| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Soil Science, Univ. of Saskatchewan, 51 Campus Dr, Saskatoon, SK, Canada, S7N 5A8
* Corresponding author (huangp{at}sask.usask.ca).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Abbreviations: CEC, cation-exchange capacity • EGME, ethylene glycol mono-ethyl ether • FTIR, Fourier transform infrared • HAS1, hydroxyaluminosilicate with Si/Al molar ratio = 0.24 • HAS2, hydroxyaluminosilicate with Si/Al molar ratio = 0.48 • HGAAS, hydride generation atomic absorption spectrometry • HyA, hydroxyaluminum • LMWOAs, low molecular weight organic acids • Mt, montmorillonite • p-jump, pressure jump • SAP, selenite adsorption potential • SE, standard error • Vt, vermiculite • XRD, X-ray diffraction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Adsorption–desorption at the soil particle–water interface is extremely important in affecting the transport, bioavailability, and fate of different metals, metalloids, and other inorganics in soil and associated environments. Soils and sediments have a remarkable ability to adsorb Se. Therefore, it is essential that one understands the kinetics and mechanisms of Se adsorption on soil colloidal surfaces including both pure clays and natural soil systems.
Like other anions, selenite 
can adsorb on the surface of Al oxides by ion-pair formation with positively charged surface sites or by ligand exchange with surface hydroxyls after protonation (Hingston et al., 1972; Okazaki et al., 1989; Goldberg et al., 1996; Schulthess and Hu, 2001). Selenium adsorption on short-range ordered hydrous oxides has been studied to a considerable extent, establishing that short-range ordered hydrous Al and Fe oxides are an important soil constituent in SeO3–2 adsorption (Levesque, 1974; John et al., 1976; Rajan, 1979; Balistrieri and Chao, 1987; Zhang and Sparks, 1990; Hayes et al., 1987). Selenite is also adsorbed onto an allophane clay, which consists of an aluminosilicate core coated with HyA (Rajan and Watkinson, 1976). Low-molecular weight organic acids (LMWOAs) showed remarkable competitiveness with SeO3–2 for adsorption onto a short-range ordered Al hydroxide (Dynes and Huang, 1995, 1997). Such competitive effects were, however, greatly influenced by the structure and functionality of the LMWOAs (Dynes and Huang, 1995, 1997).
The mechanisms of SeO3–2 adsorption onto the hydrous oxides of Al and Fe have also been postulated by different investigators. According to Dynes and Huang (1995), for short-range ordered Al hydroxides, the electrostatic attraction between HSeO3– and edge Al-OH2+ groups leads to the initial formation of an outer-sphere surface complex, AlOH2+–HSeO3–. Subsequent deprotonation from the surface complexed HSeO3– could form AlOH2+–SeO3–2 surface complex, which could finally form the inner-sphere surface complex, Al-SeO3–, through a ligand exchange reaction (Dynes and Huang, 1995). A similar mechanism has been proposed for adsorption on goethite by Zhang and Sparks (1990) using pressure-jump (p-jump) relaxation techniques. Spectroscopic studies have also shown that SeO3–2 adsorption by oxides occurs via ligand exchange forming inner-sphere surface complexes (Parfitt, 1978; Hayes et al., 1987).
The two active Al species, HyA (Rich, 1968; Barnhisel and Bertsch. 1989; Inoue and Yoshida, 1990) and HAS (Lou and Huang, 1988, 1994, 1995; Huang et al., 2002) ions are ubiquitous in acidic soils and aquatic environments. These HyA and HAS ions interact with both organic and inorganic soil components and play a significant role in regulating the mobility of plant nutrients and pollutants in acidic soil environments (Huang et al., 2002). In 2:1 type expansible silicate clays, not only HyA, but also the HAS interlayers are common in acid to slightly acid soils (Lou and Huang, 1988, 1994, 1995; Matsue and Wada, 1988; Bautista-Tulin and Inoue, 1997). Laboratory investigations concluded that irreversible adsorption of HyA and HAS on the silicate surface results in very significant modification of electrochemical and mineralogical properties of the host clays (Inoue and Satoh, 1992, 1993; Sakurai and Huang, 1998), which in turn results in a significant change in the ion adsorption and exchange properties of the host clays. Such changes include phosphate retention (Saha and Inoue, 1997a; Saha et al., 1998), K+ and NH4+ fixation (Saha and Inoue, 1997b, 1998), K+/Ca2+ and NH4+/Ca2+ selectivities (Kozak and Huang, 1971; Saha et al., 1999), adsorption of Cu (Harsh and Doner, 1984), adsorption–desorption of Cd (Keizer and Bruggenwert, 1991; Sakurai and Huang, 1995, 1996; Lothenbach et al., 1997, 1998; Taniguchi et al., 2000), and adsorption selectivity of Pb, Zn, and Cd (Saha et al., 2001).
However, due to the absence of any direct evidence, our current understanding on the adsorption behavior of SeO3–2 onto HyA-Mt and HAS-Mt materials relies on mere gazing through extension of the existing literature dealing with the adsorption of other anions (phosphate, sulfate, and organic ligands) onto these clays or from the SeO3–2 adsorption onto the amorphous hydrous oxides of Al and Fe and HAS type discrete adsorbents. Therefore, the adsorption kinetics of SeO3–2 onto Mt, HyA-Mt, and HAS-Mt were investigated in this study to elucidate the influence of HyA- and HAS-interlayering and/or coatings on Mt on the SeO3–2 adsorption phenomena. The effect of temperature on SeO3–2 adsorption kinetics was also included to evaluate the pre-exponential factor (an index of the frequency of SeO3–2 collision with clay surface) and activation energy involved in the adsorption phenomena.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The HyA and HAS ionic solutions were prepared by mixing orthosilicic acid, AlCl3, and NaOH solutions in the following manner. For the HyA ionic solution, 0.1 M AlCl3 was titrated with 0.1 M NaOH at the rate of 0.2 to 0.5 mL min–1 with continuous stirring to give a NaOH/Al molar ratio of 2.0. For the two HAS ionic solutions, orthosilicic acid prepared from tetraethyl orthosilicate (Farmer et al., 1979) was mixed with a 0.1 M AlCl3 solution to obtain a Si/Al molar ratio of 0.50 and 1.00. The solutions were then titrated with 0.1 M NaOH in the manner described above. The OH/Al and Si/Al molar ratios, pH, and Al concentrations of the HyA and HAS parent solutions before reaction with Mt are presented in Table 1.
|
Characterization of the HyA/HAS-Mts
The negative charge characteristics of Mt and HyA/HAS-Mts were determined through estimation of Ca2+ retained in the pH range of 4 to 7.5 following the procedure described by Wada and Okamura (1980), except that, instead of NH4Cl, CaCl2 was used for saturating the samples to avoid consequence of cation fixation (if any). The point of zero salt effect (PZSE) of the HyA/HAS-Mts was determined by a modified salt titration method (Sakurai et al., 1988). The total surface area of the clay samples was determined by the ethylene glycol mono-ethyl ether (EGME) method (Eltantawy and Arnold, 1973) and the external surface area by adsorption of N2 gas at 78 K using a BET surface area analyzer (Shibata P-850, Shibata-Kagaku Co., Tokyo). The internal surface area was calculated as the difference between total and external surface areas.
The X-ray diffraction (XRD) analysis of K- and Mg-saturated clay specimens was performed with a Rigaku X-ray diffractometer (RAD-1A) (Rigaku Co., Tokyo) using Fe-filtered CoK
radiation generated at 30 kV and 10 mA. To further characterize the HyA/HAS-Mts, the K-saturated (and air-dried) specimens were heated at 383, 573, and 823 K and the Mg-saturated specimens were solvated with glycerol, and their X-ray diffractograms recorded after each treatment.
The clay specimens were also examined by infrared (IR) absorption spectroscopy. One milligram of sample was mixed with 200 mg of KBr and the KBr pellets were then examined by a Fourier-transform infrared (FTIR) spectrometer (Bio-Rad, Cambridge, MA).
Kinetics of Selenite Adsorption
The SeO32– adsorption kinetics on Mt and HyA/HAS-Mts was investigated at 288, 298, 308, and 318 K by the conventional batch method. The sodium salt of SeO32– (Na2SeO3) was used and all the experiments were performed in duplicate.
One hundred milligrams of the clay specimen was added to a 500-mL Pyrex Erlenmeyer flask containing 170 mg NaNO3 dissolved in 180 mL of double-deionized water and mixed using a magnetic stirrer. The pH of the suspension was adjusted to 4.5 using 0.1 M HNO3 or 0.1 M NaOH. The suspension was then shaken for 24 h in a constant temperature water bath with another interim adjustment of pH to 4.5. After shaking for 24 h, the suspension pH was readjusted to 4.5 and the volume of water was adjusted to 195 mL by weighing the reaction vessel. A 5-mL aliquot of 1.0 mM Na2SeO3 solution (having pH 4.5) was then carefully added to the suspension while being stirred (magnetic stirrer) and the reaction vessel was then immediately transferred to a constant temperature water bath and shaken. Thus, the resultant suspension in the reaction vessel had a clay concentration of 0.5 g L–1 and a SeO3–2 concentration of 0.025 mM. At various times between 0.083 and 24 h (0.083, 0.167, 0.333, 0.5, 0.75, 1, 1.5, 2, 3, 8, 16, and 24 h), a 10-mL aliquot of the suspension was withdrawn from the system while being stirred (to maintain a constant solid/solution ratio) and filtered through a 0.1-µm millipore membrane filter in <15 s using vacuum. The pH and SeO3–2 concentration in the filtrates were determined. The amount adsorbed after a particular reaction period was determined from the difference between the concentration of SeO3–2 in the parent solution and that remaining in the filtrate. Determination of SeO3–2 was performed by hydride generation atomic absorption spectrometry (HGAAS) as described by Huang and Fujii (1996). A Varian VGA-77 continuous flow hydride generator (Varian Australia Pty Ltd., Mulgrave, Victoria) coupled to a Spectra AA-220 AAS (Varian Australia Pty Ltd) was used. A freshly prepared sodium borohydride solution, 0.6% (w/v) and 7 M HCl were used for hydride vapor generation by the VGA apparatus. The detection limit was 1.27 x 10–8 M.
Statistical Analysis
The linear forms of different kinetic equations were applied to the adsorption data and their goodness of fit was evaluated based on the r2, level of significance (p), and standard error (SE). The comparisons of the amounts of SeO3–2 adsorbed, adsorption rate constants, activation energies, and pre-exponential factor values in different cases were performed using the least significant difference (LSD) test. The LSD values were calculated based on SE and t values at appropriate degrees of freedom at 95 and 99% confidence levels. The LSD values for the rate constants and those of activation energies were calculated based on the standard errors of the slopes for the rate equation and Arrhenius equation, respectively. The LSD values for the pre-exponential factor were calculated based on the standard error of the intercept of the Arrhenius equation.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The mass of the HyA component adsorbed on Mt can be approximated by using Al(OH)3 for the structural formula as adopted by Harsh and Doner (1984). The mass of the adsorbed Al in the case of the HyA-Mt then becomes 0.088 g of Al(OH)3 g–1 of Mt or 0.081 g of Al(OH)3 g–1 of HyA-Mt. In the present systems, the mass of the HyA ions adsorbed per unit mass of the Mt was considerably lower than the HyA-Mts prepared by Harsh and Doner (1984), Sakurai and Huang (1998), and Taniguchi et al. (2000). Such differences may be attributed to the differences in preparation methods, cation-exchange capacity (CEC) of the host clays, the amounts of Al added per unit mass of the clay, and the initial basicities of the original solutions.
As calculated using the method of Hsu (1968), the HyA and HAS materials adsorbed on Mt had OH/Al ratios ranging from 2.64 to 2.79 (Table 1). Barnhisel and Bertsch (1989) calculated the OH/Al ratios for a variety of preparations and reported that all the values were around 2.7. The HyA/HAS-Mts used by Harsh and Doner (1984), Sakurai and Huang (1998), and Taniguchi et al. (2000), however, were somewhat higher with the average OH/Al ratios ranging from 2.80 to 2.93.
The CEC of untreated Mt shows only a slight pH-dependence (please see the change in CEC of Mt with pH rise from 4.0 to 7.5 in Table 1). Fixation of both HyA and HAS ions resulted in a substantial reduction in permanent negative charge (CECs at pH 4.0 in Table 1) and a substantial increase in pH-dependent negative charge (Table 1). However, the CEC of the two HAS-Mts exhibited weaker pH-dependence than that of the HyA-Mt (detailed data not shown). A drastic reduction in the CEC of Mt due to HyA and HAS interlayering and coatings is attributed to a nonexchangeable occupancy of the cation-exchange sites and physical blocking or steric hindrance of the cation-exchange process by the HyA/HAS polymers (Rich, 1968; Inoue and Satoh, 1993). A substantial increase in the pH-dependent charge in the HyA/HAS-Mts is attributable to the formation of AlOH–0.5 groups at the edges of the HyA/HAS polymers through a deprotonation process in response to increasing pH.
The PZSE values for the HyA/HAS-Mts varied within a narrow range of 4.4 to 4.7. The PZSE values for the two HAS-Mts were to some extent lower than that of the HyA-Mt. Hydroxides or oxides of Si have much lower PZSE values (<2.0) than those of Al (>8.5) (Parks, 1965). Huang et al. (2002) reported that the presence of specifically adsorbed phosphate, silicate, or organic acids could lower the point of zero charge (PZC) of gibbsite and poorly crystalline Al hydroxides. The lower PZSE values of the HAS-Mts compared with the HyA-Mt as described above is, thus, logical.
The adsorption of HyA or HAS on Mt resulted in lower total and internal surface areas and a considerable increase in external surface area (Table 1). The marked reduction of internal surface area is predominantly due to the occupancy of the interlayer spaces of Mt by the HyA or HAS ions (Rich, 1968; Inoue et al., 1988; Lou and Huang, 1988). Moreover, the penetration of large EGME molecules into the Mt interlayers could be hindered through physical blocking due to an "atoll" like distribution of the HyA and HAS ions in the Mt interlayers, where the interlayer materials remain concentrated near the edges (Frink, 1965). On the contrary, a marked increase of external surface area of the HyA/HAS-Mts may be attributed to the possible formation of irregular external planar surfaces and edges by the adsorption of HyA and HAS ions on Mt and/or the formation of discrete amorphous Al hydroxides or aluminosilicates during the interaction of Mt with HyA and HAS ions. Similar results were also observed in previous studies (Inoue and Satoh, 1993; Sakurai and Huang, 1998).
The K-saturated HyA-Mt had an expanded d(001) spacing of 1.47 nm in comparison with 1.22 nm for the K-saturated Mt (Table 2), suggesting that at least a portion of the HyA was fixed in the interlayer. The d(001) spacings of K-saturated and heated (383–823 K) HyA-Mt indicate a considerable heat stability of the synthesized interlayers. When heated to 573 K, a broad band from 1.19 to 1.32 nm appeared. This implies that the distribution of HyA ions was not uniform throughout the interlayer space.
|
The XRD results of the air-dried and glycerated Mg-saturated clays show that upon solvation, the HyA/HAS-Mts expanded from 1.55 to 1.79 nm, indicating that the presence of HyA and HAS ions in the interlayer did not lead to any irreversible bonding between the silicate layers (Table 2). This is consistent with the behavior observed for most preparations of hydroxy-interlayered smectites (Harsh and Doner, 1984).
The FTIR spectrum of the original Mt (Fig. 1a) exhibited major IR absorption bands at 3637 cm–1 (Al-OH stretching), 3458 cm–1 (OH stretching of adsorbed water), 1640 cm–1 (OH bending of adsorbed water), 1048 cm–1 (Si-O stretching), and 919 cm–1 (Al-OH bending), which are very close to those reported by White and Roth (1986) and Sakurai and Huang (1998). Since the systems were not free from atmospheric CO2, some solution carbonates may be adsorbed on the mineral phase, which would partially account for the broad peaks at 3400 to 3500 cm–1 on the FTIR spectra (Fig. 1 and 2) as explained by Wijnja and Schulthess (1999) using DRIFT spectra for carbonate adsorbed on 
-Al2O3. The Si-O-Al and Si-O-Mg bending vibrations were seen at 525 and at 473 cm–1, respectively, which were also very close to those reported by Lou and Huang (1994) and Sakurai and Huang (1998). The IR absorption bands at 1085, 794, and 627 cm–1 are characteristic to the mineral cristobalite (Marel and Beutelspacher, 1976), which was present in the Mt as impurity (Inoue and Satoh, 1993). The spectra of HyA-Mt, HAS1-Mt, and HAS2-Mt (Fig. 1b, c, and d, respectively), however, did not show any special features for the adsorbed HyA and HAS species.
|
|
Amount of Selenite Adsorbed
The amounts of SeO3–2 adsorbed onto the Mt, HyA- and HAS-Mts after selected reaction periods at 298 K are given in Table 3. It should be noted that SeO3–2 adsorption invariably reached a quasi-equilibrium at the end of the 24-h reaction period (Fig. 3). During any particular reaction period, the amounts of SeO3–2 adsorbed onto the different clays were considerably different from each other. The Mt always adsorbed a much lower amount of SeO3–2 than the HyA/HAS-Mts, suggesting that the HyA or HAS polymers were actively involved in SeO3–2 adsorption. A greater adsorption of SeO3–2 on Mt compared with kaolinite as observed by Frost and Griffin (1977) was also attributed to the presence of interlayer HyA polymers in Mt.
|
|
Out of 50 mmol of SeO3–2 added per kilogram of clay, the SeO3–2 adsorbed by the end of the 24-h reaction period was around 9% for Mt, 25% for HAS2-Mt, 31% for HAS1-Mt, and 50% for HyA-Mt, indicating that the Mt adsorbed the lowest fraction of the SeO3–2 added among the samples studied (Table 3). The negatively charged nature of Mt would impede the approach of the negatively charged SeO3–2 anion to the adsorption sites on its edges due to anion exclusion effects. This is in accord with the earlier observations by Huang (1975) in the case of arsenate adsorption by micas with different degrees of K-depletion. As compared with muscovite and biotite, a lower adsorption of arsenate by K-depleted muscovite and biotite was attributed to the exclusion of arsenate anions by the K-depleted, partially opened mica interlayers (Huang, 1975). Therefore, adsorption of SeO3–2 on Mt as an uncharged selenious acid (H2SeO3) molecule merits attention as a possible adsorption mechanism. Unlike SeO3–2 anions, the uncharged H2SeO3 molecules would not have to overcome the repulsive action of Mt while approaching the adsorption sites on its edges. Selenious acid in solution exists in the following equilibrium:
 |
Thus, at pH 4.5, 98.7% of the total Se in solution exists as HSeO3–2 anions, while only 1.3% as undissociated H2SeO3 molecules. However, once the adsorption of uncharged H2SeO3 molecules begins, its level in the solution depletes; the equilibrium shifts to the left inducing regeneration of undissociated H2SeO3 molecules, and the adsorption can continue. But, as an alternative mechanism, the possibility of adsorption via a direct ligand exchange between the SeO3–2 anion and OH2 and/or OH (after protonation) ligands on the edges of Mt should not be ruled out.
On the HyA/HAS-Mts, the polymeric HyA or HAS cations were adsorbed in the interlayer space (Tables 1 and 2) and apparently also on the external planar surfaces and/or on the edges of Mt, substantially blocking the permanent negatively charged surfaces of Mt but considerably increasing the variable charged surfaces by the interlayerings and coatings with the polymers. The HyA/HAS interlayerings and coatings also caused aggregation of the particles and reduction in the total surface area (Table 1).
A substantial reduction in negative charge and total surface area due to HyA/HAS interlayerings and coatings (Table 1) should minimize the anion exclusion effects of Mt. Further, the adsorption experiment was performed at pH 4.5, which were below the PZSE of HyA-Mt and either equal or slightly above the PZSE of the two HAS-Mts. Therefore, the surfaces of the HyA/HAS-Mts carried a substantial positive charge at the pH of the adsorption experiments. Consequently, the impeding effect of anion exclusion on the approach of the SeO3–2 anions to the adsorption sites on the HyA/HAS-Mts should be substantially decreased. This would facilitate the electrostatic adsorption of SeO3–2 anion on AlOH2+ sites at the edges or even through direct ligand exchange with the edge AlOH2+ groups of the adsorbed HyA or HAS polymers. However, the residual negative charges (CEC) in the Mt (Table 1) may still impede the adsorption of SeO3–2 by the adsorption sites on the HyA or HAS polymers residing in the interlayer spaces of Mt due to anion exclusion effect. This is consistent with the explanation given by Huang (1975) for the difference in arsenate adsorption on K-depleted and original micas treated with HyA. In the original micas with completely collapsed interlayers, the HyA adsorption took place on the edges and external planer surfaces only (Huang and Kozak, 1970), while, a partial opening of the mica interlayers due to K-depletion allowed HyA adsorption in the interlayers as well as on the edges and external planer surfaces (Kozak and Huang, 1971). As compared with HyA-treated original micas, a lower adsorption of arsenate anion by the HyA-treated K-depleted micas was attributed to the inaccessibility of the arsenate anions to adsorption sites on HyA polymers due to the anion exclusion effect (Huang, 1975).
Bowden et al. (1980) reported that HSeO3– and SeO3–2 ions adsorbed on goethite occupied an area of 15 and 50 Å2, respectively. This implies that the HSeO3– and SeO3–2 anions would have their diameters as 4.37 and 7.98 Å, respectively. In view of the d(001) spacings of the air-dried HyA/HAS-Mts and Mt (Table 2), the sizes of the SeO3–2 anions could be another important factor impeding their accessibility to the HyA/HAS in the Mt interlayers. Thus, the adsorption of SeO3–2 anion on the edge AlOH2+ and AlOH (after protonation) groups of the HyA or HAS polymers residing on the external planer surfaces and on the edges of Mt is a more possible mechanism in the case of the HyA/HAS-Mt.
It has been shown that the edge surfaces [Al(H2O) (OH)] of gibbsite are much more reactive than its 001 faces (Al-OH-Al hydroxyls) (Parfitt et al., 1977). However, the poorly ordered HyA polymers are expected to have substantial structural defects on their 001 faces and the Al-OH groups after protonation around these defect sites would be as reactive as their edge counterparts. In HAS ionic solution, the Si(OH)4 molecules would be complexed through the reaction of Si-OH groups with edge sites of HyA ions, that is, Al-OH/OH2 groups. Thus, silication in HyA ions should reduce the frequency distribution of Al-OH or Al-OH2 groups on the edge and defect sites (which could otherwise be highly reactive to SeO3–2) on the HyA ions (Wada and Wada, 1980; Sterte and Shabtai, 1987). Consequently, HAS-Mts were lower in their reactivity to SeO3–2 compared with HyA-Mt counterpart. The substantial reduction in SeO3–2 adsorption by the HAS-Mt with the higher Si/Al ratio (Table 3) indeed supports this explanation.
As reported by Dynes and Huang (1995) for short-range ordered Al hydroxides, the reaction of HSeO3– with the edge Al-OH2+ groups of HyA and HAS components of the HyA/HAS-Mts in the present study could lead to the initial formation of AlOH2+–HSeO3– surface complex due to electrostatic attraction. The surface complexed HSeO3– could be deprotonated rapidly to form AlOH2+–SeO32– surface complex, which could finally form the inner-sphere surface complex, Al–SeO3–, through a ligand exchange reaction (Dynes and Huang, 1995). Zhang and Sparks (1990) using p-jump relaxation techniques indicated that a similar surface complex was formed when SeO3–2 was adsorbed onto goethite as did Mikami et al. (1983) when phosphate was adsorbed onto a -Al2O3 surface. Spectroscopic studies have also shown that SeO3–2 adsorption by oxides occurs via ligand exchange forming inner-sphere surface complexes (Parfitt, 1978; Hayes et al., 1987). The ligand exchange of HSeO3– with the protonated surface hydroxyl groups 
should decrease surface positive charge of the HyA/HAS polymers, which in turn reduces the electrostatic attraction of SeO3–2 for the HyA/HAS surfaces with increasing surface coverage by SeO3–2, thereby slowing down the adsorption process.
Based on the pH rise after an initial fast adsorption of SeO3–2 on the Al precipitate, Dynes and Huang (1995) speculated that adsorption took place through the ligand exchange of SeO3–2 with surface hydroxyl (Al-OH) groups. However, they stated that ligand exchange of SeO3–2 with surface hydroxyl (Al-OH) groups could occur only after protonation of the surface hydroxyls by protons dissociated from H2O molecules (Hingston et al., 1972), since inception of adsorption takes place through the formation of an initial surface complex 
by electrostatic attraction. In agreement with the study of Dynes and Huang (1995), in the present study also, the pHs of the adsorption systems of Mt, HyA-Mt, HAS1-Mt, and HAS2-Mt after a 24-h reaction period raised from a common initial value of 4.50 to 4.62 ± 0.05, 4.63 ± 0.07, 4.62 ± 0.06, and 4.64 ± 0.07, respectively. This observation indicates that ligand exchange of SeO3–2 with surface hydroxyl (Al-OH) groups occurred after protonation of the surface hydroxyls in the adsorption systems of the present study, which is in accord with the findings of Schulthess and Hu (2001).
Rates of Selenite Adsorption
As all the systems invariably reached a quasi-equilibrium with respect to SeO3–2 adsorption by the end of the 24-h reaction period (Table 3, Fig. 3), these figures provide an indication of the selenite adsorption potential (SAP) of the clays. The time functions of SeO3–2 adsorption on the clays at different temperatures are shown in Fig. 3. The adsorption followed a typical kinetic process with an initial rapid decrease in SeO3–2 concentration in the bathing solution followed by a slower decrease and a quasi-equilibrium at all temperatures
Selenite adsorption onto the clays was very fast, especially for the HyA/HAS-Mts, where 39 to 54 and 64 to 89% of their SAP was satisfied within the first 0.083- and 0.5-h reaction periods, respectively. In contrast, only 13 and 44% of the SAP of Mt was satisfied after these reaction periods (Table 3). However, after 3 h, 86 to 100% of the SAP was achieved for both the Mt and the HyA/HAS-Mts. Generally, the concentration of SeO3–2 in solution decreased rapidly in the 0.083- to 0.5-h period and more slowly in the 0.5- to 3-h reaction period, indicating that SeO3–2 adsorption consisted of a multiple rate process (Fig. 3). Therefore, based on the steepness of the adsorption curves, the reaction kinetics for SeO3–2 adsorption was divided into an initial fast reaction (0.083–0.5 h) followed by a slow reaction (0.5–3 h). The multiple rate characteristics of SeO3–2 adsorption may be related to the heterogeneity of the adsorption sites. Site heterogeneity is believed to result from the different accessibility of surface pores and sites with different adsorption affinity and binding strength (Benjamin and Leckie, 1981; Madrid and de Arambarri, 1985; Heimestra et al., 1989; Liu and Huang, 2000).
Further, SeO3–2 adsorption by a ligand exchange reaction with protonated surface hydroxyl groups of HyA/HAS polymers would decrease the positive surface potential and increase the negative charge as was reported for phosphate adsorption onto goethite (Parfitt and Atkinson, 1976). This would enhance the electrostatic repulsion as the adsorption proceeded forward (Hansmann and Anderson, 1985; Shang et al., 1993). A decrease in the more accessible and reactive sites, coupled with a decrease in the positive charge on the HyA/HAS polymers, may account for the slower adsorption of SeO3–2 following the initial fast reaction (Fig. 3).
The kinetic and empirical equations, including the zero-, first-, and second-order rate equations, the power function (modified Freundlich), Elovich, and the parabolic diffusion equations were applied to the SeO3–2 adsorption data. The goodness of fit of the equations to the data was evaluated by using the correlation coefficient (r2), probability (p), and SE of linear regression analysis. The values of these three parameters obtained by applying different models to the eight sets of kinetic data obtained in the fast (0.083–0.5 h) and slow (0.5–3 h) reactions for four different clays at 298 K (the values for other three temperatures are not shown) are given in Table 4.
|
0.05) with three exceptions. This is consistent with the existing literature dealing with heterogeneous systems where it has often been shown that a number of kinetic and empirical models seem to describe satisfactorily the adsorption rate data, if r2, p, and SE are used as the indices to judge their goodness of fit (Chien and Clayton, 1980; Onken and Matheson, 1982; Sparks and Jardine, 1984). Despite this, there often is not a consistent relation between the equation that gives the best fits and the physicochemical and mineralogical properties of the sorbents being studied (Sparks, 2003). There is no correlation between the applicability of any of these equations and the nature of the processes actually involved; dissimilar processes are often fitted by the same equation and similar processes are fitted by different equations (Aharoni and Sparks, 1991). Nevertheless, the parameters of the chosen model do provide some meaningful tools for comparison of the rates of adsorption processes on different adsorbents. It is evident in Table 4 that in the case of slow reaction with HyA-Mt, the power function, Elovich, and the parabolic diffusion models gave the r2 values with p > 0.05. Such observation precludes the use of these three models as a common one to describe the whole set of kinetic data. Moreover, the parameters derived by the Elovich and power-function models are not well defined physicochemically and the equations may not provide the rate constants (Kuo and Lotse, 1973; Bolan et al., 1985). The parabolic diffusion equation is often used to indicate that diffusion-controlled phenomena are rate-limiting and only "apparent" diffusion coefficients can be obtained (Sparks, 1999).
All of the three ordered models (zero-order, first-order, and second order) gave r2 values at p 0.05. Among the three ordered models, the second-order model, however, gave the highest r2 and lowest p values (Table 4). Among the six models tested, the SE values normalized to the same unit (amount adsorbed per unit mass) were the lowest for the second-order model. To our belief, more emphasis should be given on the normalized SE value as a criterion for evaluating the model performance, since it is indeed related to the closeness of the model predictions with the observed untransformed data (amount adsorbed per unit mass in this study). On the other hand, the r2 and p values give a fair measure of linearity between the dependent and independent variables of the model concerned; either one or both of which are generated through transformation of the original data. Thus, they are actually good tools for evaluating the model performance in describing the variation in the transformed data sets. As far as the data obtained for other temperatures are concerned (data not shown), a similar conclusion about the best fit of the second-order rate equation was established.
The ranges and mean values of r2, p, and SE (Table 5) obtained by applying different models to the 32 sets of kinetic data (obtained in the fast and slow reactions for four different clays at four different temperatures) individually also reflect that among the three ordered models, the second-order rate equation had the highest mean r2 coupled with the lowest mean p and SE values. Furthermore, the ranges of these parameters (Table 5) show that for both the lower and upper limits, the second-order rate equation had the highest r2 and the lowest SE values. The second-order rate equation, however, had the lowest p value only for the upper limits of the ranges.
|
|
|
|
on the surfaces of HyA/HAS-Mts is higher. Both of these two factors apparently resulted in the enhanced rates of SeO3–2 adsorption on the HyA/HAS-Mts. A substantially slower SeO3–2 adsorption due to silication of the HyA polymers is attributable to the reduction in the frequency distribution of the potential adsorption sites through masking of the Al-OH and Al-OH2+ groups at the edge surfaces [Al(H2O)(OH)] and defect sites of HyA polymers and to lowering of their positive charge. The lower PZSE values of the HAS-Mts than the HyA-Mt (Table 1) support this interpretation.
Activation Energy and Pre-exponential Factor
The effect of temperature on SeO3–2 adsorption by the clays was examined in greater detail by applying the Arrhenius equation on the rate constant data. The Arrhenius equation is: k = Ae–Ea/RT. Therefore, ln k = ln A – Ea/RT, where k is the rate constant, A is the pre-exponential factor (frequency factor), which is a measure of the accessibility of the reactive sites to the reactant, Ea is the Arrhenius activation energy, which must be overcome before adsorption can take place, R is the universal gas constant (8.314 J K–1 mol–1), and T is the absolute temperature. A plot of ln k (on the y axis) versus 1/T (on the x axis) yields a straight line, from which the Ea and A can be obtained based on the slope and intercept, respectively.
Figure 5 shows the Arrhenius plots for the fast and slow reactions of SeO3–2 adsorption on the Mt and HyA/HAS-Mts. All the plots yielded a linear relationship (p 0.05) with r2 values ranging from 0.934 to 0.999, except for the slow reaction of SeO3–2 adsorption onto HAS1-Mt. In this case, the r2 and p values were 0.873 and 0.07, respectively. The estimated values of Ea and A are given in Table 8. The activation energy values of SeO3–2 adsorption ranged from 11 to 53 kJ mol–1 of SeO3–2 adsorbed by the clays. Except for a higher degree of silication of HyA-Mt (HAS2-Mt) in the slow reaction, the activation energy for SeO3–2 adsorption onto Mt was always substantially higher than that for the HyA- and HAS-Mts.
|
|
For the slow reactions, the Ea obtained for all three HyA/HAS-Mts were statistically identical (p 0.05) to each other (Table 8). But compared with Mt (53 kJ mol–1), it was much lower in the cases of HyA- (32 kJ mol–1) and HAS1-Mt (27 kJ mol–1). The Ea for the slow reaction in the cases of Mt (53 kJ mol–1) and HAS2-Mt (37 kJ mol–1) were, however, statistically identical (p 0.05) to each other.
The activation energy of a diffusion-controlled process in solution is about 25 kJ mol–1 (Sparks, 1999). However, in heterogeneous systems such as mineral–water interfaces, diffusion occurs not only in the bulk solution but also in micropores and macropores, in the films around solid particles, on the solid surface, and inside solid particles (Sparks, 1989). Therefore, the activation energy for diffusion processes in heterogeneous systems is higher than that in solutions. Film diffusion typically has an activation energy of 17 to 21 kJ mol–1 and intraparticle diffusion has Ea values of 21 to 42 kJ mol–1 (Sparks, 1999). Thus, low Ea values (<42 kJ mol–1) indicate diffusion-controlled processes whereas higher Ea values (>42 kJ mol–1) indicate chemically controlled processes (Sparks, 1989). Therefore, the rate-limiting step for most of the SeO3–2 adsorption reactions onto the Mt and HyA/HAS-Mts was a diffusion process, except for the slow reaction of SeO3–2 adsorption onto the Mt (Table 8). A 53 kJ mol–1 Ea for the slow reaction of SeO3–2 adsorption onto Mt indicates that the rate-limiting step in this case might be a chemically controlled process (Table 8).
The positive effects of the protonated hydroxyl groups of HyA/HAS interlayers and coatings on the accessibility of SeO3–2 anions to the adsorption sites due to minimization of the negative charge of Mt have been discussed. This would result in a lower Ea values, especially of a diffusion-controlled SeO3–2 adsorption on the HyA/HAS-Mts. Moreover, the adsorption of SeO3–2 to the protonated surface hydroxyl groups of the HyA/HAS polymers should be less energy requiring than that with surface hydroxyl (Al-OH) groups on the edges of Mt, since the ligand exchange of SeO3–2 with surface hydroxyl (Al-OH) groups could occur only after they become protonated by the protons dissociated from H2O molecules (Hingston et al., 1972). As the HyA/HAS-Mts had higher frequency distribution of the protonated surface hydroxyl groups on their surfaces, a lower Ea of SeO3–2 adsorption on these clays than on the Mt is, thus, logical.
The pre-exponential factor, an index of the frequency of collision of SeO3–2 with the reactive sites on the clay surface was substantially lower for the HyA/HAS-Mts compared with Mt (Table 8). This was attributable to the fact that, HyA/HAS interlayerings and coatings brought the Mt particles closer together (aggregation) as reflected in the reduction in the total surface area (Table 1).
Agro-environmental Significance
Hydroxy-interlayering in the 2:1 expansible phyllosilicates in acidic soils and sediments is a widely occurring pedogeochemical phenomenon that gives rise predominantly to HyA- and HAS-interlayered Mt and vermiculite (Vt). This process substantially alters the charge and surface properties, and interfacial reactions of the parent Mt and Vt with plant nutrients and pollutants in soil and associated environments. The present study presents the first report of original investigation describing the alteration of SeO3–2 adsorption kinetics on Mt as a consequence of HyA- and HAS-interlayering/coating. The results accomplished in the present study reveal that HyA- and HAS-interlayering/coating brings about a clear increase in the SeO3–2 adsorption affinity, kinetics, and capacity of Mt. Silication of the HyA adsorbed on Mt, which is common in soil environments, results in a remarkable reduction in the SeO3–2 adsorption affinity, kinetics, and capacity. Thus, the findings of the present study are of vital agro-environmental significance in understanding Se transformation and transport in the soil and related environments.
Received for publication April 8, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
-FeOOH. Science (Washington, DC) 238:783–786.-Al2O3 surface using the pressure-jump thechnique. J. Phys. Chem. 87:1454–1462.-FeOOH). Nature (London) 264:740–742.
