Modeling the Adsorption of Oxalate onto Montmorillonite - American

Oct 7, 2015 - Modeling the Adsorption of Oxalate onto Montmorillonite. M. Elena Ramos,. †. Caglayan Emiroglu,. ‡. David García,. §. C. Ignacio S...
5 downloads 0 Views 3MB Size
Article pubs.acs.org/Langmuir

Modeling the Adsorption of Oxalate onto Montmorillonite M. Elena Ramos,† Caglayan Emiroglu,‡ David García,§ C. Ignacio Sainz-Díaz,† and F. Javier Huertas*,† †

Instituto Andaluz de Ciencias de la Tierra, CSIC-Universidad de Granada, Avenida de las Palmeras 4, 18100 Armilla, Granada, Spain Giresun University, 28200 Giresun, Merkez/Giresun, Turkey § Amphos XXI Consulting S.L., Passeig de Garcia i Faria, 49-51, E08019, Barcelona, Spain ‡

ABSTRACT: In this work, a multiscale modeling of the interaction of oxalate with clay mineral surfaces from macroscale thermodynamic equilibria simulations to atomistic calculations is presented. Previous results from macroscopic adsorption data of oxalate on montmorillonite in 0.01 M KNO3 media at 25 °C within the pH range from 2.5 to 9 have been used to develop a surface complexation model. The experimental adsorption edge data were fitted using the triple-layer model (TLM) with the aid of the FITEQL 4.0 computer program. Surface complexation of oxalate is described by two reactions: >AlOH + Ox2− + 2H+ = >AlOxH + H2O (log K = 14.39) and >AlOH + Ox2− + H+ = >AlOx− + H2O (log K = 10.39). The monodentate complex >AlOxH dominated adsorption below pH 4, and the bidentate complex >AlOx− was predominant at higher pH values. Both of the proposed innersphere oxalate species are qualitatively consistent with previously published diffuse reflectance FTIR spectroscopic results for oxalate on montmorillonite edge surface (Chem. Geol. 2014, 363, 283−292). Atomistic computational studies have been performed to understand the interactions at the molecular level between adsorbates and mineral surface, showing the atomic structures and IR frequency shifts of the adsorption complexes of oxalate with the edge surface of a periodic montmorillonite model.

1. INTRODUCTION The interaction of organic acids with mineral surfaces in electrolyte solutions is of great interest in a wide range of geochemical topics and processes. One of the properties of low molecular weight organic anions is to promote the mineral dissolution. Their capacity to catalyze the dissolution reaction arises from their ability to bind strongly to mineral surfaces in an inner-sphere manner.2−5 The polarization of the surface metal−organic anion bonds gives rise to the release of the metal cations from the surface through a rate-determining detachment step. Although extensive work has been done to identify the type and structure of surface complexes of relatively simple organic acids at mineral/water interfaces,6−9 a molecular-level understanding of these interactions is not complete. Two major types of surface complexes have been suggested: inner-sphere (direct bond between carboxylate oxygen and surface cations) and outer-sphere complexes (carboxylate oxygen held at the surface through a combination of hydrogen bonding and electrostatic interactions). In a previous study, diffuse reflectance Fourier-transform infrared (DR-FTIR) spectroscopy was used to identify the types of oxalate surface complexes on the montmorillonite surface.1 However, due to the limited amount of oxalate sorbed, it was difficult to make a statement about the type of complexes formed. The quantitative adsorption was also studied at different oxalate concentrations. In order to better understand the oxalate−montmorillonite interaction, the experimental © XXXX American Chemical Society

results obtained are integrated with a surface complexation model, which establishes the stoichiometry of the adsorption reactions and provides a thermodynamic characterization of the equilibria involved. Atomistic studies can be a useful tool to understand experimental phenomena related with the interactions of organics with crystalline mineral surfaces,10 especially clay mineral surfaces.11−14 The phyllosilicate mineral montmorillonite is the focus of the present study because of its prevalence in soils and sediments in temperate regions. Montmorillonite dissolution mechanism and reactivity have been widely investigated in previous studies.15−25 In the present work, the adsorption of oxalate at the edge surface of montmorillonite has been successfully modeled on the basis of results from previous investigations.1 The results of this study allow one to make quantitative predictions that can facilitate evaluation of the potential role of mineral surface chemistry in several geochemical processes involving the interactions of organic molecules and mineral surfaces. Received: July 29, 2015 Revised: September 28, 2015

A

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Both inner- and outer-sphere complexation at the montmorillonite surface may be considered according to the experimental DR-FTIR results previously reported by Ramos et al.1 The DR-FTIR results also indicated that oxalate adsorbs in a mononuclear manner; therefore, binuclear complexes have not been considered in our conceptual model. The triple-layer model28 (TLM) can distinguish inner- and outer-sphere complexes: sorbates forming inner-sphere complexes are typically placed on the o-plane, whereas those forming outer-sphere complexes are placed on the β-plane (Figure 2). The inner layer

2. MATERIALS AND METHODS 2.1. Materials. The montmorillonite sample used in the present study was the same as that used by Ramos et al.25 The specific surface area measured by BET26 was 111 m2 g−1 with an associated uncertainty of 10%. The edge surface area was estimated to be 6.5 m2 g−1.20 For a complete montmorillonite characterization, the reader is referred to the work of Ramos et al.25 2.2. Adsorption Experiments. In this study, we used adsorption data previously reported by Ramos et al.1 For the sake of clarity, the experimental information that is relevant to this data is briefly provided. Quantitative adsorption of oxalate on montmorillonite was studied at 25 °C using the batch equilibrium method. Batch samples with a solid concentration of 10 g L−1 and a total concentration of oxalate of 0.1 mmol L−1 were prepared in 30 mL Corex centrifuge tubes with polytetrafluoroethylene (PTFE) caps. In order to cover a wide range of pH values (from 3 to 9 units), precise volumes of HCl or KOH were added to each sample using 10 mmol L−1 KCl as background electrolyte. Preliminary kinetic experiments indicated that the adsorption of oxalate reached a steady state within the first 5 h after addition of oxalate to a montmorillonite suspension. After the filtration of every supernatant, oxalate was analyzed using a Dionex ion chromatograph equipped with an AS50 autosampler, a GP50 gradient pump, an AS50 thermal compartment, EG40 eluent generator, and ES50 electrochemical detector. By calculating the difference between the known total concentration and the remaining concentration in the supernatant after equilibration, the quantity of oxalate adsorbed on the surface of montmorillonite was determined in each sample. The adsorption modes of oxalate on montmorillonite have been examined in a previous work1 using DR-FTIR spectroscopy. 2.3. Modeling the Adsorption Data. The macroscopic adsorption edge data shown in Figure 1 can be tentatively linked with the DR-FTIR data results using a suitable surface complexation model.

Figure 2. Schematic representation of the electrical properties and location of the complexes formed at the mineral−solution interface.28 (between the surface plane and the β-plane) and the outer layer (between the β-plane and the bulk solution) have their own constant capacitance values, C1 and C2, respectively. In our study, the surface site density and protolysis constants for aluminol edge sites optimized by Rozalen et al.23 were adopted. This set of constants was obtained by taking into account the smectite permanent charge due to isomorphic substitutions as well as pHdependent charge at the crystal edges. Protonation and deprotonation reactions are assumed to occur on the inner plane and are described by the following reactions:

Figure 1. Experimental (symbols) and modeled (lines) data of adsorption of oxalate on montmorillonite as a function of pH at varying ligand concentrations (experimental data from the work of Ramos et al.;1 modeled data were calculated in this study).

>AlOH 2+ = >AlOH + H+

>AlOH = >AlO− + H+ The octahedral sheet contains Al3+ as the main component, with Mg2+ and Fe3+ as minor components due to isomorphic substitutions. The so-called aluminol sites correspond to the average octahedral cation in our sample, which should be very close to >AlOH. The free concentration of the background electrolyte ions (K+, NO3−), 0.01 mol L−1, is assumed to remain constant in the experiments. The equilibrium constants for the oxalate acid/base reactions and the montmorillonite surface protonation/deprotonation, as well as the density of surface aluminol groups and the outer and inner capacitance for the electrostatic double layer, were considered as known parameters and used without modifications (Table 1).

The montmorillonite edge surface contains functional groups with oxygen atoms, which are coordinated by different numbers of protons and silicon and aluminum. These groups can accept and release protons and also take part in complexation reactions with metal ions and ligands. Because of the low ability of Si to form complexes in solution,27 complexation reactions of silanol groups of the surface have not been considered in our conceptual model. The reactive edge surface can be described as a homogeneous surface with only one type of reactive hydroxyl group (>AlOH), which is responsible for all surface complexation reactions. This relatively simple model of surface groups is often satisfactory for modeling of adsorption data.7 B

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

harmonic force constants derivated from finite atomic displacements at the same theoretical level to confirm the nature of the stationary points, finding only positive eigenvalues for minima. First-principles energy calculations of the periodic crystal model were performed using DFT methods based on numerical atomic orbital (NAO) methodology with periodical boundary conditions implemented in the SIESTA code.37 The basis sets are made of strictly localized NAOs with a localization cutoff radii corresponding to an energy shift of 270 meV. These basis sets used in this work are doubleZ polarized (DZP) following the perturbative polarization scheme. We used the generalized gradient approximation (GGA) with the Perdew−Burke−Ernzerhof (PBE) parametrization of the exchange− correlation functional. Core electrons were replaced by normconserving pseudopotentials that simulate the interaction between valence electrons and cores (nuclei plus core electrons). By means of this approximation, the valence wave functions are substituted by pseudowave functions that do not present strong oscillations in the core region. A uniform mesh with a certain plane-wave cutoff energy (Emesh cutoff) was used to represent the electron density, the local part of the pseudopotential, and the Hartree and exchange−correlation potentials. Pseudopotentials and basis sets of the mineral atoms were previously optimized.38 After preliminary optimization of calculation parameters reported elsewhere,39 we used a Emesh cutoff = 200 Ry and two k points in the irreducible wedge of the Brillouin zone for all calculations. In all structures, all atoms were relaxed by means of conjugated gradient minimizations with a force tolerance of 0.01 eV/Å. The vibration frequencies of the periodical models were obtained with the VIBRA code that is included in the SIESTA package. Several derivatives of oxalic were calculated: oxalic acid, monosodium oxalate, and disodium oxalate. Each molecule optimized as isolated gas phase with Gaussian was placed in a cubic periodical box of 20 Å axis to be optimized with SIESTA, avoiding intermolecular interactions. A periodic model of the crystal structure of phyllosilicate was generated on the basis of experimental atomic coordinates and lattice cell parameters of illite/smectite.40 Two phyllosilicate models were generated: one of illite with one tetrahedral Al3+ substitution per unit cell and K+ as interlayer cation, K(Al4)(Si7Al1)O20(OH)4, and one of pyrophyllite where all octahedral cations are Al, all tetrahedral cations are Si, and no interlayer cation was included, (Al4)(Si8)O20(OH)4. The H atom coordinates were included manually and optimized previously.41 These structures are dioctahedral in the transvacant crystal form and completely dry in our simulations. For the study of the adsorption process, we created a clean surface corresponding to the (100) plane of pyrophyllite as an edge surface of mineral. This is not the most stable surface and it does not correspond to the largest specific surface; however, it is the most reactive surface. The anion exclusion volume formed by the excess of negative charge on the basal surface prevents the adsorption of oxalate.1 In order to generate a surface along the (100) plane, an additional vacuum space of 20 Å over this surface was created by applying periodical boundary conditions with u = 9.07 Å and v = 10.06 Å (related with the axis b and c of the pristine bulk mineral crystal). The depth of the slab is about 13 Å (two unit cells in the pristine axis a of crystal bulk), and the real vacuum space is about 17 Å. This model allows us to study the interaction of adsorbate on only one surface neglecting the effect on the other surface. All cut dangling bonds were completed with H, OH, and OH2 ligands, obtaining a neutral surface model (Figure 3a). Nevertheless, an additional protonated model was also generated by forming a Al−OH2+ group on the surface (Figure 3). To validate the calculation parameters, the illite model was optimized at variable volume, obtaining a unit cell of a = 5.20 Å, b = 8.97 Å, c = 10.02 Å, α = 90.2°, β = 102.9°, and γ = 89.8°, according with experimental data. Hence, this methodology and calculation parameters are valid for our studies.

Table 1. Parameters of the Surface Complexation Model log K

ref

Protonation of Oxalate Ox2− + H+ = HOx− 4.27 Ox2− + 2H+ = H2Ox 5.52 Surface Reactions (I = 0.01 mol L−1) + >AlOH2 = >AlOH + H+ 6.44 >AlOH = >AlO− + H+ −9.94 >AlOH + H+ + NO3− = >AlOH2+−NO3− 8.30 >AlOH + K+ = >AlO−−K+ + H+ −9.20 >AlOH + Ox2− + H+ = >Al−Ox− + H2O 10.39 >AlOH + Ox2− + 2H+ = >AlOxH + H2O 14.39 Other Parameters site density for >AlOH (sites nm−2) 3.55 Sedges (m2 g−1) 6.5 inner capacitance (F m−2) 1.0 outer capacitance (F m−2) 0.2

29 29 23 23 31 32 this study this study 23 23 28 28

The protonation reaction constants of oxalate were collected from the work of Filius et al.29 Preliminary tests with the chemical speciation program MEDUSA30 showed that the formation of Al−oxalate aqueous complexes is negligible under our experimental conditions (oxalate concentration = 0.1 mmol L−1); thus, they were not considered in the model. The equilibrium constants for the outersphere complexes >AlOH2+−NO3− and >AlO−−K+ published previously for γ-Al2O331 and illite,32 respectively, were used to complete the set of constants required to implement the TLM. Usually, background electrolyte ions are assumed to be bind in the dplane. Johnson et al.33 observed a poor data convergence with FITEQL 4.0,34 modeling oxalate adsorption onto corundum, and suggested that placing the electrolyte ions at the β-plane improves the convergence. Their experimental conditions of low background electrolyte concentration (0.01 mol L−1), wide pH range, and very low oxalate adsorption are similar to ours. The experimental data were evaluated using the program FITEQL 4.0.34 Since oxalate was present in the solution as H2Ox, HOx−, and Ox2− and Ox2− is used as main component in the definition of the equilibrium model for the surface complexation, it is convenient to represent the measured aqueous oxalate concentration by a parameter that is independent of the aqueous speciation. Therefore, the entering of data into FITEQL was done as described in Westall and Herbelin35 using a dummy component, Ox(ads), which represents the total concentration of adsorbed oxalate. Ox(ads) was calculated for each data point as the difference between the total concentration of oxalate added and the analyzed total concentration in solution (see section 2.2). The evaluation of the experimental data in the system H+ → AlOH−oxalate consisted of a systematic test of combinations of inneror outer-sphere complexes with different compositions and an optimization of the corresponding equilibrium constants. The combination of complexes giving the lowest average WSOS/DF [V(Y)], which is the weighted sum of squares divided by the degrees of freedom of the problem, was considered the best fitting model. WSOS/DF values between 0.1 and 20 are indicative of good agreement between model and experiment. Moreover, special attention was focused on making sure that the final model was in good semiquantitative agreement with previously reported FTIR spectroscopic measurements.1 2.4. Computational Methodology and Atomistic Models. Quantum mechanical calculations of the isolated oxalic acid derivatives were performed using the Gaussian 03 program package.36 The electronic structure was calculated with a valence double-ζ basis set augmented with diffuse functions and polarization functions for all atoms including the second-order Moeller−Plesset approximation at the all-electron MP2 level (MP2/6-31+G**). All geometries were fully optimized using the Berny analytical gradient method with no geometry constraint. Normal mode analyses were carried out through

3. RESULTS 3.1. Adsorption Model. The TLM fit for oxalate sorption to montmorillonite is satisfactory [V(Y) = 21.7], although C

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

whole pH range studied. However, the introduction of outersphere complexes in the model led to poor convergence characteristics. On the other hand, the addition of the monodentate inner-sphere complex >AlOxH in the model led to a good fitting at pH AlOxH dominates from pH 2.5 up to pH 4. The bidentate inner-sphere complex >AlOx− is significant above pH 4.

Figure 4. Speciation diagram of oxalate adsorption at 25 °C from 0.1 mmol L−1 solution onto montmorillonite, calculated from the surface complexation model. Model parameters are detailed in Table 1. Filled circles: experimental adsorption edge data previously reported by Ramos et al.1 Black solid lines: model results for the individual contribution of each oxalate surface complex. Dashed black line: overall model fit to experimental data for oxalate sorption. Figure 3. Crystal structure of the (100) surface of pyrophyllite optimized in the neutral (a−c) and protonated form (d). (a) View from the (100) plane of a periodical supercell, (b) View of a supercell from the (010) plane showing the size of the slab created and the vacuum over the surface. (c) Details of the neutral surface and (d) details of the protonated surface. The H, O, Al, and Si are represented in cyan, red, pink, and yellow. The atoms of the surface are highlighted as balls. Distances are in angstroms.

In general, there is a good fit to all data. Nevertheless, at pH >6 the model fit probably overestimates the extent of oxalate adsorption. This phenomenon can be due to competitive solution complexation of oxalate by dissolved Al3+.1 The model fit for adsorption edges with higher concentrations of oxalate (0.3, 1, and 1.5 mmol L−1) is also satisfactory, and the general trend for the model fit is in good agreement with the experimental observations, taking into account the data dispersion. The adsorption reaction constant for the complex >AlOx− (log K = 10.39) is very similar to that reported for Al2O3 by Zutic and Stumm2 (log K = 11.00) and for α-Al2O3 by Johnson et al.33 (log K = 10.22). 3.2. Atomistic Calculations. 3.2.1. Adsorbates. The molecular structure of oxalic acid optimized at both calculation levels, MP2/6-31+G** and GGA/PBE, corresponds to the most stable conformer, where the hydroxyl group is oriented toward the carbonyl group of the vicinal carboxylic moiety with an intramolecular H bond of 1.94 Å (Figure 5) according to previous theoretical43 and spectroscopic works.43,44 The monosodium oxalate shows the same coplanar molecular structure for both calculation levels in the most stable conformer with the H atom oriented toward the vicinal carboxylate group with an intramolecular H bond of 1.75 Å (at GGA/PBE level) according to previous works43 (Figure 5). Slight differences were observed in the O···Na distances: 2.16 Å at GGA/PBE level and 2.62−2.69 Å at MP2/6-31+G** level. The optimized disodium oxalate molecule has d(O···Na) = 2.14

montmorillonite is not a strong oxalate sorbent. The best fit is achieved when oxalate sorption is represented by the reactions >AlOH + Ox 2 − + 2H+ = >AlOxH + H 2O >AlOH + Ox 2 − + H+ = >AlOx − + H 2O

These reactions support the hypothesis that oxalate sorption to montmorillonite is dominated by specific sorption to amphoteric aluminol edge sites in the whole pH range studied (3−9). The DR-FTIR data obtained in the previous work1 agree with the proposed model under acidic conditions, since the spectra suggested that the formation of the complex >Al− Ox− was favored at low pH. However, modelization results indicate that this complex exists in the whole pH range studied, with a maximum concentration at pH 5. This complex involves one surface Al atom to form a mononuclear, bidentate, fivemember ring, which has been also observed for adsorption of oxalate on Al2O3 (corundum)33 and γ-AlO(OH) (boehmite).9,42 The infrared spectra also suggested the formation of the monodentate outer-sphere complex >AlOH···Ox2− in the D

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Redington and Redington45), and 1.000 (from Maçôas et al.44) for values calculated at the GGA/PBE level. Therefore, the values obtained with GGA/PBE are closer to experimental data than those obtained at the MP2/6-31+G** level. This fact validates the extension of the GGA/PBE calculations to the adsorption complexes of periodical models (Table 2). For monosodium oxalate, linear relationships between experimental47 and calculated frequency values of the main vibration modes were found, with correlation coefficient R > 0.998. The scale factors extracted from these correlations, 0.965 (at MP2/6-31+G** level) and 0.974 (at GGA/PBE level), are lower than in oxalic acid, owing to the experimental data come from aqueous dissolutions. In disodium oxalate, the comparison between theoretical and experimental48 frequencies yields also a linear relationship, with a scale factor of 0.977 (R = 0.9988) at the GGA/PBE level and 0.966 (R = 0.9916) at the MP2/6-31+G** level. Again, the frequency values obtained with the periodical GGA/PBE method in our crystalline models fit better for our purposes. The assignments of the main vibration normal modes were similar to those proposed from experimental studies (Table 2). However, we detected some clear differences: the bands at 1337, and 1314 cm−1 were assigned by Frost48 to ν(C−C) and δ(OCO) modes with several multiple bands, possibly due to the presence of several structures, whereas we assign them to the symmetrical ν(CO) mode. 3.2.2. Adsorption Complexes. The neutral surface was optimized at constant volume (surf) showing several kinds of OH groups on the surface: a AlOHAl group maintaining the existing bulk structure, two AlOH groups, two SiOHAl groups, and four SiOH groups (Figure 3a). The AlOH groups are stabilized by internal H bondings, d(AlOH···O(H)Al) = 1.888 Å and d(SiOH···O(H)Al) = 1.589, 1.645 Å. A protonated surface model was also prepared by adding one H atom to a AlOH group, forming a Al−OH2 group (Figure 3d), which is stabilized with stronger H bonds than in the neutral surface, d(AlO(H)H···O(H)Al) = 1.488 Å, d(AlO(H)H···O(H)Si) = 1.685 Å, and d(SiOH···O(H)Al) = 1.748 Å. To compare calculated and experimental frequencies of the main vibration modes for these surfaces and adsorption complexes, we applied the same scale factor as that with disodium oxalate (0.977). The calculated stretching ν(OH) mode of the AlOHAl group in the bulk, far from the surface effect and the edges with dangling bonds, appears at 3649 cm−1, which is close to the experimental

Figure 5. Optimized (at GGA/PBE level) molecular structures of the oxalic acid (a), monosodium oxalate (b), and disodium oxalate (c). The H, O, C, and Na are represented in cyan, red, gray, and purple. Distances are in angstroms.

Å (GGA/PBE) and 2.23 Å (MP2/6-31+G**) and shows a twisted conformer where both carboxylate groups form 90° to each other, according with previous studies43 (Figure 5). In order to validate the calculations at the GGA/PBE level, we compare the vibration frequencies calculated at this level with those obtained at the MP2/6-31+G** level for oxalic acid. For both calculations levels, the main vibration modes were easily assigned with slight differences. To compare with experimental data, we choose those spectroscopic data related to gas phase, high dilution, and completely dry samples owing to the high tendency of oxalic acid to be associated as dimer and with water molecules.44−46 In all cases we found a completely linear relationship between experimental and theoretical frequencies, with correlation coefficient R > 0.999. The slopes of these relationships give us the scale factors to apply to the theoretical values to compare with experimental ones, being 0.964 (with experimental data from the work of Stace et al.46), 0.965 (from Redington and Redington45), and 0.951 (from Maçôas et al.44) for values calculated at the MP2/ 6-31+G** level and 1.003 (from Stace et al.46), 1.004 (from

Table 2. Calculated (at the GGA/PBE level) and Experimental Frequencies (in cm−1) of the Main Vibration Modes of Oxalic Acid Derivatives oxalic

sodium monooxalate

mode

calcd

expl

calcd

ν(OH) ν(CO)as ν(CO)s δ(O−H)s δ(O−H)as ν(C−O) γ(OH) δ(OCO) γ(CO)

3392, 3350a 1850 1796 1487 1396 1218 853 661 457

3461,b 3470,c 3472d 1812,b,c 1817d 1800c,d 1423c,d 1321,b 1325,c 1329d 1267,b 1275,c 1273d − 657,b 660,c 651d 459,b 460c,d

3202 1778f 1679g 1442 1357 1124 − 862,a 700 807,h 473

sodium oxalate

exple

calcd

expli

1725 1623

1634 1419,j 1338

1632 1416,j,k 1337k

1394 1245

1314k 911,a 764 783

818, 766

a

In the symmetric mode. bIn the gas phase.44 cIn the gas phase.46 dIn the gas phase, from Redington and Redington.45 eIn aqueous solution.47 fFrom the carboxylic group. gCarboxylate anion. hγ(OCCO) mode. iFrom the solid state.48 jν(OCCO) mode. kNo clear assignment was reported, with multiple bands attributed to the existence of several structures and coupling with δ(OCO) mode. E

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Table 3. Relative Energy (in eV) and Calculated (at the GGA/PBE level) Frequencies (in cm−1) of the Main Vibration Modes of the Adsorption Complexes complex relative energy ν(OH) (SiOH surface) ν(OH) (AlOHAl surface) ν(OH) (SiOHAl surface) ν(OH) (AlOH) ν(OH)_H2OOxa ν(OH) (COH) δ(OH) water δ(OH) Al(OH2) ν(CO) ν(CC) + skel ν(C−O) δ(OCO) + ν(C−O) δ(OH) (AlOHAl) δ(OH) CO2H δ(OH) SiOH γ(OH) CO2H γ(OH) (AlOHAl) δ(OCO)

H-oxa2 4.098 3673, 3423a

oxamono 0.617 3674, 3459, 3034

oxadiAl 3.646 3138, 2938, 2894b 3697, 3571,c 3501d 3629, 3613

3641, 3602,c 3573c 3352

3638, 3577,c 3533,c 3517d 3362, 2181b

3560, 3214e

3638 3205b



1687

1710

1696, 1673

1692, 1601i 1468, 1402

1410, 1245, 1222 1178,b 921,c 878c 1026

1398h 1253, 1149

3610, 3586, 3335,b 3152b

surf − 3707, 3555, 3205a 3703, 3649,c 3505d 3611, 3271

3689, 3594,c 3557,c 3547d −

3632, 3607,c 3524d



3160,e,f 2134e,f

1730 1652 1674, 1662 1375, 1421 1323i

1339

871c

1064

dioxadi 3.849 3079, 2911

3562,e 3286e,f 3270b 2489g 1662 1634 1707, 1688 1396 1343,h1235j

2613g

1710, 1641 1486

dioxamono 0.0 3625, 3521, 3497

1046, 1016

894,d 844, 838

909,c 873,d 854

1243

1143, 1042,b 1026,b1012,b 1006, 1002

986 1241e

939 880

896

873

a

Forming a H bond with AlOH. bForming a H bond with one carboxylate group. cAlOHAl bulk. dThe same AlOHAl group of the surface but the OH group oriented toward the bulk. eFrom the AlOH2 surface group. fWith a H bond with a SiOH group. gWith a strong H bond with an O(H)Al group. hCOAl group coupled with δ(OH) of SiOH. iJoined to Al. jCO2H group.

value of pyrophyllite (3645 cm−1). Similar values are found in both pristine (neutral and protonated) surfaces. Therefore, this validates our approach to calculate frequencies on the main vibration modes in these systems. The AlOHAl groups localized close to the surface show lower frequencies (3566, 3618 cm−1). The SiOH groups of the surface appear at different frequencies, depending on the local interactions. The SiOH groups with minimal interactions appear at high frequency, at 3687 and 3707 cm−1 for neutral and protonated surfaces, respectively, according to experimental data of silanol groups in hydrated silica.49 The other SiOH groups appear at lower frequency due to the H bonding interactions with other groups, at 3507 and 3555 cm−1 for neutral and protonated surface, respectively. The AlOH groups of the neutral surface appear at 3582 and 3458 cm−1 (OH group forming a H bond with the vicinal AlOH group). In the protonated surface, the ν(OH) of Al(OH2) group appears at low frequency (3160 and 2134 cm−1) due to the strong H bonds with the vicinal SiOH groups (Table 3). Several models of adsorption complexes have been prepared on the basis of the protonated surface that is more likely to exist under the experimental conditions. The anion monosodium oxalate was approached with the carboxylate anion group oriented to the surface and the NaO2C moiety oriented outward to the surface forming the H-oxa2 complex, where the carboxylate group is at d(CO···H2OAl) = 2.2−2.4 Å in a perpendicular orientation with respect to the surface. After optimization, a H transfer from the AlOH2 group to the carboxylate is produced with d(CO−H) = 1.034 Å and a strong H bond with the AlOH group, d(COH···O(H)Al) = 1.501 Å. The carbonyl group of this new carboxylic moiety forms two H bonds with two HOAl groups with d(CO···HOAl) = 2.149 and

2.302 Å. One SiOH group forms another H bond with a vicinal HOAl group, d(SiOH···O(H)Al) = 1.832 Å. In this complex, the external sodium carboxylate moiety maintains the noncoplanarity with the carboxylic group with a dihedral angle O C−CO close to 90° and a d(Na···OC) = 2.14 Å (Figure 6a). This monosodium oxalate can be adsorbed also by forming a bond directly with an octahedral Al cation in the complex oxamono, where the oxalate molecule is oriented partially perpendicularly to the surface. During the optimization an intermediate was obtained, where a bidentate−mononuclear complex was formed following the nomenclature of IUPAC for coordination bonding structures.50 The oxalate molecule was adsorbed with both carboxylate groups joined to one Al atom. However, this structure was not stable, and a further optimization yielded a monodentate−mononuclear adsorption complex where one O atom of one carboxylate group is joined to the Al atom that had the OH2 ligand. This Al atom loses the water ligand, which goes to coordinate the Na cation. The Al− OC bond is very strong, with a bond length of 1.819 Å, much shorter than in the bulk, d(Al−O) = 1.912−1.919 Å. The other O atom of the carboxylate group joined to surface is stabilized with H bonds with OH groups of the surface, d(CO···HOSi) = 1.728 Å, d(CO···HOAl) = 2.546 Å. The other carboxylate group is coplanar with the surface-joined one and forms H bonds with the SiOHAl groups, d(CO···HOSi) = 1.483 Å, and AlOHAl groups, d(CO···HOAl) = 1.660 Å, of the surface and another H bond with the water molecule, d(CO···HOH) = 1.695 Å. The Na cation is coordinated with the SiOH groups of the surface, d(Na···O(H)Si) = 2.291 and 2.352 Å, and with the water molecule, d(Na···OH2) = 2.155 Å (Figure 6b). Another possible adsorption complex model can be with the oxalate molecule in a perpendicular orientation with respect to F

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 6. Optimized structure of the adsorption complexes of oxalate on the protonated (100) surface of a dioctahedral 2:1 phyllosilicate, H-oxa2 (a), oxamono (b), dioxamono (c), oxadiAl (d), and dioxadi (e). The atoms involved in the adsorption interactions are highlighted as balls. The H, O, C, and Na are represented in cyan, red, gray, and purple.

with Na cation, d(CO···Na) = 2.31, 2.32 Å. The coordination sphere of Na cation is completed with one water molecule coming from the Al−OH2 group, d(Na···OH2) = 2.14 Å, and one OH anion coming from the surface Al−OH group, d(Na··· OH) = 2.06 Å. This anion and the water molecule are strongly joined by a strong H bond, d(HO···HOH) = 1.34 Å, and the water molecule is highly asymmetric, d(O−H) = 0.96 and 1.13 Å. Finally, another possible adsorption complex can be with each carboxylate group joined with one O atom to different Al cations, i.e., dioxadi complex (Figure 6e), forming strong CO− Al bonds. The other O atoms of adsorbate form H bonds with the surface SiOH groups, d(CO···HOSi) = 1.62, 1.77 Å. These last carboxylate O atoms coordinate the Na cation, d(Na···OC) = 2.23, 2.24 Å. The coordination sphere of Na has a tetrahedral form with the O atoms of two carboxylate groups, one water molecule [d(Na···OH2) = 2.13 Å], and the hydroxy anion [d(Na···OH) = 2.04 Å]. As above, the OH group is joined to the water molecule by means of a strong H bond, d(HO··· HOH) = 1.34 Å. Calculating the adsorption energies of these complexes (Table 3), we observe that the most stable is the dioxamono complex. The oxamono complex is slightly less stable, 0.62 eV, and the rest form a group with a similar energy level at 3.6 eV, more energetic than the most stable one.

the surface with one O atom of each carboxylate group oriented to the surface but to different Al atoms, d(CO···Al) = 2.30− 2.42 Å, i.e., the dioxamono complex. However, this structure was not stable, and a further optimization yielded a monodentate−mononuclear adsorption complex with a strong Al−OC bond, d(CO−Al) = 1.88 Å, with the Al atom that had the initial OH2 ligand. The AlOHSi group of the Al joined with oxalate donates the H atom to the other carboxylate group, forming a carboxylic acid and maintaining a strong H bond, d(COH···OSi) = 1.47 Å (Figure 6c). The other SiOHAl H atom jumps to the AlOH group forming the AlOH2 group that is stabilized with the vicinal SiOH group by means of the H bond, d(AlOH2···O(H)Si) = 1.19 Å. The water molecule is coordinated with the O atom of oxalate not oriented to the surface, d(CO···HOH) = 1.71 Å. The oxalate molecule is twisted with a dihedral OCCO angle of 37°. The carbonyl O atom of the new carboxylic group is coordinated with Na, d(CO···Na) = 2.38 Å. The water molecule is also coordinated with Na, d(Na···OH2) = 2.17 Å. Another possible adsorption complex can be with one carboxylate group of the oxalate adsorbate coordinated with two Al atoms of the surface, i.e., oxadiAl complex (Figure 6d), forming strong O−Al bonds, d(CO−Al) = 1.88 Å. Another carboxylate group forms H bonds with the surface SiOH groups, d(CO···OHSi) = 1.56, 1.60 Å, and it is coordinated G

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir For the frequency analysis of the adsorption complexes of oxalate, we used the same scale factor as for the disodium oxalate salt (0.977) (Table 3). The stretching ν(OH) modes of the OH groups of the surface, SiOH, AlOH, AlOHAl, and SiOHAl, appear in the range of 3707−2894 cm−1. These frequency values depend on the H bonding and electrostatic interactions of these H atoms with O atoms of vicinal OH or carboxylate groups, where stronger interactions yield lower ν(OH) frequencies. When the H bonds are very strong, this mode appears at very low frequency, 2134 (AlO(H)−H··· O(H)Si) and 2181 (Si(Al)O−H···OC) cm−1. In the cases where the carboxylic acid is formed, H-oxa2 and dioxamono, the H atom of the COH group forms strong H bonds with the O(H)Al group, and its ν(OH) mode appears at low frequency. The stretching ν(CO) mode appears in the range of 1710− 1601 cm−1, and its bands can be overlapped with those from the δ(OH) mode of the water molecule or AlOH2 group. The ν(CO) mode of those adsorption complexes with a carboxylic acid group, which has no significant nonbonding interactions, appears at high frequency, 1710 and 1707 cm−1 in H-oxa2 and dioxamono, respectively. In some adsorption complexes, the carbonyl groups [d(CO) = 1.23−1.24 Å] in oxamono are clearly distinguished from the C−O bonds [d(C− O) = 1.30−1.31 Å] in oxamono. Hence, the ν(CO) mode of these carbonyl groups appears at 1696−1673 cm−1, whereas the ν(C−O) mode of the C−O group joined to the Al atom appears from 1398 (oxamono) to 1343 (dioxamono) cm−1. On the contrary, in another adsorption complex, oxadiAl, the C−O bonds of the carboxylate groups are symmetrically conjugated, d(CO) = 1.263 Å for those joined to H atoms and d(CO) = 1.279 Å for those joined with Al atoms, and only ν(CO) modes are detected, where those joined to surface Al show lower frequency.

Figure 7. DRIFT spectra of oxalate adsorbed to montmorillonite at pH 2.2, 5.2, 7.2, 8.6, and 11.0 (modified from Ramos et al.1).

For the broad band at 1640−1570 cm−1 with a maximum at 1617 cm−1, we can assign tentatively the δ(OH) mode of the AlOH2 group with a calculated frequency of 1634 cm−1 in the dioxamono complex. In this zone, we can assign the ν(CO) band of the bidentate−mononuclear adsorption complex oxadiAl, the calculated value of which appears at 1601 cm−1 for the CO joined directly to the Al atom. This band can appear along with the ν(CO) band of the other carboxylate group, the calculated value of which is 1692 cm−1 and could be in the right shoulder of the main band. The shoulder that appears on the left side of the 1702 and 1617 cm−1 bands could be assigned to the presence of a certain outer-sphere adsorption complex, H-oxa2, the calculated ν(CO) bands of which appear at 1710 and 1641 cm−1. The bidentate−binuclear adsorption complex is very unlikely to be present at low pH, owing to the absence of bands in the 1660−1674 zone, where the calculated ν(CO) bands of the dioxadi complex appear at 1662 and 1674 cm−1. Nevertheless, this complex can be present at high pH, where a broad band appears in this zone. Different models of these systems with several protonation grades and the presence of several water molecules could be explored. Although this study is out of the scope of this work, our calculations can open the possibility of further studies to explore a wide range of experimental conditions. Our modeling is focused mainly on the experimental conditions that, as showed in Figures 1 and 4, most of the oxalate adsorption occurs at low pH. Besides, at high pH both parts, oxalate and surface, would be negatively charged, disfavoring the adsorption. On the other hand, we have explored the adsorption on edge surfaces with only octahedral Al cations. In montmorillonite, the proportion of isomorphous substitution is low and the probability of finding other octahedral cations on the edge surfaces is very low. Nevertheless, these substitutions could alter the frequencies of the vibration bands of these complexes, but the resolution of the DRIFT spectra is too low to distinguish and interpret these possible small alterations. To explore all possible combinations of cation isomorphous susbtitution on the edge surfaces and in other positions close

4. DISCUSSION From our atomistic calculations, we find that the monodentate and mononuclear adsorption complexes are more stable than the bidentate complexes. Hence, the stabilization does not come from the number or strength of Al−O−C bonds, but it is due to the H bond interactions between the hydroxy groups of the surface and the O carboxylate atoms. This is consistent with our experimental results where the adsorption of oxalate increases at lower pH. A decrease of pH yields a higher protonation of the surface, more OH groups, and more sites for adsorption and H bonding. Previous DR-FTIR studies of mineral surface treated with oxalate1 (Figure 7) showed that the mononuclear adsorption complex was more probable, and this is consistent with the relative energy of these complexes, where the mononuclear ones are more stable (dioxamono and oxamono) (Table 3). In these DR_FTIR spectra an interesting band was observed at 1702 cm−1 that agrees with the calculated ν(CO) frequency of the mononuclear complexes, 1707 and 1696 cm−1 in dioxamono and oxamono complexes, respectively. This band appears along with a broad, complex band at 1440 cm−1 with another at 1420 cm−1 and a shoulder around 1400 cm−1. The relative intensity of these last bands increases upon decreasing the pH, and we can assign them to the ν(C−OAl) mode that appears at 1398 and 1343 cm−1 in the calculated oxamono and dioxamono complexes, respectively. Nevertheless, these assignments are tentative, because in this frequency region several coupled modes of ν(CC), ν(C−O), and δ(OCO) of the skeleton of oxalate molecule appear, overlapping each other. H

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

boehmite (γ-AlOOH) interface: A combined potentiometric and IR spectroscopic study. Langmuir 1998, 14, 3655−3662. (8) Dobson, K. D.; McQuillan, A. J. In situ infrared spectroscopic analysis of the adsorption of aliphatic carboxylic acids to TiO2, ZrO2, Al2O3, and Ta2O5 from aqueous solutions. Spectrochim. Acta, Part A 1999, 55, 1395−1405. (9) Axe, K.; Persson, P. Time-dependent surface speciation of oxalate at the water-boehmite (γ-AlOOH) interface: Implications for dissolution. Geochim. Cosmochim. Acta 2001, 65, 4481−4492. (10) Iuga, C.; Sainz-Diaz, C. I.; Vivier-Bunge, A. On the OH initiated oxidation of C2-C5 aliphatic aldehydes in the presence of mineral aerosols. Geochim. Cosmochim. Acta 2010, 74, 3587−3597. (11) Sainz-Díaz, C. I.; Francisco-Márquez, M.; Vivier-Bunge, A. Molecular structure and spectroscopic properties of polyaromatic heterocycles by first principle calculations: spectroscopic shifts with the adsorption of thiophene on phyllosilicate surface. Theor. Chem. Acc. 2010, 125, 83−95. (12) Sutton, R.; Sposito, G. Molecular simulation of humic substance−Ca-montmorillonite complexes. Geochim. Cosmochim. Acta 2006, 70, 3566−3581. (13) Swadling, J. B.; Coveney, P. V.; Greenwell, H. C. Clay minerals mediate folding and regioselective interactions of RNA: A large-scale atomistic simulation study. J. Am. Chem. Soc. 2010, 132, 13750−13764. (14) Yu, C.-H.; Newton, S. Q.; Norman, M. A.; Miller, D. M.; Schäfer, L.; Teppen, B. J. Molecular dynamics simulations of the adsorption of methylene blue at clay mineral surfaces. Clays Clay Miner. 2000, 48, 665−681. (15) Zysset, M.; Schindler, P. W. The proton promoted dissolution kinetics of K montmorillonite. Geochim. Cosmochim. Acta 1996, 60, 921−931. (16) Baeyens, B.; Bradbury, M. H. A mechanistic description of Ni and Zn sorption on Na-montmorillonite 1. Titration and sorption measurements. J. Contam. Hydrol. 1997, 27, 199−222. (17) Bauer, A.; Berger, G. Kaolinite and smectite dissolution rate in high molar KOH solutions at 35° and 80 °C. Appl. Geochem. 1998, 13, 905−916. (18) Cama, J.; Ganor, J.; Ayora, C.; Lasaga, A. C. Smectite dissolution at 80°C and pH 8.8. Geochim. Cosmochim. Acta 2000, 64, 2701−2717. (19) Huertas, F. J.; Caballero, E.; Jimenez de Cisneros, C.; Huertas, F.; Linares, J. Kinetics of montmorillonite dissolution in granitic solutions. Appl. Geochem. 2001, 16, 397−407. (20) Amram, K.; Ganor, J. The combined effect of pH and temperature on smectite dissolution rate under acidic conditions. Geochim. Cosmochim. Acta 2005, 69, 2535−2546. (21) Metz, V.; Amram, K.; Ganor, J. Stoichiometry of smectite dissolution. Geochim. Cosmochim. Acta 2005, 69, 1755−1772. (22) Rozalén, M. L.; Huertas, F. J.; Brady, P. V.; Cama, J.; GarciaPalma, S.; Linares, J. Experimental study of the effect of pH on the kinetics of montmorillonite dissolution at 25°C. Geochim. Cosmochim. Acta 2008, 72, 4224−4253. (23) Rozalén, M.; Brady, P. V.; Huertas, F. J. Surface chemistry of Kmontmorillonite: ionic strength, temperature dependence and dissolution kinetics. J. Colloid Interface Sci. 2009, 333, 474−484. (24) Rozalén, M.; Huertas, F. J.; Brady, P. V. Experimental study of the effect of pH and temperature on the kinetics of montmorillonite dissolution. Geochim. Cosmochim. Acta 2009, 73, 3752−3766. (25) Ramos, M. E.; Cappelli, C.; Rozalen, M.; Fiore, S.; Huertas, F. J. Effect of lactate, glycine and citrate on the kinetics of montmorillonite dissolution. Am. Mineral. 2011, 96, 768−780. (26) Brunauer, S.; Emmett, P. H.; Teller, E. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 1938, 60, 309−319. (27) Pokrovski, G. S.; Schott, J. Experimental study of the complexation of silicon and germanium with aqueous organic species: Implications for germanium and silicon transport and Ge/Si ratio in natural waters. Geochim. Cosmochim. Acta 1998, 62, 3413−3428. (28) Hayes, K. F.; Redden, G.; Ela, W.; Leckie, J. O. Surface complexation models: An evaluation of model parameter estimation using FITEQL and oxide mineral titration data. J. Colloid Interface Sci. 1991, 142, 448−469.

to the surface is out of the scope of this work. However, our study allows the extension of this exploration in further works. In general, we find a good agreement between experimental and calculated results in the identification of structures of adsorption complexes, where the most probable one is an inner sphere−mononuclear complex. However, the use of the IUPAC nomenclature for our complex, protonated mineral surface is too simple, because in addition to the CO−Al bond, we have to consider also other H-bonding interactions of the adsorbate functional groups and the OH groups of the mineral surface that justify the relative energy of these complexes.

5. CONCLUSIONS The macroscopic behavior of the adsorption of oxalate on montmorillonite was successfully modeled using the TLM. Two geometries were identified for the surface complexes: the first one, the monodentate complex >AlOxH, dominated adsorption below pH 4, and the bidentate complex >AlOx− was predominant at higher pH values. Both of the proposed inner-sphere oxalate species were qualitatively consistent with DR-FTIR spectroscopic data and the calculated atomistically results.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +34 958 552 620. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support was obtained from projects CGL2005-00618 (MEC), CGL2008-01652 (MICINN), CGL2011-22567, and FIS2013-48444-C2-2-P (MINECO) and grants RNM-264, ́ RNM363, and RNM1897 (Junta de Andaluci a), with contribution of FEDER funds. M.E.R. benefited from a FPI grant (MEC). C.E. was granted by the Erasmus Mundus program, FJH, amdg.



REFERENCES

(1) Ramos, M. E.; García-Palma, S.; Rozalen, M.; Johnston, C. T.; Huertas, F. J. Effect of oxalate on the kinetics of montmorillonite dissolution. Chem. Geol. 2014, 363, 283−292. (2) Zutic, V.; Stumm, W. Effect of organic acids and fluoride on the dissolution kinetics of hydrous alumina. Geochim. Cosmochim. Acta 1984, 48, 1493−1503. (3) Furrer, G.; Stumm, W. The coordination chemistry of weathering: I. Dissolution kinetics of Al2O3 and BeO. Geochim. Cosmochim. Acta 1986, 50, 1847−1860. (4) Stumm, W.; Furrer, G. The dissolution of oxides and aluminum silicates; examples of surface-coordination-controlled kinetics. In Aquatic Surface Chemistry; Stumm, W., Ed.; Wiley-Interscience: New York, 1987 pp 197. (5) Chin, P. K. F.; Mills, G. L. Kinetics and mechanism of kaolinite dissolution: effects of organic ligands. Chem. Geol. 1991, 90, 307−317. (6) Nordin, J.; Persson, P.; Laiti, E.; Sjoberg, S. Adsorption of ophthalate at the water-boehmite (γ-AlOOH) interface: Evidence for two coordination modes. Langmuir 1997, 13, 4085−4093. (7) Nordin, J.; Persson, P.; Nordin, A.; Sjoberg, S. Inner-sphere and outer-sphere complexation of a polycarboxylic acid at the waterI

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(44) Maçôas, E. M. S.; Fausto, R.; Pettersson, M.; Ralsalnen, M. Infrared-induced rotamerization of oxalic acid monomer in argon matrix. J. Phys. Chem. A 2000, 104, 6956−6961. (45) Redington, R. L.; Redington, T. E. Infrared matrix-isolation spectra of monomeric oxalic acid. J. Mol. Struct. 1978, 48, 165−176. (46) Stace, B. C.; Oralratmanee, C. Infrared and Raman spectra of gaseous oxalic acid. J. Mol. Struct. 1973, 18, 339−342. (47) Jung, Y. M. Characterization of pH-dependent IR spectra of oxalic acid: Comparison of self-modeling curve resolution analysis with calculation of IR frequencies. Bull. Korean Chem. Soc. 2004, 24, 1410− 1412. (48) Frost, R. L. Raman spectroscopy of natural oxalates. Anal. Chim. Acta 2004, 517, 207−214. (49) McMillan, P. F.; Remmele, R. L., Jr. Hydroxyl sites in SiO, glass: A note on infrared and Raman spectra. Am. Mineral. 1986, 71, 772− 778. (50) Muller, P. Glossary of terms used in physical organic chemistry (IUPAC Recommendations 1994). Pure Appl. Chem. 1994, 66, 1077− 1184.

(29) Filius, J. D.; Hiemstra, T.; Van Riemsdijk, W. H. Adsorption of small weak organic acids on goethite: Modeling of mechanisms. J. Colloid Interface Sci. 1997, 195, 368−380. (30) Puigdomenech, I. HYDRA and MEDUSA chemical equilibrium software. Software and documentation; 2004. (31) Reich, T. J.; Koretsky, C. M. Adsorption of Cr(VI) on γ-alumina in the presence and absence of CO2: Comparison of three surface complexation models. Geochim. Cosmochim. Acta 2011, 75, 7006− 7017. (32) Kulik, D. A.; Aja, S. U.; Sinitsyn, V. A.; Wood, S. A. Acid-base surface chemistry and sorption of some lanthanides on K+-saturated Marblehead Illite: II. A multisite-surface complexation modeling. Geochim. Cosmochim. Acta 2000, 64, 195−213. (33) Johnson, S. B.; Yoon, T. H.; Slowey, A. J.; Brown, G. E., Jr Adsorption of organic matter at mineral/water interfaces: 3. Implications of surface dissolution for adsorption of oxalate. Langmuir 2004, 20, 11480−11492. (34) Herbelin, A. L.; Westall, J. C. FITEQL 4.0: A Computer Program for Determination of Chemical Equilibrium Constants from Experimental Data. Report 99-01; Department of Chemistry, Oregon State University: Corvallis, OR. 1999. (35) Westall, J. C.; Herbelin, A. Addendum to FITEQL 2.0 Manual; Oregon State University: Corvallis, OR, 1992. (36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A. Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.: Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Danneberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2004. (37) Soler, J. M.; Artacho, E.; Gale, J. D.; García, A.; Junquera, J.; Ordejón, P.; Sánchez-Portal, D. The SIESTA method for ab-initio order-N materials simulation. J. Phys.: Condens. Matter 2002, 14, 2745−2779. (38) Ortega-Castro, J.; Hernández-Haro, N.; Hernández-Laguna, A.; Sainz-Díaz, C. I. DFT calculation of crystallographic properties of low charge dioctahedral 2:1 phyllosilicates. Clay Miner. 2008, 43, 351−361. (39) Sainz-Díaz, C. I.; Escamilla-Roa, E.; Hernández-Laguna, A. Quantum mechanical calculations of trans-vacant and cis-vacant polymorphism in dioctahedral 2:1 phyllosilicates. Am. Mineral. 2005, 90, 1827−1834. (40) Tsipursky, S. I.; Drits, V. A. The distribution of octahedral cations in the 2:1 layers of dioctahedral smectites studied by obliquetexture electron diffraction. Clay Miner. 1984, 19, 177−193. (41) Sainz-Díaz, C. I.; Palin, E. J.; Dove, M. T.; Hernández-Laguna, A. Monte Carlo simulations of ordering of Al, Fe, and Mg cations in the octahedral sheet of smectites and illites. Am. Mineral. 2003, 88, 1033−1045. (42) Yoon, T. H.; Johnson, S. B.; Musgrave, C. B.; Brown, G. E. Adsorption of organic matter at mineral/water interfaces: I. ATRFTIR spectroscopic and quantum chemical study of oxalate adsorbed at boehmite/water and corundum/water interfaces. Geochim. Cosmochim. Acta 2004, 68, 4505−4518. (43) Chen, C.; Shyu, S. Conformers and intramolecular hydrogen bonding of the oxalic acid monomer and its anions. Int. J. Quantum Chem. 2000, 76, 541−551. J

DOI: 10.1021/acs.langmuir.5b02776 Langmuir XXXX, XXX, XXX−XXX