| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Bonding
a Department of Geology and Geophysics, Texas A&M University, College Station, TX 77843
b School of the Environment, Clemson University, Clemson, SC 29634-0919
c Present address: Connecticut Agricultural Experiment Station, 123 Huntington Street, New Haven, CT 06504
* Corresponding author (zhud{at}purdue.edu).
Received for publication June 17, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
donors and metal cations in aqueous solutions. The objective of this study was to characterize potential cation– interactions between donors and exchangeable cations accumulated at mineral surfaces via both spectroscopic and batch sorption methods. Quadrupolar splitting in deuterium nuclear magnetic resonance (2H NMR) spectroscopy for d2–dichloromethane, d6–benzene, and d8–toluene (C6D5– moiety) in aqueous suspensions of a Na-saturated reference montmorillonite unambiguously indicated the ordering of solute molecules with respect to the clay surface. The half line broadening (
1/2) of 2H NMR of d6–benzene in montmorillonite suspensions showed that soft exchangeable cations generally resulted in more benzene sorption compared with harder cations (e.g., Ag+ > Cs+ > Na+ > Mg2+, Ba2+). In batch sorption experiments, saturating minerals (e.g., porous silica gels, kaolinite, vermiculite, montmorillonite) with a soft transition metal or softer base cations generally increased the polycyclic aromatic hydrocarbon (PAH) sorption relative to harder cations (e.g., Ag+ >> Cs+ > K+ > Na+; Ba2+ > Mg2+). Sorption of phenanthrene to Ag+–saturated montmorillonite was much stronger compared with 1,2,4,5-tetrachlorobenzene, a coplanar non- donor having slightly higher hydrophobicity. In addition, a strong positive correlation was found between the cation-dependent sorption and surface charge density of the minerals (e.g., vermiculite, montmorillonite >> silica gels, kaolinite). These results, coupled with the observations in 2H NMR experiments with montmorillonite, strongly suggest that cation– bonding forms between PAHs and exchangeable cations at mineral surfaces and affects PAH sorption to hydrated mineral surfaces.
Abbreviations: 2H NMR, deuterium nuclear magnetic resonance • NOC, nonionic organic chemical • PAH, polycyclic aromatic hydrocarbon • , quadrupolar splitting • 1/2, half line broadening
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Potential mechanisms controlling NOC sorption to hydrated mineral surfaces have been advanced based primarily on batch sorption experiments because of the difficulty in direct spectroscopic observations. Huang et al. (1996) proposed that differences in sorption behaviors of phenanthrene exhibited by external or internal porous surfaces of nonswelling minerals (e.g., 
-Al2O3, quartz, kaolinite) and the internal interlayer surfaces of a swelling bentonite clay were caused by their different relative accessibilities to the sorbate, which was mainly affected by near-surface or pore geometry and by the preferential sorption of water. In recent studies by Hundal et al. (2001), the lack of a correlation between Freundlich sorption constants and indices of charge or hydrophobicity of the clay led them to suggest that sorption of phenanthrene by smectites was primarily a physical process via capillary condensation in interlayer nano- or micropores.
Theoretically, several models have been outlined to describe the driving forces for sorption of NOCs by mineral surfaces. Nonionic organic chemical sorption to hydrated mineral surfaces has been ascribed to the "hydrophobic effect," where sorption is driven by a substantial thermodynamic gradient due to a combination of relatively small van der Waals forces and large entropy differences (Tanford, 1980; Israelachvili, 1985; Goss and Eisenreich, 1995). Chiou et al. (1985) have suggested that the sorption of NOCs, such as PAHs, to minerals is a competitive adsorption process by which water molecules and NOCs compete for positions at the mineral surface. It has also been suggested that the nature of adsorbed water at the mineral–water interface, the so-called vicinal water (Drost-Hansen, 1969; Schlautman and Morgan, 1994; Schwarzenbach et al., 2003), serves as a more favorable environment for NOCs to partition from the bulk aqueous phase. None of the models above, however, have been verified by spectroscopic and/or microscopic characterization.
The interaction between small, deuterated molecules and natural organic matter (e.g., humic substance, soil organic matter) or mineral surfaces can be reflected by changes in the deuterium nuclear magnetic resonance (2H NMR) relaxation times (Nanny and Maza, 2001; Zhu et al., 2003a, 2003b) or quadrupolar splitting (Grandjean and Laszlo, 1989; Delville et al., 1991; Weiss and Gerasimowicz, 1995) of the 2H nuclei. Deuterium NMR has the advantage of being very sensitive to solute interactions because the relaxation of deuterium is dominated by the quadrupole relaxation mechanism (Smith, 1983). This creates a relatively simple correlation between the relaxation rate and molecular correlation time (
c), providing a method to directly characterize the complexation or sorption of PAHs in geologically relevant systems.
Recent evidence has suggested that cation– bonds can form between the electrons of aromatic organics and cations, especially in systems with restricted geometries, such as ion channels in cell membranes, the interiors of synthetic receptors, and natural biomolecules in aqueous systems (Ma and Dougherty, 1997). The magnitude of enthalpies of these interactions (approximately 20 kcal/mol) can be quite substantial, approximately one-fifth that of a covalent bond and about five times stronger than a hydrogen bond (Dougherty and Stauffer, 1990; Kumpf and Dougherty, 1993; Gokel et al., 2001), which is competitive with the expected strongest noncovalent binding forces. Because minerals accumulate cations at the mineral–water interface, the formation of cation– interactions between PAHs and the exchangeable cations may affect PAH sorption to mineral surfaces.
Spectroscopic studies in interactions of NOCs with hydrated mineral surfaces are very limited. The objectives of this study were to: (i) quantify the sorption of PAHs to different minerals saturated with different cations, (ii) apply 2H NMR techniques to characterize molecular interactions of deuterated solutes with mineral surfaces, and (iii) identify the sorption mechanism(s) controlling PAH sorption to mineral surfaces.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
donors, and 1,2,4,5-tetrachlorobenzene (Aldrich Chemical) as a non- donor were used for batch sorption experiments. Aqueous solubility and n octanol–water partitioning coefficient (Kow) of these compounds are listed in Table 1. ACS-grade NaCl (EM Science, Gibbstown, NJ), MgCl2·6H2O (EM Science), KCl (EM Science), CsCl (Fisher Scientific, Hampton, NH), BaCl2 (Fisher Scientific), NaNO3 (Fisher Scientific), and AgNO3 (Fisher Scientific) were used to prepare solutions for NMR experiments and/or saturate minerals for batch sorption experiments. The minerals used in this study included two kinds of silica gels (Davisil, Grade 633, 500 m2/g, pore size = 60 Å and Grade 643, 300 m2/g, pore size = 150 Å; Fisher Scientific), kaolinite (Kga-lb; Source Clay Minerals Repository, University of Missouri, Columbia), vermiculite (matted, Transvaal, Africa; Ward's Natural Science Establishment, Rochester, NY), and a reference montmorillonite (STx-1; Source Clay Minerals Repository, University of Missouri, Columbia). The silica gels were saturated by cations directly without any pretreatment. Kaolinite, vermiculite, and montmorillonite were pretreated to remove amorphous iron oxides using dithionite–citrate–bicarbonate (DCB) and organic matter using hydrogen peroxide (Kunze and Dixon, 1986). With the exception of Ag+, the minerals were saturated with different metal cations after being washed four times with a 1.0 M metal chloride solution and four times with a background solution of 0.1 M ionic strength. After the final rinse, the minerals were oven-dried at 105°C and placed in a desiccator until used. For Ag+ saturation experiments, the minerals (except silica gels) were first washed two times with 0.5 M NaNO3 to remove residual anions (e.g., Cl–) that may have reacted with Ag+ to form precipitates and then two times with 1.0 M AgNO3 solution, during which time the pH of the suspension was adjusted to approximately 6.5 using 0.5 M HNO3 to avoid precipitation of Ag+. The obtained minerals were then rinsed two times by a solution mixed by 0.01 M AgNO3 and 0.09 M NaNO3. After the final rinse the minerals were oven-dried at 70°C, and then placed in a desiccator until used. During the process, samples were protected from light to avoid possible photolysis of AgNO3. Specific surface areas of the parent minerals kaolinite (9.5 m2/g) and vermiculite (5.2 m2/g) were determined by single point N2 Brunauer–Emmett–Teller (BET) analysis using a Quantasorb Jr. surface area analyzer (Quantachrome Corp., Greenvale, NY). For the swelling montmorillonite, measured specific surface areas were 90, 80, and 47 m2/g for the Na+–, Cs+–, and Ag+–saturated mineral, respectively, as determined by multipoint BET nitrogen isotherms using a Micromeritics ASAP 2010 surface area analyzer (Micromeritics Instrument Corp., Norcross, GA).
|
) were internally referenced to the natural abundance of deuterated water for clay suspension samples ( < 0.02 ppm, checked with an external reference of deuterated chloroform). No detectable chemical shift (i.e., >0.05 ppm) was observed for any deuterated probe.
Batch Sorption Experiments
Batch equilibrium sorption experiments of naphthalene, phenanthrene, pyrene, and 1,2,4,5-tetrachlorobenzene were performed on a constant solution chemistry basis (0.1 M ionic strength; pH 7.0; 25°C). Twenty- or forty-milliliter glass centrifuge tubes (Kimax; Kimble Glass, Cleveland, OH) with Teflon-lined screw caps were used as batch reactors. The pH values of suspensions of sorbents in background solutions were adjusted to 7.0 with 1.0 M NaOH and 1.0 M HCl, or 1.0 M HNO3 for suspensions containing Ag+. Solutes were added with methanol carrier, which was kept at levels (<0.1% v/v) sufficiently low to eliminate cosolvent effects. Blanks (sorbent plus background solution, no PAH) were prepared for background corrections, and control samples (PAH plus background solution, no sorbent) were prepared to quantify sorbate loss due to other processes. Based on previous studies, centrifuge tubes were agitated in the dark on an orbital shaker at 150 rpm for at least 72 h for montmorillonite (Hundal et al., 2001) and 24 h for the other minerals (Grimaldi, 1999; Zhu, 2001) to reach sorption equilibrium.
Following centrifugation at 5000 rpm for 45 min, 10 mL of supernatant from all batch systems except those containing Ag+ or montmorillonite was mixed with high performance liquid chromatography (HPLC)-grade methanol in a 1:1 volume ratio for PAH fluorescence intensity measurements. For samples containing Ag+, 10 mL of supernatant was first mixed with 10 mL of a solution containing 3.0 M NaCl and methanol (in a 1:1 volume ratio) and the resulting AgCl precipitate was then removed from solution by centrifugation at 2800 rpm for 45 min before making the PAH fluorescence intensity measurement on the final supernatant. The efforts taken to remove Ag+ from samples before fluorescence measurements were done because previous studies have shown that Ag+ is a very effective quencher of PAH fluorescence even at very low concentrations (Lakowicz, 1999; Lee et al., 2004). The added methanol prevented further PAH sorption to glassware, fluorescence cells, and any AgCl precipitates. Concentrations of naphthalene and pyrene were measured using fluorescence spectroscopy (Photon Technology International Inc., South Brunswick, NJ). The excitation and emission wavelengths were 277 and 336 nm for naphthalene, and 336 and 374 nm for pyrene, respectively. Standards used to determine fluorescence calibration curves were prepared with the same background solutions used in the batch sorption experiments, and were treated by the same method as mentioned above for Ag+–containing samples. For sorption experiments with montmorillonite, after centrifugation the solute was extracted from an aliquot with hexane and analyzed by gas chromatography (GC)–flame ionization detection (FID) for phenanthrene and GC–electron capture detection (ECD) for 1,2,4,5-tetrachlorobenzene using a DB-624 capillary column (J&W Scientific, Folsom, CA).
Five-point (fluorescence) or seven-point (GC analyses) calibration curves spanning the range of expected concentrations were used to determine sorbate concentrations in the supernatants. Sorbate loss due to other processes (e.g., sorption to glassware) was accounted for by determining a loss coefficient from the control samples for each background solution used. Adsorbed mass at each equilibrium concentration was calculated as the difference between total and solution-phase mass. Apparent distribution coefficients (Kd) were calculated on a unit-surface-area basis (L/m2) for the non- and low-swelling clays (e.g., porous silica gels, kaolinite, vermiculite).
Batch Desorption Experiments
Upon completion of the sorption period, desorption for selected concentrations was conducted for phenanthrene/Ag+–montmorillonite using a single-step, centrifuge-withdraw-refill method (Huang et al., 1998). About 85% of the supernatant was replaced with fresh background solution. At least 72 h were used to reach desorption equilibrium.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
) can be expressed as (Halle and Wennerstrøm, 1981):
 | [1] |
LD is the angle between the magnetic field (B0) and the local order director vector of the oriented platelets. Whenever fast exchange occurs with isotropic orientation of either solute molecules or of the clay platelets, equals zero (Grandjean and Laszlo, 1989; Delville et al., 1991; Weiss and Gerasimowicz, 1995). Quadrupolar splitting was observed for d2–dichloromethane, d6–benzene, d8–toluene (C6D5– moiety), and d2–water (at natural abundance) in suspensions of Na-saturated montmorillonite when pH was >7 in the present study (Fig. 1).
|
Candidates for these forces include cation-induced charge–dipole interactions for dichloromethane or cation– interactions for benzene and toluene and surface-water-induced H-bonding (described in Fig. 2). For dichloromethane, charge–dipole interactions are expected to be more important because of the relative weakness of H-bonding ability of dichloromethane. Likewise, although benzene has been reported to form hydrogen bonds with water, the binding energies are relatively low (<2 kcal/mol) (Suzuki et al., 1992) compared with that of cation– interactions (approximately 20 kcal/mol) (Kumpf and Dougherty, 1993). Thus, it can be concluded that the ordering of benzene at mineral surfaces may result from cation– interactions with exchangeable cations.
|
interaction over a simple trapping mechanism. First, if the ordering were only caused by a physical trapping, d8–toluene would have had a larger than d6–benzene because the steric effects due to the CD3– substituent further restrain free molecular rotation. Note this trend would be further strengthened when considering the greater sorption of toluene resulting from its higher hydrophobicity compared with benzene, as justified by the aqueous solubility and Kow (Schwarzenbach et al., 2003). However, a reversed trend was in fact observed (i.e., although quadrupolar splitting was observed for d8–toluene ( deuterium of C6H5– moiety, shown in Fig. 1f, the value is so small that it cannot be calculated by the instrument-associated software). Because the difference in quadrupolar coupling constants (e2Qq/h) between these two compounds is negligible (i.e., identical if no measurement errors are considered [Mantsch et al., 1977]), the lower of d8–toluene is probably caused by less ordering of the solute with respect to the clay surface. This hypothesis is consistent with a cation– directed sorption mechanism.
Because of the steric effects of the CD3– group, the C6D5– moiety has to tip at a certain angle for effective cation– interactions to occur with exchangeable cations at clay surfaces. This conformation adjustment decreases the quadrupole interaction of the 2H nuclei with the external B0 (i.e., higher LD in Eq. [1]), and thus lowers the observed value. Unlike C6H5–, no quadrupolar splitting was observed for the CD3– moiety of d8–toluene (Fig. 1f), providing further support for the cation– sorption mechanism. It is the C6H5– rather than the CD3– moiety that is directly involved in the cation– interaction with exchangeable cations at the clay surface, while the free rotation of the C–C 
bond averages A to zero and thereby results in zero-valued for CD3– in Eq. [1]. Conversely, if trapping alone were responsible for the quadrupolar splitting of C6H5–, this phenomenon should also be observed for CD3– because they are both restrained from free molecular rotation by the same restricted local geometry.
In addition, the observed quadrupolar splitting for dichloromethane makes the trapping mechanism less tenable as a cause for the ordering of the different solutes. This is because unlike benzene and toluene, which have planar structures and tend to be aligned parallel with clay surfaces when trapped in restricted geometries, dichloromethane possesses a tetrahedral structure and has difficulty being reoriented without interacting with certain directed forces. However, it should be made clear that although directed specific interactions such as cation– bonding are required for the ordering of the solutes, this observation does not necessarily exclude the existence of a trapping mechanism. For example, solutes can be trapped in restricted geometries but simultaneously interact with interlayer surfaces via directed molecular forces to order their structures.
Under the same experimental conditions (1:200 clay to solution mass ratio, pH 9.0, 0.01 M ionic strength), however, no quadrupolar splitting was observed for montmorillonite saturated by the other cations (e.g., Cs+, Ag+, Mg2+, Ba2+). This may be because clay minerals saturated by different cations have different abilities to form single-layer platelets. It is worth noting that the clays saturated by cations other than Na+, especially the bivalent cations (e.g., Mg2+, Ba2+), were found to settle relatively quickly in their aqueous suspensions. It is well known that exchangeable cations significantly affect the stability of clay suspensions and the swelling of clays (McBride, 1989). Flexible platelets of montmorillonite having the thickness of a single layer can associate in stacked, roughly parallel alignment to form so-called quasicrystals (Quirk and Aylmore, 1971). Bivalent cations such as Ca2+, which are solvated by water molecules in an outer-sphere surface complex, cross-link clay layers together through electrostatic forces to form montmorillonite quasicrystals composed of stacks of four to seven layers (Sposito and Prost, 1982). Evidence for quasicrystal particles in suspensions of hectorite saturated by a high-radius monovalent base cation, Cs+, is also available (Weiss, 1990). For small monovalent base cations such as Li+ and Na+, however, macroscopic swelling becomes greater as the difference in osmotic pressure between the interlayer and external layer increases because of strong hydration of these cations (McBride, 1989). Compared with single-layer platelets, quasicrystals are much larger in particle size and hence are easier to settle and less effective to be ordered when interacting with B0 in suspension. It is believed that under the experimental conditions used in the present study, smectites saturated by Na+ are more single-layered but are quasicrystals when saturated by the other cations.
Tables 2 and 3 show measured values of d6–benzene and d2–water as a function of clay to solution mass ratio and pH, respectively. Under fast chemical exchange of solute molecules, the quadrupolar splitting values measured are averages of the different states (e.g., free and adsorbed). Thus, a higher value is expected when more solute is adsorbed with increasing loading level of the clay. As pH decreases, the negative charge at the clay surface is neutralized by H+ introduced, which in turn causes coagulation of clay platelets to form tactoids having larger particle sizes. The reduced charge density and the increased particle size of clay platelets reduce their interactions as dipoles with B0. This explains why the quadrupolar splitting was not observed under neutral and acidic conditions. Note the magnitude of is also strongly affected by the intensity of B0. Clay platelets become more oriented in suspension with increasing B0 because of the increased interaction, resulting in a higher observed value. We indeed observed the quadrupolar splitting of d6–benzene using a Varian-400 NMR spectrometer (61.3 MHz for 2H nuclei) in suspension of Na–montmorillonite at pH 5 (data not shown).
|
|
1/2) under the extreme narrow condition is described as (Zens et al., 1976):
 | [2] |
is the resonance frequency (30.7 MHz), e2Qq/h is the quadrupolar coupling constant in Hz, f and s are the molecular correlation times for free and sorbed solutes, respectively, and 
is the mole fraction for sorbed solutes. An increased 1/2 value in general indicates a higher s (slower molecular rotation) and a higher . Thus, solute sorption behaviors can be related to observed 1/2 values.
The values of 1/2 of d6–benzene in suspensions of montmorillonite saturated by different cations are shown in Table 4. It is important to note that any paramagnetic induced relaxation (i.e., structured iron not removed by the DCB treatment) causes only a baseline effect because of the same loading level of clay. The relative sorption intensity as indexed by 1/2 (e.g., Ag+ > Cs+ > Na+ > Mg2+, Ba2+) indicates that a soft transition metal (e.g., Ag+) or a softer base cation (e.g., Cs+) generally favors benzene sorption, which is consistent with the hypothetical cation– sorption mechanism.
|
interactions in aqueous solution based on the measurements of spin–lattice relaxation times (T1) of d6–benzene. It has also been reported in the literature that calix arenes can selectively remove certain cations such as Ag+ and Cs+ from aqueous solution through the formation of cation– complexes (Ikeda and Shinkais, 1994, 1997; Bocchi et al., 1995). For soft cations, participation of d orbitals and contributions from polarization and dispersion of the cation may be important to the overall cation– interaction in addition to the normal electrostatic interaction (Ma and Dougherty, 1997; Caldwell and Kollman, 1995). Also, the relatively weak hydration of soft base cations due to their high radii further makes the cation– interaction more competitive in the aqueous phase. Conversely, bivalent hard base cations (e.g., Mg2+, Ca2+) would be expected to induce very weak cation– interactions because of their very high hydration energies.
Nonetheless, mineral surfaces accumulate and concentrate cations, especially those soft transition metals and soft base cations, which provide interactive sites for cation– interactions with -electron-rich compounds including PAHs. In fact, previous studies have revealed some cation-induced specific interactions occurring at mineral surfaces. Complexation of benzene and alkylbenzene with transition metals (e.g., Cu2+, Fe3+, Ru3+) on dry montmorillonite was reported decades ago (Doner and Mortland, 1969; Pinnavia et al., 1974). Structural alteration of chlorophenols and chloroanisole (e.g., polymerization and oxidation) catalyzed by Cu2+ in dehydrated smectite has also been reported (Boyd and Mortland, 1986; Govindaraj et al., 1987; Johnston et al., 1991). More recently, site-specific interactions of 4,6-dinitro-o-cresol (DNOC) and 4,6-dinitro-2-sec-butylphenol with exchangeable cations have been implicated by cation-induced wavelength shifts of the NO2– groups in Fourier transform infrared (FTIR) spectra and molecular simulation (Johnston et al., 2002; Li et al., 2003). The investigators also showed that weakly hydrated cations (e.g., K+, Cs+) resulted in greater sorption compared with more strongly hydrated cations (e.g., Na+, Ca+), and the DNOC molecules were oriented parallel to the clay surface. The relatively strong cation– bonding capability of Ag+ and Cu+ has been directly utilized to prepare cation-charged minerals as adsorbents for effective olefin separation (Takahashi et al., 2001; Padin and Yang, 2002).
Batch Sorption and Desorption Studies
Non- and Low-Swelling Minerals
Linear equilibrium sorption isotherms were observed for all batch sorption experiments with the non- and low-swelling minerals (e.g., porous silica gels, kaolinite, vermiculite). Unfortunately, because these minerals suspended very poorly in aqueous solution, no consistent and reliable results could be collected for 2H NMR experiments with d6–benzene. The apparent Kd (L/m2) values for sorption of naphthalene/silica gels, pyrene/silica gels, pyrene/kaolinite, and pyrene/vermiculite are shown in Fig. 3. In general, PAH sorption to minerals saturated by soft cations (e.g., Cs+, Ba2+), especially by Ag+, was greater than that to minerals saturated by hard cations. These results correspond very well with the overall binding energy sequence Ag+ > Cs+ > K+ > Na+, Li+ for cation– interactions obtained in our previous studies (Zhu et al., 2004), suggesting that cation– interactions are important in sorption of PAHs to hydrated minerals.
|
Swelling Mineral
A reference montmorillonite (STx-1) was the swelling clay tested in this study. Sorption isotherms for phenanthrene and/or 1,2,4,5-tetrachlorobenzene with Na+–, Cs+–, and Ag+–montmorillonites are shown in Fig. 4, along with desorption points for selected concentrations of phenanthrene with Ag+–montmorillonite. Values of the Freundlich isotherm parameters (KF and n) are listed in Table 5. Like vermiculite, strong effects of exchangeable cations on PAH sorption (i.e., Ag+ >> Cs+ > Na+) were also observed for the high-CEC (approximately 800–1500 mmol/kg [Sparks, 1995]) montmorillonite. It is interesting to note that sorption of 1,2,4,5-tetrachlorobenzene, a coplanar non- donor, to Ag+–montmorillonite is much lower compared with phenanthrene, a donor, despite the fact that 1,2,4,5-tetrachlorobenzene is slightly more hydrophobic, as shown by its lower aqueous solubility and higher Kow (Table 1). These results clearly demonstrate that specific interactions occur between phenanthrene and Ag+ at the montmorillonite surface.
|
|
Conversely, a strong correlation relationship was observed in the present study between Freundlich sorption constants (KF and n) and indices of cation– bonding ability of the monovalent exchangeable cations (Ag+ >> Cs+ > Na+) for phenanthrene sorption to montmorillonite, which is consistent with cation– bonding sorption mechanism. We suggest strong cation– bonding occur between PAHs and soft exchangeable cations concentrated at montmorillonite interlayers, which may be progressively opened up to create more sorption sites (conditioning of mineral surfaces) with increasing PAH loading. This is why stronger sorption and a more significant upward-curving isotherm were observed for soft exchangeable cations relative to hard cations (e.g., Ag+ >> Cs+ > Na+) and donors relative to non- donors (e.g., phenanthrene >> 1,2,4,5-tetrachlorobenzene). Note the significant discrepancy in sorption behavior between phenanthrene and 1,2,4,5-tetrachlorobenzene clearly cannot be explained by a physical sorption mechanism alone. It is also interesting to note that the desorption points for phenanthrene/Ag+–montmorillonite were displaced exactly on the original sorption isotherm, indicating that the conditioning process of mineral surfaces driven by cation– bonding is fully reversible. It should be pointed out that such mineral conditioning characterized with an upward-curving isotherm was not observed for nonswelling vermiculite despite its high cation exchange capacity.
For smectites saturated by bivalent hard base cations such as Ca2+, which have extra strong ability to cross-link clay interlayers and form quasicrystals (Sposito and Prost, 1982), PAH molecules are mainly physically entrapped in the nano- and micropores, which can also be conditioned (i.e., enlarging existing pores with increasing solute loading level) driven by nonspecific hydrophobic effects.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Hydrophobic effects, whereby sorption is controlled by a combination of relatively small van der Waals forces and a substantial entropic difference (Tanford, 1980; Israelachvili, 1985), have been used to explain observed correlations between the logarithms of Kd values and PAH aqueous activity coefficients (
w) (Schwarzenbach and Westall, 1981; Mader et al., 1997). This explanation would also be consistent with the relative sorption of naphthalene and pyrene observed in our batch experiments. However, an entropic-driven sorption mechanism implies that chemical bonds formed between mineral surfaces and PAHs are weak (or nonexistent) and do not contribute significantly to the enthalpy change (H) and the Gibbs free energy change (G) of the reaction. If this mechanism were true, PAH sorption would not be affected by the type of exchangeable cation or the mineral surface chemistry for a given PAH, which contradicts the observations from our batch sorption experiments. Therefore, an entropic sorption mechanism cannot be the only process that governs PAH sorption to hydrated mineral surfaces.
Chiou et al. (1985) have suggested that the sorption of NOCs to minerals is a competitive adsorption process by which water molecules and NOCs compete for sites at the mineral surface. Conceptualizing the sorption mechanism in this way, the authors concluded that NOC sorption to mineral surfaces in aqueous solutions is energetically unfavorable and therefore significant sorption would not be expected. This conclusion was based on a comparison of both the relative interaction energies of NOCs and water with mineral surfaces and their relative concentrations in aqueous solutions.
However, in the present study we have shown that PAHs are a special type of NOC in that they form relatively strong interactions with cations through cation– bonding. Interactions between PAHs and exchangeable cations, especially those soft cations that favor cation– interactions, at mineral surfaces are strong enough to drive PAH sorption to mineral surfaces through enthalpic effects on the overall free energy change (G) of the reaction. Therefore, the best conceptual mechanistic model of PAH sorption to mineral surfaces is one that combines the enthalpic effects due to the formation of specific interactions between exchangeable cations on the mineral surface and the PAH molecules with the influence of entropy on PAH activity coefficients. This model then accounts for the wide variation of PAH sorption as a function of PAH properties, solution chemistry, and mineral surface chemistry. Note circumstantial changes in mineral properties, especially the formation of quasicrystals, should also be considered for swelling clays.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
interactions–nonadditive effects are critical in their accurate representation. J. Am. Chem. Soc. 117:4177–4178. interactions. Coord. Chem. Rev. 222:127–154.-donor participation in complexation of cations and evidence for metal-tunneling through the calix[4]arene cavity. J. Am. Chem. Soc. 116:3102–3110. interactions. Science (Washington, DC) 261:1708–1710. interaction. Chem. Rev. 97:1303–1324.[ISI][Medline] interactions in aqueous solution using deuterium NMR spectroscopy. J. Environ. Qual. 33:276–284.Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  Y. Zhang, D. Zhu, and H. Yu Sorption of Aromatic Compounds to Clay Mineral and Model Humic Substance-Clay Complex: Effects of Solute Structure and Exchangeable Cation J. Environ. Qual., May 1, 2008; 37(3): 817 - 823. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Li, B. J. Teppen, D. A. Laird, C. T. Johnston, and S. A. Boyd Effects of Increasing Potassium Chloride and Calcium Chloride Ionic Strength on Pesticide Sorption by Potassium- and Calcium-Smectite Soil Sci. Soc. Am. J., September 20, 2006; 70(6): 1889 - 1895. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Roberts, H. Li, B. J. Teppen, and S. A. Boyd SORPTION OF NITROAROMATICS BY AMMONIUM- AND ORGANIC AMMONIUM-EXCHANGED SMECTITE: SHIFTS FROM ADSORPTION/COMPLEXATION TO A PARTITION-DOMINATED PROCESS Clays and Clay Minerals, August 1, 2006; 54(4): 426 - 434. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. A. Schoonen, C. A. Cohn, E. Roemer, R. Laffers, S. R. Simon, and T. O'Riordan Mineral-Induced Formation of Reactive Oxygen Species Reviews in Mineralogy and Geochemistry, January 1, 2006; 64(1): 179 - 221. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||
