Exponential Sensitivity and Speciation of Al(III), Sc(III), Y(III), La(III

ARTICLE pubs.acs.org/JPCA

Exponential Sensitivity and Speciation of Al(III), Sc(III), Y(III), La(III), and Gd(III) at Fused Silica/Water Interfaces David S. Jordan, Sarah A. Saslow, and Franz M. Geiger Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States

bS Supporting Information ABSTRACT: The binding contants, adsorption free energies, absolute adsorbate number densities, and interfacial charge densities of Al(III), Sc(III), Y(III), La(III), and Gd(III) interacting with fused silica/water interfaces held at pH 4 were determined using second harmonic generation and the Eisenthal χ(3) technique. By examining the relationship between the measured adsorption free energies and the electric double layer interfacial potential at multiple electrolyte concentrations, we elucidate the charge state and possible binding pathways for each ion at the fused silica surface. Al(III) and Sc(III) ions are found to bind to the fused silica surface as fully hydrated trivalent species in a bidentate geometry. In contrast, the Y(III), La(III), and Gd(III) ions are each shown to adsorb to the silica surface in a decreased charge state, but the extent and mode of binding varies with each ion. By quantifying the exponential sensitivity of the surface coverage of the adsorbed ions to their charge state directly at the fused silica/water interface, we provide benchmarks for theory calculations describing the interactions of metal ions with oxide interfaces in geochemistry and hope to improve the prediction of trivalent metal ion transport through groundwater environments.

I. INTRODUCTION Adsorption processes at mineral/water interfaces are important for determining the mobility and transport of aqueous pollutants in groundwater systems.13 Understanding these interactions is therefore key to successful groundwater pollutant remediation and risk assessment. In this work, we use second harmonic generation (SHG) to assess the binding of several trivalent metal ions to the fused silica/water interface. Specifically, we analyze the relationship between the adsorption free energy of the trivalent metal interactions with fused silica/water interfaces and the initial interfacial potential of those interfaces to gain insight into the speciation and chemical nature of the surface active species. This information is crucial to understanding and predicting how the trivalent metal pollutants are transported throughout the environment. Metal ion speciation is a key factor when assessing the free energy gained upon adsorption.35 According to Coulomb’s law, a trivalent cation should interact more strongly with a charged interface than a divalent cation at the same distance and surface potential. Bulk solution thermodynamics often predict that trivalent ions exist as fully hydrated species in solution unless electrolyte concentrations approach those of brines,6 and they are thus treated as trivalent ions at interfaces as well. However, charge repulsion between adsorbed ions at interfaces can change the charge state of the ions through complex ion formation, a situation that is not predicted in bulk speciation calculations. This is an important distinction, as changes in the charge state of an adsorbing ion would have a large impact on pollutant mobility,3,7 as we will show here by quantifying the free energy r 2011 American Chemical Society

relationships in the electrical double layer at the fused silica/ water interface through an experimental model study that allows us to determine the speciation of adsorbed metal cations. Specifically, we determine the binding constants, adsorption free energies, absolute surface coverage, and speciation of Al(III), Sc(III), Y(III), La(III), and Gd(III), adsorbing to the fused silica/water interface as a function of background electrolyte concentration at constant pH. These elements were chosen due to their environmental significance, as well as their lack of redox chemistry and clean bulk solution speciation under the experimental conditions employed in this work. Yttrium and the rare earth elements (YREEs) have emerged as important materials in research and industry.8 Examples include the use of YREEs as emitters in electroluminescent devices,9,10 as experimental reagents in organic synthesis and catalysis,11,12 as MRI contrast agents,13 and in the magnets and batteries that power technological devices such as cell phones, laptop computers, hybrid cars, and wind turbines.14,15 YREEs are also regarded as natural analogues to the more volatile actinide series16,17 and are used as tracers of geochemical processes.1820 High demand for these materials has led to greater environmental exposure,21 resulting in groundwater contamination22,23 and posing a threat to plants and animals.24,25 Their toxicity has been previously reviewed.26 Additionally, aluminum, while a common constituent in the Earth’s crust,27 can be harmful to both plants and Received: September 13, 2011 Revised: November 1, 2011 Published: November 03, 2011 14438

dx.doi.org/10.1021/jp208843v | J. Phys. Chem. A 2011, 115, 14438–14445

The Journal of Physical Chemistry A

ARTICLE

animals if soil concentrations are elevated.28 Excess aluminum can enter groundwater through various natural and anthropogenic activities,27,2932 posing a threat to both plants, inhibiting root growth and function,1,33 and humans, acting as a neurotoxin.28,3436 Due to the potential threat posed to both humans and the environment, adsorption studies of these trivalent ions at the mineral oxide/water interface are crucial to pollutant risk assessment and remediation efforts. Previous work on the adsorption of these trivalent ions to interfaces has focused on using batch techniques to study their interactions with biological membranes3739 and various mineral oxide materials.4043 X-ray spectroscopy has also been used to assess the binding of several lanthanide ions with rutile44 and other mineral oxides.45 Additionally, solid-state NMR has been used to investigate the association of Al(III) with silica, where the Al(III) ion was shown to interact with the material through an inner-sphere mechanism.46 Our prior SHG results on the adsorption of Al(III), La(III), Gd(III), and Lu(III) to the fused silica/water interface showed that the lanthanide ions exhibited adsorption parameters similar to those of divalent cations, whereas Al(III) showed a much higher binding affinity.47 This SHG work complements previous batch studies which showed the lanthanide ions coordinate weakly with mineral surfaces, while Al(III) and Sc(III) are shown to coordinate to a stronger extent. This leads to the hypothesis that the lanthanide ions are binding to the silica interface in a decreased charge state, which would greatly affect the various surface complexation models used to evaluate pollutant mobility.7 However, a quantitative experimental study evaluating the surface speciation of these ions is lacking. To bridge this knowledge gap, we present an analysis of the relationship between the adsorption free energies of Al(III), Sc(III), Y(III), La(III), and Gd(III) and the interfacial potential at pH 4 and various electrolyte concentrations. As outlined below, we use second harmonic generation and the Eisenthal χ(3) technique in order to quantify the thermodynamics of trivalent metal adsorption directly at the fused silica/water interface, under flow conditions, and without the use of labels.

intrinsic properties of the interface which undergo a negligible change throughout the experiments. Thus, eq 1 can be simplified to pffiffiffiffiffiffiffiffi ð2Þ ISHG ¼ ESHG µ P2ω ¼ A þ BΦ0 Here, A and B are constants that include Eω and the nonlinear susceptibilities. The square root of the second harmonic signal intensity, ISHG, is directly proportional to the initial interfacial potential, illustrating that a change in the interfacial potential directly results in a change in the SHG signal intensity. II.B. Triple Layer Model. At pH 4 and under the ionic strengths employed in these studies, the fused silica surface exhibits a net negative charge.3 The addition of positive cations to this interface should decrease the magnitude of the interfacial potential and thus cause a decrease in the second harmonic signal intensity proportional to the concentration of the metal cation. In order to quantify metal ion adsorption from χ(3) adsorption isotherms, the triple layer model is employed as an analyte-dependent expression for the interfacial potential.3 This framework is widely used to model adsorption to mineral oxides due to its applicability over a wide range of ionic strengths and analytes.2,7,52,53 The relevant possibilities and limitations of the approach used here are evaluated and quantified in detail in ref 45 and its supporting information. According to the triple layer model, the interfacial region is taken to consist of the zero plane, which represents the surface of the solid substrate, the β plane, and the diffuse layer plane. The acid/base chemistry of the zero plane is taken to determine the surface charge density, σ0, and the corresponding interfacial potential, Φ0. The intermediate β plane represents the location of specifically adsorbed ions and carries the charge density σβ and the potential Φβ. Finally, the outermost diffuse plane represents a layer of charge-balancing ions and carries the potential Φd. The innermost layer, bound on either side by the zero and β planes, is treated with a constant capacitance approach (C1), as is the layer bound by the β and diffuse planes (C2).53,54 The following is an expression for the interfacial potential (Φ0) based on the triple layer model (a derivation of eq 3 can be found in the Supporting Information):

II. BACKGROUND AND THEORY

Φ0 ¼

II.A. Eisenthal χ Technique. Second harmonic generation (SHG)48,49 is a coherent nonlinear optical process wherein two photons of frequency ω are converted to one photon at twice the frequency, 2ω, in a single quantum mechanical process.50 The Eisenthal χ(3) technique51 makes this method highly sensitive to the electrostatic potential generated by charged interfacial species. As we describe in detail in ref 45, the χ(3) technique allows us to quantify the interaction of metal cations with mineral surfaces directly and without the use of electrochemical or photochemical labels. The SHG E-field (ESHG) is proportional to the induced second order polarization (P2ω) and is expressed as follows: (3)

ESHG µ P2ω ¼ χð2Þ Eω Eω þ χð3Þ Eω Eω Φ0

ð1Þ

Here, Eω is the applied electric field at frequency ω, Φ0 is the initial interfacial potential obtained by integrating the third-order electric field from 0 to ∞ in the direction perpendicular from the surface, and χ(2) and χ(3) are the second- and third-order nonlinear susceptibilities, respectively. The incident electric field is held constant throughout our experiments, and the second- and third-order nonlinear susceptibilities are constants related to

σβ þ σ0 σ0 þ þ Φd C1 C2

ð3Þ

The potential of the diffuse layer (Φd) is assumed to decay according to GouyChapman theory.53,54 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi! 2kB T π 1 sinh Φd ¼ ðσ0 þ σ β Þ ð4Þ ze 2εTkB Celec Here, kB, T, and e have their usual meanings, z is the valency of the screening ions in solution, ε is the dielectric constant for water at 25 °C, and Celec is the bulk concentration of the screening electrolyte in solution. By substituting eqs 3 and 4 into eq 2, we arrive at a triple layer expression for the SHG response as a function of the bulk metal concentration. ESHG ¼ A0 þ B ( rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!) σm θ 2kB T π 1 sinh þ ðσ0 þ σm θÞ C2 ze 2εTkB Celec

ð5Þ In eq 5, the charge density at the β plane, σβ, is modeled as the maximum charge density established by metal ion adsorption, 14439

dx.doi.org/10.1021/jp208843v |J. Phys. Chem. A 2011, 115, 14438–14445

0.0045(6) 1.4(2) 1012 0.004(1) 1.3(3) 1012

All values calculated assuming +3 charge on the Al(III) and Sc(III) ions. b A range of values for Y(III) and La(III) are reported to account for either a +2 or +3 ion charge. c All values calculated assuming a +2 charge on Gd(III). a

21(1) +1.55(9) 0.0039(3) 1.22(9) 1012 0.0036(3) 1.12(9) 1012 c

Gd(III)

Sc(III)

0.004(3) 1.3(9) 1012

0.009(4) 1.9(8) 10123(1) 1012 29.9(9) 0.006(2) 1.3(4) 10121.9(6) 1012 32.5(6) 0.0042(8) 0.8(2) 10121.3(3) 1012 36.9(7) 0.0035(4) 0.73(8) 10121.1(1) 1012 42.5(6) 0.0041(4) 0.85(8) 10121.3(1) 1012 b 46.4(9)

La(III)

25.4(6) +1.15(5) 0.0036(3) 0.75(6) 10121.1(1) 1012 31.1(7) 0.0045(2) 0.85(4) 10121.40(6) 1012 35(1) 0.004(1) 0.8(2) 10121.3(3) 1012 37.2(3) 0.0040(2) 0.83(5) 10121.25(7) 1012 41.1(6) 0.0042(3) 0.87(6) 10121.3(1) 1012 b 44(1)

25.6(3) +1.15(3) 2(1) 1012 31.5(2) 1.3(4) 1012 34.5(4) 1.0(2) 1012 38(1) 8(2) 1011 41.1(3) 7(2) 1011 a 44.2(6) Y(III)

32.8(7) 8(4) 1011 37(1) 0.007(4) 4(2) 1011 39.1(9) 0.006(2)

37.2(6) 0.004(2) 38.5(6) 0.002(1) Al(III)

+0.79(7) 1.1(8) 1012 40.2(8) 0.005(1) 8(1) 1011 43.3(4) 0.0040(7) 7(1) 1011a 46.0(8) 0.0033(7)

+0.9(1) 40.9(3) 0.0055(4) 44.8(6) 0.0037(5)

31.6(9)

ARTICLE

47.1(6) 0.0032(5)

ΔGchem Δz 10 mM 54.0 mV 3 mM 81.4 mV 1 mM 108.7 mV 0.3 mM 138.8 mV 0.1 mM 168.9 mV metal

Table 1. Summary of the Measured Adsorption Free Energies (kJ/mol), Charge Densities (C/m2), and Adsorbate Number Densities (Ions/cm2) for Each Metal Ion Binding to the Fused Silica/Water Interface at pH 4 and 298 K as a Function of Screening NaCl Concentration (Column Headings) and Corresponding Interfacial Potential


σm, scaled by surface coverage55 where θ is given by the Langmuir model.56 θ¼

Kads ½Cmetal 1 þ Kads ½Cmetal

ð6Þ

II.C. Free EnergyPotential Relation. A metal ion adsorbing to a charged fused silica substrate from an electrolyte solution must first pass through the electrical double layer. The adsorption free energy for this situation can be described as follows:2,3,7,5759

ΔGads ¼ ΔGelectrostatic þ ΔGchemical

ð7Þ

Here, the observed adsorption free energy is taken to be the sum of the intrinsic chemical free energy associated with the adsorption process and the free energy due exclusively to electrostatic interactions, which corresponds to the work required to bring a charged particle to a charged interface,2,7,57according to ΔGads ¼ FðΔzÞΦ0 þ ΔGchemical

ð8Þ

Here, F is Faraday’s constant, Δz is the change in the charge state of the surface site upon adsorption, and Φ0 is the initial interfacial potential, which depends on the background electrolyte concentration (eq 4). From the linear relationship between adsorption free energy and initial interfacial potential shown in eq 8, the value of Δz can be determined, which, in turn, can help elucidate the charge state of the species bound to the fused silica/water interface.

III. EXPERIMENTAL METHODS III.A. Substrate and Solution Preparation. Fused silica hemisphere lenses (ISP Optics) were used as the adsorption substrates. The lenses were cleaned for 1 h using NoChromix solution (Godax Laboratories), thoroughly rinsed in Millipore water (18.2 MΩ), submerged in methanol, placed in a sonicator for 6 min, dried in a 100 °C oven for 30 min, oxygen plasma cleaned (Harrick Plasma) for 30 s, and finally stored under Millipore water until further use. Solutions of Al(III), Sc(III), Y(III), La(III), and Gd(III) were prepared in Millipore water using AlCl3 3 6H2O (Alfa Aesar, 99.9995%), ScCl3 3 6H2O (Alfa Aesar, 99.99%), YCl3 3 6H2O (Aldrich, 99.99%), LaCl3 3 6H2O (Aldrich, 99.999%), and GdCl3 3 6H2O (Alfa Aesar, 99.9%) salts, respectively. A background electrolyte was added using additions of solid NaCl (Alfa Aesar, 99+%). Each solution was maintained at pH 4 using a dilute solution of HCl (EMD, ACS grade). ChemEQL60 calculations indicate that the metals under investigation are fully dissolved in solution and no precipitation of the hydroxide salt occurs. III.B. Laser and Flow System. Descriptions of the lasers used here are available elsewhere.6164 Briefly, a regeneratively amplified Ti:sapphire laser (Hurricane, Spectra-Physics) pumping an optical parametric amplifier (OPA-CF, Spectra-Physics) at a kilohertz repetition rate was used to conduct the SHG experiments. The beam, tuned to a fundamental frequency of ω = 600 ( 5 nm, was focused onto the fused silica/water interface after attenuating the power to 0.55 ( 0.05 μJ per 120 fs pulse, sampling an area of the fused silica substrate of approximately 0.8 μm2. The second harmonic signal was isolated using a UV-grade Schott filter and a monochromator set to frequency 2ω, directed 14440



ARTICLE

Figure 2. Proposed Coulombic adsorption pathways for Al(III) and Sc(III) (represented by “M3+”) binding to the fused silica/water interface at pH 4.

Figure 1. Observed adsorption free energy as a function of interfacial potential for Al(III) ion adsorbing to the fused silica/water interface at pH 4 and 298 K. Al(III) concentrations ranged from 0.1 μM to 2 mM. The initial interfacial potential was calculated using the Gouy Chapman expression, and the solid line represents a least-squares fit to the data.

into a photomultiplier tube, amplified, and recorded using a gated single photon detection system. The SHG measurements were carried out under flow conditions to simulate environmental groundwater flow using a previously published flow setup.6164 The fused silica hemisphere was clamped to a custom-built Teflon flow cell with an approximate volume of 7 mL using a Viton O-ring to ensure a leakproof seal. Details of the flow cell are given in the supporting information of ref 65. Peristaltic pumps draw at a flow rate of ∼1 mL/s from a reservoir containing Millipore water and the background NaCl electrolyte adjusted to pH 4.0 ( 0.1, and another reservoir containing the metal analyte solution adjusted to the same pH and background electrolyte. Total analyte concentration was determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES, Varian).

IV. RESULTS AND DISCUSSION IV.A. χ(3) Adsorption Isotherms. Adsorption/desorption studies were performed to assess the binding of Al(III), Sc(III), Y(III), La(III), and Gd(III) to the silica/water interface at pH 4 under dynamic flow conditions. A detailed explanation of how these adsorption/desorption studies are carried out can be found in previous publications.47,63,66 Briefly, SHG χ(3) adsorption isotherms were recorded by collecting adsorption/desorption traces for a range of metal analyte concentrations. This SHG E-field, which is due to the presence of the adsorbing analyte (ESHG), is then plotted against the metal ion concentration. In these experiments, analyte concentrations range from 0.1 μM in low NaCl concentrations to 7 mM in high NaCl concentrations for each metal ion. The ESHG values were obtained by normalizing the square root of the SHG signal intensity in the presence of a trivalent metal analyte to the square root of the average background water signal level before and after analyte flow. These isotherms are then fit with the Langmuir-modified triple layer model shown in eq 5. The value for σ0, the initial surface

charge density of the bare silica surface at pH 4, was previously determined to be 0.004(1) C/m2.47 IV.B. Free EnergyPotential Relation. Adsorption isotherms for each trivalent metal ion adsorbing to the silica/water interface were measured in background NaCl concentrations ranging from 0.1 to 10 mM, corresponding to an initial interfacial potential range of 170 to 54.0 mV at pH 4 on fused silica. These initial interfacial potential values were calculated using the GouyChapman expression.2,3,7,51 The isotherms were fit with the Langmuir-modified triple layer expression described in eq 5. The binding constants, Kads, were obtained from the fit and used to calculate the adsorption free energies with reference to 55.5 M water.4 Table 1 shows the calculated adsorption free energy at each initial interfacial potential value, the background electrolyte concentration, absolute number densities, and the Δz and ΔGchem values obtained from the free energy vs potential analysis (described below). From these data, it becomes apparent that our group of trivalent ions can be sorted into three distinct subgroups when they interact with the silica/water interface in an electrolyte solution at pH 4: one in which the change in the charge state of the surface site upon adsorption (Δz) is e1.0, another where Δz = 1.5, and a third where Δz = 1.15. Al(III) and Sc(III) are represented in the first regime, Gd(III) is represented in the second, and Y(III) and La(III) are represented in the third. Each of these binding regimes is discussed in detail below. 1. Al(III) and Sc(III). Figure 1 shows the calculated adsorption free energy vs initial interfacial potential for Al(III) binding to the fused silica/water interface at pH 4. As expected from eq 8, this plot depicts a linear relationship between the adsorption free energy and the initial interfacial potential. As the initial interfacial potential approaches zero due to increased charge screening by increasing the background electrolyte concentration, adsorption becomes less favorable and the corresponding free energy of adsorption decreases in magnitude. According to eq 8, the slope of the line of best fit to these data is equal to FΔz. Thus, the net change in charge of the silica surface site upon adsorption (Δz) can be calculated from the slope of the SHG adsorption data and used to elucidate possible binding modes. From Figure 1, a Δz value equal to 0.9(1) (the number in parentheses represents the error (1σ) that is associated with the last digit in the value) was calculated for the Al(III) ion adsorbing to the fused silica surface. A very similar trend is observed for Sc(III) adsorbing to fused silica at pH 4. In this case, the free energy/potential analysis yields a Δz equal to 0.79(7). Due to the similarity of the binding parameters of these two ions, and due to several similarities in the chemical properties of the ions 14441



Figure 3. Proposed Coulombic adsorption pathways for Gd(III) ion binding to the fused silica/water interface at pH 4.

(discussed below), binding mechanisms for both Al(III) and Sc(III) are taken to be similar. To explain the noninteger Δz values, the adsorption of Sc(III) and Al(III) is believed to occur through an average of two distinct pathways: one leading to a Δz = 0 and the other with a Δz = +1. As shown in Figure 2, we propose a bidentate surface complex in which the fully hydrated ion initially associates with one deprotonated surface site (SiO) and one neutral surface site (SiOH), which becomes deprotonated upon adsorption. In Figure 2A, the trivalent metal ion adsorbs to one bare SiO surface associated with a sodium ion and one neutral surface site, leading to a net change in the charge of the surface site of +1. In Figure 2B, the metal hydroxide is adsorbed in a similar bidentate mode as in Figure 2A, resulting in a net change in charge of the surface species of 0. Speciation calculations show that, in the aluminum system, 90% of the total dissolved species is the free ion, whereas Al(OH)2+ accounts for 9% of the total dissolved species in solution. Likewise for the scandium system, 60% of the total dissolved species is the free ion, and Sc(OH)2+ accounts for 30% of the total dissolved species in solution. The abundances of the hydroxide species in each system are reflected in the Δz values: the scandium system yields a Δz = 0.79(7), indicating that Figure 2B plays a more dominant role than in the aluminum system, which possesses a higher Δz value of 0.9(1). This proposed binding mechanism is supported by several studies. It has been shown that the aluminum ion does not coordinate with chloride even under high chloride concentrations.67,68 Charlet et al. have shown that, at pH 4, one silica surface site is deprotonated for every adsorbed aluminum ion,69 consistent with the mechanism shown in Figure 2. Houston and co-workers used several NMR techniques to show that Al(III) adsorbs to the amorphous silica substrate in a bidentate fashion;46 however, they were not able to rule out coordination of higher denticity. We therefore conclude, based on the adsorption free energy and interfacial potential analysis described above, that the most probable binding pathway involves a bidentate complex when Al3+ interacts with fused/silica water interfaces maintained under 0.110.0 mM NaCl. Scandium is quite similar to aluminum due to its small ionic radius, noble gas electronic structure, and similar coordination chemistry.27 In this study, we demonstrated that Sc(III) and Al(III) bind to the fused silica substrate in an identical manner under the experimental conditions employed. 2. Gd(III). The Δz value for Gd(III) binding to the fused silica interface was determined using the same method described above. In this instance, however, the adsorption free energy vs initial interfacial potential analysis yields a Δz value equal to +1.55(9). Again, this noninteger value indicates that adsorption

ARTICLE

occurs through two different pathways, shown in Figure 3. Here, we propose that Gd(III) adsorbs to a single deprotonated silica surface site in a decreased charge state due to complexation of a chloride anion. Specifically, Figure 3A describes the Gd(III) 3 Cl complex adsorbing to a single SiO site, resulting in a Δz value of +2. In Figure 3B, the Gd(III) 3 Cl complex adsorbs to the single SiO site by displacing a sodium ion, resulting in a Δz value of +1. The overall measured adsorption is taken to be an average of these two pathways. The binding scheme for Gd(III) proposed in Figure 3 is consistent with studies showing that Gd(III) forms ion pairs with bulk chloride ions under high chloride concentrations.70 While brine conditions are not employed here, bulk Gd(III) chloride complexes are still present in the bulk solution in non-negligible concentrations (concentrations of the chloride complex are 1 order of magnitude lower than the free ion, which is the most abundant species) and their formation is further enhanced at the silica interface due to thermodynamic considerations.71 The adsorption free energies for Gd(III) binding to fused silica calculated from experimental data in this study are comparable to adsorption free energy data taken for different divalent metal cations using this same technique.66 Further, the adsorption free energy vs initial interfacial potential analysis was employed on alkaline earth metal divalent cations and yielded the same Δz value of +1.5,71 lending support to our finding that Gd(III) is effectively binding to fused silica as a divalent cation under the experimental conditions used in this study. 3. Y(III) and La(III). Thermodynamic binding parameters were also measured for both Y(III) and La(III) adsorbing to the fused silica/water interface in varying background electrolyte concentrations. The adsorption free energy vs initial interfacial potential analysis was then performed for these two metal ions, yielding a Δz value equal to +1.15 for both Y(III) and La(III). The specific binding mechanism for these two metal ions remains somewhat unclear. However, due to the intermediate ΔGchem value of both Y(III) and La(III) relative to the other metal ions studied, and due to the intermediate chemical properties of Y(III) and La(III), we propose that these ions bind to the fused silica/water interface as an average of the pathways described in Figures 2 and 3. Bulk studies have shown that both Y(III) and La(III) only weakly associate with chloride ions under high chloride concentrations.72,73 Further, Diniz and Volesky found that La(III) adsorbing to a biomass material was only partially complexed with chloride anions.74 Zhang and co-workers have found that Y(III) adsorbs to the rutile interface in a multisite binding mode,75 furthering the possibility that the mechanism for Al(III) and Sc(III) in Figure 2 applies to Y(III) and La(III). There are many factors affecting ion adsorption to an interface.42,76 In this study, many of the external factors are held constant while differences in the properties of the ions themselves determine variations in binding. These differences are discussed below, as well as how they impact the mode in which the ion adsorbs to the silica interface. IV.C. Nature of Chloride Complexation. In order to explain the differences in chloride complexation among the ions studied here, one must examine several intrinsic properties of the ions themselves. First, we have shown experimentally that Sc(III) and Al(III) do not complex to chloride under the experimental conditions employed here. Al(III) ions have a very large charge to radius ratio, resulting in strongly hydrated ions in aqueous solution and causing association with a chloride ion to be thermodynamically unfavorable.67,68 Due to its similarly small ionic radius and high charge, the Sc(III) ion also exhibits a large 14442



ARTICLE

Figure 5. Equation 9 plotted as a function of absolute interfacial potential with varying values of n (1, 2, and 3) represented with red curves. The dotted vertical line represents an interfacial potential of 110 mV, and the dotted horizontal dotted lines are guides to the eye and represent where 110 mV intersects the function with n = 2 and n = 3.

Figure 4. (A) Adsorption free energy determined from the experiments for each metal ion studied in 10 mM NaCl and pH 4 plotted against the atomic number for the ion. The solid line represents a least-squares fit to the data. (B) Adsorption free energy determined from the experiments for each metal ion studied in 10 mM NaCl and pH 4 plotted against the Gibbs energy of formation for each aqueous ion. Ion concentrations ranged from 0.2 μM to 3 mM. The solid line represents a least-squares fit to the data, excluding the Gd point.

hydration enthalpy along with a similar aversion to chloride complexation.27 While Sc(III) chloride complexes have been observed, these complexes are described as very weak and only exist in highly concentrated chloride solutions.77,78 The ionic radii of the remaining ions, Y(III), La(III), and Gd(III), are very similar.27 In order to explain the differences in binding seen in section IV.B, variations in electronic configuration must be considered. In 1954, Seaborg and co-workers found that trends in chloride complexation among the lanthanides and actinides could not be rationalized through ionic interactions alone.79 Instead, they discovered that complexation of a covalent nature was found to play a role. Additionally, Ikeda et al. found

that the lanthanide ions interact covalently with a tertiary pyridine resin.80 Further, the mobilities of the ions through the resin were affected by the electron configuration of the elements: La(III) and Y(III) showed increased mobility through the resin with respect to the rest of the lanthanide series. These covalent interactions have been found to involve the outermost s and p orbitals of the metal ion and not the f orbitals, which are considered to be too shielded to participate in complexation interactions.81 Adamo and Maldivi used computational methods to show that Gd(III) possesses greater covalent character than the early lanthanide ions due to changes in orbital energy levels.82 Specifically, the energy gap between the 6s orbital of the metal and the p orbitals of the halide decreases across the lanthanide series. If chloride complexation with trivalent ions contains a small covalent character, this degree of covalency is then dependent on the electronic configurations of the ions themselves, which, in turn, results in greater chloride complexation in Gd(III) than in the early lanthanide ions, and a different binding mechanism for Y(III)/La(III) and Gd(III), as discussed in section IV.B. Predictability is an important tool when studying metal ion adsorption to a mineral oxide interface. Figure 4A shows the adsorption free energy for each ion measured in a 10 mM NaCl plot against the atomic number of that ion. A linear trend is observed as the magnitude of the adsorption free energy decreases from Al(III) to Gd(III), allowing for a degree of predictability within the trivalent metal ions. The atomic number parameter encompasses the changes in both ionic size and electron configuration, two of the main factors affecting the binding mechanisms of these ions under the experimental conditions employed here. A linear trend is also observed in Figure 4B, which shows the adsorption free energy measured in 10 mM NaCl against the Gibbs free energy of formation for the aqueous ions.83 Gd(III) is an outlier in this trend due to the effects of the lanthanide contraction and was not included in the linear fit to the data.

V. CONCLUSIONS AND ENVIRONMENTAL IMPLICATIONS In this work, we have investigated the binding of Al(III), Sc(III), Y(III), La(III), and Gd(III) to the fused silica/water interface held at pH 4 and varying electrolyte concentration by 14443



ARTICLE

using the Eisenthal χ(3) technique. We quantified the adsorption free energy, binding constant, absolute number density, and interfacial charge density for each adsorbed ion in real time and under dynamic flow conditions. The analysis of the relationship between the adsorption free energy of each ion binding to the fused silica/water interface and the initial interfacial potential of that interface is consistent with the notion that Al(III) and Sc(III) adsorb to the fused silica/water interface as trivalent ions. On the other hand, Gd(III) is likely to adsorb to the fused silica surface as a divalent cation due to the presence of a complexing chloride anion, while both Y(III) and La(III) exhibit an average of two different binding modes. Changes in the charge state of an ion at an interface would have a large impact on electrostatic adsorption models, as shown in the following equation: nþ ΨF=RT n nþ ¼ Xbulk ðe Þ Xads

ð9Þ

Here, Xadsn+ and Xbulkn+ represent the surface coverage and bulk concentration of an ion with charge n, respectively. Ψ is the interfacial potential, F represents Faraday’s constant, and R and T have their usual meanings. In Figure 5, eq 9 is plotted as a function of changing interfacial potential. For a potential at the adsorption plane of 110 mV, which is typical for fused silica under water at pH 4 and 1 mM NaCl solution, Figure 5 shows that assigning Gd(III) a charge of +3 instead of +2 results in a difference in the output of eq 9 of almost 2 orders of magnitude. This difference becomes exponentially larger as the absolute value of the interfacial potential increases. Therefore, assuming the charge state of an adsorbed ion to be same at a surface as in the bulk solution can lead to drastic miscalculations of absolute surface coverages, and thus rates. By quantifying the exponential sensitivity of eq 9 to the charge state of trivalent ions under slightly acidic environmental conditions directly at the fused silica/water interface, we provide benchmarks for theory calculations describing the interactions of metal ions with oxide interfaces in geochemistry and hope to improve the prediction of trivalent metal ion transport through groundwater environments.

’ ASSOCIATED CONTENT

bS

Supporting Information. Representative adsorption isotherms for each free energy vs interfacial potential analysis plot for Al(III), Sc(III), Y(III), La(III), and Gd(III), as well as the derivation of eq 3. This material is available free of charge via the Internet at http://pubs.acs.org.

’ ACKNOWLEDGMENT This material is based upon work supported by a National Science Foundation Graduate Research Fellowship to S.A.S. under Grant 1000118395. This work was also supported by the National Science Foundation Environmental Chemical Sciences program under Grant CHE-0950433 and the Director, Office of Biological and Environmental Research, of the U.S. Department of Energy under Grant DE-PS02-ER09-07. F.M.G. gratefully acknowledges an Irving M. Klotz professorship. We thank Spectra-Physics Lasers, a division of Newport Corp., for equipment support. The ICP-AES analysis was completed at the Northwestern University Integrated Molecular Structure Education and Research Center (IMSERC).

’ REFERENCES (1) Evangelou, V. P. Environmental Soil and Water Chemistry; John Wiley & Sons: New York, 1998. (2) Morel, F. M. M.; Hering, J. G. Principles and Applications of Aquatic Chemistry; John Wiley & Sons, Inc.: New York, 1993. (3) Stumm, W.; Morgan, J. J. Aquatic Chemistry, 3rd ed.; John Wiley & Sons, Inc.: New York, 1996. (4) Adamson, A. W. Physical Chemistry of Surfaces, 5th ed.; John Wiley & Sons: New York, 1990. (5) Somorjai, G. A. Chemistry in Two Dimensions; Cornell University: Ithaca, NY, 1981. (6) M€uller, B. ChemEQL, 3.0 ed.; EAWAG: Kastanienbaum, Switzerland, 1996. (7) Langmuir, D. Aqueous Environmental Geochemistry; Prentice Hall: Upper Saddle River, NJ, 1997. (8) Kagan, H. B. Chem. Rev. 2002, 102, 1805–1806. (9) Gawryszewska, P.; Sokolnicki, J.; Legendziewicz, J. Coord. Chem. Rev. 2005, 249, 2489–2509. (10) Kalinowski, J.; Mezyk, J.; Meinardi, F.; Tubino, R.; Cocchi, M.; Virgili, D. Chem. Phys. Lett. 2008, 453, 82–86. (11) Molander, G. A.; Romero, J. A. C. Chem. Rev. 2002, 102, 2161–2185. (12) Shibasaki, M.; Yoshikawa, N. Chem. Rev. 2002, 102, 2187–2209. (13) Duimstra, J. A.; Femia, F. J.; Meade, T. J. J. Am. Chem. Soc. 2005, 127, 12847–12855. (14) Folger, T. The Secret Ingredients of Everything. In National Geographic, 2011, Vol. 219, pp 136145. (15) Haxel, G. B.; Hedrick, J. B.; Orris, G. J. Rare Earth Elements; Critical Resources for High Technology; U. S. Geological Survey: Reston, VA, 2002. (16) Allen, P.; Bucher, J.; Shuh, D.; Edelstein, N.; Craig, I. Inorg. Chem. 2000, 39, 595–601. (17) Sawada, K.; Abdel-Aal, K.; Tan, K.; Satoh, K. Dalton Trans. 2005, 3291–3296. (18) Bau, M. Geochim. Cosmochim. Acta 1999, 63, 67–77. (19) Johannesson, K. H.; Hawkins, J. D. L.; Cortes, A. Geochim. Cosmochim. Acta 2006, 70, 871–890. (20) Sholkovitz, E. R. Earth Planet. Sci. Lett. 1992, 114, 77–84. (21) Kucera, J.; Mizera, J.; Randa, Z.; Vavrova, M. J. Radioanal. Nucl. Chem. 2007, 271, 581–587. (22) Bau, M.; Dulski, P. Earth Planet. Sci. Lett. 1996, 143, 245–255. (23) Protano, G.; Riccobono, F. Environ. Pollut. 2002, 117, 499–514. (24) Weltje, L.; Heidenreich, H.; Zhu, W. Z.; Wolterbeek, S.; Korhammer, S.; De Goeij, J. J. M.; Markert, B. Sci. Total Environ. 2002, 286, 191–214. (25) Zhang, H.; Feng, J.; Zhu, W. F. L.; Liu, C.; Xu, S.; Shao, P.; Wu, D.; Yang, W.; Gu, J. Biol. Trace Elem. Res. 2000, 73, 1–17. (26) Hirano, S.; Suzuki, K. T. Environ. Health Perspect. 1996, 104, 85–95. (27) Housecroft, C. E.; Sharpe, A. G. Inorganic Chemistry, 2nd ed.; Pearson Education Ltd.: Essex, U.K., 2005. (28) Flaten, T. P. Brain Res. Bull. 2001, 55, 187–196. (29) Dahlgren, R. Nature 1994, 368, 838–841. (30) Driscoll, C.; Baker, J.; Bisogni, J.; Schofield, C. Nature 1980, 284, 161–164. (31) Rabadjieva, D.; Tepavitcharova, S.; Todorov, T.; Dassenakis, M.; Paraskevopoulou, V.; Petrov, M. Environ. Monit. Assess. 2009, 159, 353–366. (32) Wu, P.; Tang, C.; Liu, C.; Zhu, L.; Pei, T.; Feng, L. Environ. Geol. 2009, 57, 1457–1467. (33) Kochian, L. Annu. Rev. Plant Phys. 1995, 46, 237–260. (34) Banks, W.; Kastin, A. Neurosci. Biobehav. Rev. 1989, 13, 47–53. (35) Ferreira, P. C.; Piai, K. D. A.; Takayanagui, A. M. M.; SeguraMunoz, S. I. Rev. Latino-Am. Enfermagem 2008, 16, 151–157. (36) Walton, J. Neurotoxicology 2006, 27, 385–394. (37) Diniz, V.; Volesky, B. Water Res. 2005, 39, 239–247. (38) Lehrmann, R.; Seelig, J. Biochim. Biophys. Acta, Biomembr. 1994, 1189, 89–95. 14444


The Journal of Physical Chemistry A (39) Yermiyahu, U.; Rytwo, G.; Brauer, D. K.; Kinraide, T. B. J. Membr. Biol. 1997, 159, 239–252. (40) Huittinen, N.; Rabung, T.; Luetzenkirchen, J.; Mitchell, S. C.; Bickmore, B. R.; Lehto, J.; Geckeis, H. J. Colloid Interface Sci. 2009, 332, 158–164. (41) Koeppenkastrop, D.; Decarlo, E. Environ. Sci. Technol. 1993, 27, 1796–1802. (42) Kosmulski, M. J. Colloid Interface Sci. 1997, 195, 395–403. (43) Pepper, S.; Hull, L.; Bottenus, B.; Clark, S. Radiochim. Acta 2006, 94, 229–237. (44) Zhang, Z.; Fenter, P.; Cheng, L.; Sturchio, N. C.; Bedzyk, M. D.; Predota, M.; Bandura, A. V.; Kubicki, J. D.; Lvov, S. N.; Cummings, P. T.; et al. Langmuir 2004, 20, 4954–4969. (45) Piasecki, W.; Sverjensky, D. A. Geochim. Cosmochim. Acta 2008, 72, 3964–3979. (46) Houston, J. R.; Herberg, J. L.; Maxwell, R. S.; Carroll, S. A. Geochim. Cosmochim. Acta 2008, 72, 3326–3337. (47) Jordan, D. S.; Malin, J. N.; Geiger, F. M. Environ. Sci. Technol. 2010, 44, 5862–5867. (48) Eisenthal, K. B. Chem. Rev. 1996, 96, 1343–1360. (49) Shen, Y. R. The Principles of Nonlinear Optics; John Wiley & Sons: New York, 1984. (50) Boyd, R. W. Nonlinear Optics, 3rd ed.; Academic Press: London, 2008. (51) Yan, E. C. Y.; Liu, Y.; Eisenthal, K. B. J. Phys. Chem. B 1998, 102, 6331–6336. (52) Brown, G. E.; Henrich, V. E.; Casey, W. H.; Clark, D. L.; Eggleston, C.; Felmy, A.; Goodman, D. W.; Gratzel, M.; Maciel, G.; Mccarthy, M. I.; et al. Chem. Rev. 1999, 99, 77–174. (53) Koopal, L. K. Electrochim. Acta 1996, 41, 2293–2305. (54) Hayes, K.; Papelis, C.; Leckie, J. J. Colloid Interface Sci. 1988, 125, 717–726. (55) Salafsky, J. S.; Eisenthal, K. B. J. Phys. Chem. B 2000, 104, 7752–7755. (56) Atkins, P.; De Paula, J. Physical Chemistry, 7th ed.; W. H. Freeman and Co.: New York, 2002. (57) Castro, A.; Bhattacharyya, K.; Eisenthal, K. B. J. Chem. Phys. 1991, 95, 1310. (58) Sverjensky, D. A. Geochim. Cosmochim. Acta 2005, 69, 225–257. (59) Sverjensky, D. A.; Fukushi, K. Environ. Sci. Technol. 2006, 40, 263–271. (60) M€uller, B. ChemEQL, 3.0 ed.; Limnological Research Center, Swiss Federal Institute of Aquatic Science and Technology: Kastanienbaum, Switzerland, 2005. (61) Al-Abdableh, H. A.; Mifflin, A. L.; Musorrafiti, M. J.; Geiger, F. M. J. Phys. Chem. B 2005, 109, 16852–16859. (62) Hayes, P. L.; Gibbs-Davis, J. M.; Musorrafiti, M. J.; Mifflin, A. L.; Scheidt, K. A.; Geiger, F. M. J. Phys. Chem. C 2007, 111, 8796–8804. (63) Hayes, P. L.; Malin, J. N.; Konek, C. T.; Geiger, F. M. J. Phys. Chem. A 2008, 112, 660–668. (64) Musorrafiti, M. J.; Konek, C. T.; Hayes, P. L.; Geiger, F. M. J. Phys. Chem. C 2008, 112, 2032–2039. (65) Gibbs-Davis, J. M.; Kruk, J. J.; Konek, C. T.; Scheidt, K. A.; Geiger, F. M. J. Am. Chem. Soc. 2008, 130, 15444–15447. (66) Malin, J. N.; Hayes, P. L.; Geiger, F. M. J. Phys. Chem. C 2009, 113, 2041–2052. (67) Rudolph, W.; Mason, R.; Pye, C. Phys. Chem. Chem. Phys. 2000, 2, 5030–5040. (68) Vinogradov, E.; Smirnov, P.; Trostin, V. Russ. Chem. Bull. 2003, 52, 1253–1271. (69) Charlet, L.; Schindler, P.; Spadini, L.; Furrer, G.; Zysset, M. Aquat. Sci. 1993, 55, 291–303. (70) Beuchat, C.; Hagberg, D.; Spezia, R.; Gagliardi, L. J. Phys. Chem. B 2010, 114, 15590–15597. (71) Malin, J. N.; Holland, J. G.; Geiger, F. M. J. Phys. Chem. C 2009, 113, 17795–17802. (72) Ramos, S.; Neilson, G.; Barnes, A.; Mazuelas, A. J. Phys. Chem. B 2001, 105, 2694–2698.

ARTICLE

(73) Yaita, T.; Ito, D.; Tachimori, S. J. Phys. Chem. B 1998, 102, 3886–3891. (74) Diniz, V.; Volesky, B. Water Res. 2005, 39, 2229–2236. (75) Zhang, Z.; Fenter, P.; Cheng, L.; Sturchio, N.; Bedzyk, M.; Machesky, M.; Wesolowski, D. Surf. Sci. 2004, 554, L95–L100. (76) Kosmulski, M. J. Colloid Interface Sci. 1999, 211, 410–412. (77) Pye, C. C.; Corbeil, C. R. Can. J. Chem. 2002, 80, 1331–1342. (78) Rudolph, W.; Pye, C. J. Solution Chem. 2000, 29, 955–986. (79) Diamond, R.; Street, K.; Seaborg, G. J. Am. Chem. Soc. 1954, 76, 1461–1469. (80) Ikeda, A.; Suzuki, T.; Aida, M.; Ohtake, K.; Fujii, Y.; Itoh, K.; Hara, M.; Mitsugashira, T. J. Alloys Compd. 2004, 374, 245–248. (81) Choppin, G. J. Alloys Compd. 2002, 344, 55–59. (82) Adamo, C.; Maldivi, P. Chem. Phys. Lett. 1997, 268, 61–68. (83) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Halow, I.; Bailey, S. M.; Schumm, R. H. NBS Tech. Note (U. S.) 1968, No. 270-3.

14445


Exponential Sensitivity and Speciation of Al(III), Sc(III), Y(III), La(III

Recommend Documents