonable option to use for stoichiometric studies if various precautions are in place. The low reactivity of Cl anions in the high pH region seems to avoid complications for the pH-stat method. The pH-stat method, however, is also not reliable at low pH values when a Se-induced variable Cl adsorption situation arises. Wijnja (unpublished data, 1999) also measured the proton/selenate adsorption stoichiometry on Al oxide in 0.011 M NaCl using the pH-stat method. His very low value of 0.96:1 at pH 6 is remarkably similar to our salt-containing values discussed in the section above.
The presence of background Cl anion concentrations clearly affects the pH-stat values, which suggests a need to improve the method. As discussed in the next section, measurement of the Cl anions in solution may resolve these problems if precautions are in place that correct for any anion contamination leaching from the reference electrode. In any case, the pH-stat method is not as informative as the backtitration technique because it lacks the detail on the pH-dependent behavior of proton adsorption. The data in Fig. 4 and 5, for example, which are very closely linked to other surface charge phenomena, would not be available from the pH-stat results.
Proton/Selenium Adsorption Stoichiometries: Modified Titration Approach
As noted in Fig. 6, the adsorption of Cl anions at low pH is accompanied by the coadsorption of protons following a 1:1 stoichiometry. Accordingly, as the NaCl concentration is increased, Fig. 4A and 5A show that the amount of protons adsorbed will also increase for any given pH value below the zero point of salt effect. However, at low pH, Fig. 4 and 5 show that the increase in proton adsorption in the presence of Se anions decreases as the ionic strength increases. For example, at a given low pH value, the data for 0.1 M NaCl in Fig. 4A and 5A are shifted to the right in the presence of Se anions in Fig. 4B and 5B, but the amount of this shift to the right is less than those observed by the data with lower NaCl concentrations. This is because, as noted in Fig. 1, the amount of Cl adsorbed varies in the presence of selenate or selenite. Consequently, the impact of background Cl anions on proton adsorption stoichiometries is not a constant amount that can be easily factored out here. The suggested numerical analysis to resolve this problem is as follows
 | (4) |
Cl is the change in chloride adsorption due to the addition of selenate or selenite, and H and Se are the same as defined above for Eq. [3]. In other words, the proton isotherms illustrated in Fig. 4B and 5B are a result of the coadsorption of protons with Se anions plus the coadsorption of protons with Cl anions. The numerator in Eq. [4] adjusts the experimentally measured proton adsorption value to represent the change in proton adsorption due to Se anions only.
The H value is equal to the horizontal shift in the H values shown in Fig. 4A and 4B, or 5A and 5B, for each data set for a given pH condition. The Cl value is equal to the vertical change in the Cl values shown in Fig. 1. Accordingly,
 | (5) |
H,1 = Cl,1 (this is justified based on the 1:1 H/Cl stoichiometry previously discussed), Eq. [5] simplifies to
 | (6) |
It is advantageous to let H,1 = Cl,1 because it is difficult to reconstruct the exact background Cl concentration of a given sample with Se present in another sample without Se present. Selenite samples are particularly sensitive to this problem because of their change in aqueous speciation around pH 7.94 (Smith and Martell, 1976). The background Cl concentration originates from the fixed addition of NaCl solutions for ionic strength adjustments and from the variable amounts of added HCl for the pH adjustments.
Figure 7 plots the results based on Eq. [6] for selenate and selenite. The numerical average of the proton/selenate coadsorption stoichiometry from pH 4 to 7 is (1.96 ± 0.11):1. The average of the proton/selenite values from pH 2.5 to 10 is (1.90 ± 0.11):1.
Surely, when the change in Cl adsorption in the presence of Se anions is minimal, namely above pH 6, then the results of Eq. [3] and [6] should agree with each other. This is observed in Fig. 7 for the selenate data, where the two methods converge at pH 6.5 to 7; above pH 7 the selenate adsorption is very low and much scatter exists in the resultant calculations from both equations. For unknown reasons, this was not observed for selenite, where using Eq. [3] predicted higher stoichiometric ratios than Eq. [6].
The acidic pH value where the results of Eq. [3] begin to diverge downward from the norm is a rough indication of where the Cl impact (more specifically, the variable Cl-adsorption impact) on the system needs to be taken more seriously. The H/Se data using Lots 2 or 3 (NaOH-washed) were collected on a cleaner surface and their numerical results diverged from the norm at a lower pH value than those calculated using Lot 1 (water-washed). The impact of the variable Cl adsorption on these systems studied is clearly demonstrated by the numerical correction (using Eq. [6]) of the otherwise diverging values in Fig. 7.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The coadsorption of protons with Cl anions is an outer-sphere adsorption mechanism
 | (7) |
Note that the weakly held Cl anion is removed from solution along with the solid phase when the suspension is centrifuged and, in this sense, is a true component of the solid–liquid interface. More importantly, the adsorption of Cl can compete with the adsorption of other anions, such as selenate and selenite, particularly when present at high concentrations. This competition for surface sites typically lowers the individual adsorption levels of each anion species, while the sum of all anions adsorbed increases. When an anion modifies the adsorption levels of Cl, which is easy to do because of its weak adsorption affinity, the adsorption levels of protons are also equally modified. As Cl desorbs, protons also desorb and lower the net proton coadsorption values. If left uncorrected, then the specific proton/anion coadsorption stoichiometry value will be too low, and the subsequently suggested anion adsorption mechanisms may be erroneous.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication February 12, 1999.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
-Al2O3 studied by fourier transform Raman and infrared spectroscopy: I. Initial results. Spectrochim. Acta 49A:691–705.-Al2O3 and BeO. Geochim. Cosmochim. Acta 50:1847–1860.[Web of Science]
-FeOOH. Science 238: 783–786.-Al2O3 and kaolinite. Soil Sci. Soc. Am. J. 60:442–452.-Al2O3/water interface. Spectrochim. Acta Part A 55:861–872.This article has been cited by other articles:
|
|
 |
|
  U. K. Saha, L. M. Kozak, and P. M. Huang KINETICS OF CITRATE-INDUCED SELENITE DESORPTION FROM MONTMORILLONITE AS AFFECTED BY COMPLEXATION WITH HYDROXYALUMINUM AND HYDROXYALUMINOSILICATE IONS Clays and Clay Minerals, February 1, 2007; 55(1): 71 - 88. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. K. Saha, C. Liu, L. M. Kozak, and P. M. Huang Kinetics of Selenite Adsorption on Hydroxyaluminum- and Hydroxyaluminosilicate-Montmorillonite Complexes Soil Sci. Soc. Am. J., July 1, 2004; 68(4): 1197 - 1209. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
