X-ray Photoelectron Spectroscopy of Fast-Frozen Hematite Colloids in

24 Jan 2013 - Abstract Image ... Measurements were carried out on frozen, centrifuged wet hematite pastes that were previously equilibrated in 50 mM ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/Langmuir

X‑ray Photoelectron Spectroscopy of Fast-Frozen Hematite Colloids in Aqueous Solutions. 5. Halide Ion (F−, Cl−, Br−, I−) Adsorption Kenichi Shimizu,* Andrey Shchukarev, Philipp A. Kozin, and Jean-François Boily Department of Chemistry, Umeå University, SE-901 87 Umeå, Sweden ABSTRACT: Halide anion (F−, Cl−, Br−, and I−) adsorption and its impact on sodium adsorption at the hematite/water interface were studied by cryogenic X-ray photoelectron spectroscopy (XPS). Measurements were carried out on frozen, centrifuged wet hematite pastes that were previously equilibrated in 50 mM electrolytic solutions in the pH 2−11 range. XPS-derived halide ion surface loadings decreased in the order F− > I− ≈ Cl− > Br−, whereas sodium loadings were in the order Na(F) > Na(I) > Na(Br) > Na(Cl). The greater sodium loadings in NaF and in NaI resulted from larger anion loadings in these systems. Bromide ion had the lowest loading among all halide ions despite having a charge-to-size ratio that is intermediate between those of Cl− and I−. This unexpected result may have arisen from specific properties of the hematite/water interface, such as water structure and electric double layer thickness. Fluoride ion adsorption proceeded via the formation of hydrogen bonds with the surface hydroxo groups (e.g., Fe−OH2···F− or Fe−OH···F−). Surface-bound fluoride ions exert a greater charge-screening effect than the other halide anions, as demonstrated by considerably small zeta potential values. Fe−F bond formation was excluded as a possible interfacial process as the F 1s peak binding energy (684.2 eV) was more comparable to that of NaF (684.6 eV) than FeF3 (685.4 eV). Overall, these findings motivate further refinements of existing thermodynamic adsorption models for predicting the ionic composition of hematite particle surfaces contacted with sodium halide aqueous solutions.

1. INTRODUCTION Hematite (α-Fe2O3) is a common iron oxide in Earth’s upper crust.1 It plays a vital role in biogeochemical processes including bacterial respiration and environmental remediation,2−5 and is also considered to have promising applications in electronics and catalysis.6−9 It is therefore important to understand the intrinsic activity of the hematite/water interface, especially that related to the development of surface charge.10,11 Fundamental processes that affect charge development, such as electrolyte ion loading as well as the structure and dynamics of interfacial waters, are highly relevant in this regard. One important challenge in this area is to separate the mineral/water interface region from the bulk solution. A number of experimental approaches have been used to overcome such limitations, including X-ray spectroscopic techniques,12,13 optical reflectometry,14 infrared spectroscopic techniques,15,16 and electrochemical impedance spectroscopy.17 Cryogenic X-ray photoelectron spectroscopy (XPS) is one of such techniques that has been used for this purpose with the notable ability of recovering surface loadings of weakly binding electrolyte ions.18,19 A unique feature of this approach involves rapid freezing of wet mineral pastes for preserving the majority of interfacial waters in vacuo. We have conducted a series of studies using this method on hematite and have uncovered the pH dependence of alkali metal and chloride ions adsorption,20 particle geometry effects on Na+ and Cl− loadings,21,22 as well as previously undetected sorption properties of ammonium species.23 Our results have also supported results of molecular © 2013 American Chemical Society

models focused on ionic interactions at neutrally charged mineral surfaces.20,24 This Article focuses on the adsorption of the halide ions (fluoride, chloride, bromide, and iodide) at the surfaces of nanosized hematite particles. These anions commonly exist in hydrated forms in natural environments but also play active roles in the dissolution/precipitation of minerals as well as the formation of organo-halides. Moreover, excessive halide concentrations in natural waterways and potable water pose a number of serious health risks.25 Efforts have been made to monitor and regulate halide ions exchange with soils using various mineral-based absorbents including hematite.25−27 It is therefore important to provide a foundation for understanding mineral−electrolyte interactions, and notably based on differences in ionic properties, such as charge-to-size ratios and nucleophilicity. Classical electrical double layer models generally predict that halide ion adsorption will occur below the isoelectric point of a mineral surface. More recent experimental20 and theoretical24,28 studies have however suggested that electrolyte anions and cations can adsorb both below and above the isoelectric point, in part due to structured water layers, providing a framework for storing ions,24,28 such as in the case for the hematite/water interface.29 Although electrolyte ion loadings is expected to Received: October 8, 2012 Revised: January 23, 2013 Published: January 24, 2013 2623

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

follow Hofmeister-like series,30 departures from this concept can arise from a variety of contributions, such as interfacial electric field gradients, hydrogen bonding, and alterations of ionic hydration energies in the interfacial region. In this work, we uncover a correlation between halide and sodium ion loadings achieved both above and below the isoelectric point. We identify the Hofmeister order for these ions and discuss new details on fluoride surface binding mechanisms that have, to the best of our knowledge, not been accounted for previously.

undetectable by XPS in the frozen dialyzed pastes (Figures S2 and S3 of ref 23). 2.2. Cryogenic XPS. Full details of the cryogenic XPS technique were also presented in the previous article of this series23 and, hence, are briefly described here. Dry NaF, NaCl, NaBr, and NaI salts were added to separate aliquots of the dialyzed hematite suspensions to achieve an ionic strength of 50 mM. Aliquots (3−6 mL) of the resulting hematite suspensions (4.32 g/L) were purged with humidified N2(g) overnight in 15 mL polyethylene test tubes. The pH was then adjusted to a desired value (between pH 2 and 11) with NaOH or HX (X = F, Cl, Br, and I). An Inlab Routine-L electrode (Mettler Toledo), which was calibrated daily using standard pH buffer solutions (pH 3, 4, 7, and 9), was used for these batch experiments. The sample was equilibrated until the potential drift of the pH electrode was less than 0.3 mV/min. It was then centrifuged at 4000 rpm for 15 min and decanted to remove the supernatant. The resulting wet paste was immediately mounted on a XPS sample holder and placed under dry N2(g) on the claw of the transfer rod that was precooled at −170 °C. The paste was then flash-frozen in the spectrometer air-lock, and after a 45 s cooling period the pressure was lowered to 10−7 Torr. The frozen paste was then transferred to the precooled manipulator in the analysis chamber in which the pressure was kept less than 10−9 Torr. The sample temperature was maintained at around −155 °C throughout the course of the analyses. Measurements were carried out using Kratos Axis Ultra DLD electron spectrometer (Kratos Analytical, Manchester, U.K.). Survey spectra were first collected from 1100 to 0 eV at pass energy of 160 eV. High-resolution spectra were then collected at pass energy of 20 eV for the Fe 2p, O 1s, and C 1s photoelectron lines as well as for those of the electrolyte ions (Na 1s, F 1s, Cl 2p, Br 3d, I 3d). All spectra were subtracted with a Shirley background and fitted using the CasaXPS software using a linear combination of 70:30 Gauss-Lorentz functions. All peak positions were referred to the 530.0 eV O1s peak of hematite.20 The O 1s region required the subtraction of contributions from oxy-carbons,32,33 as explained in the previous paper of this series.20 Peak binding energies of elements in the samples that do not contain hematite are corrected for the C 1s line of aliphatic carbon at 285.0 eV. 2.3. Zeta Potential Measurements. Zeta-potentials of hematite suspensions were measured using Zen3600 Zetasizer (Malvern) with a 0.75 mL folded capillary cell (DTS1060C) at 25 °C. The original hematite suspension (4.32 g/L) was first diluted to 1−2 g/L and dry electrolyte salts were dissolved to achieve an ionic strength of 50 mM. The mixture was then transferred to an external titration vessel that was placed in a 25 °C water bath. Dissolved CO2 was removed from the suspension at pH 4 under an atmosphere of N2(g) for 30 min. The hematite suspension was continuously stirred with an overhead propeller at all time. Alkalimetric titrations were thereafter carried out with a CO2-free 50 mM NaOH titrant solution. At each pH, the suspension was stabilized for 3 min then it was pumped through the polyethylene feeding tube into the capillary cell using a syringe. The capillary cell was then electrochemically isolated from the pH electrodes to conduct the electrophoteric mobility measurements. Triplicate measurements were collected over a period not exceeding 10 s to avoid adventitious sedimentation. Zeta potentials were thereafter calculated using the Smoluchowski equation.34 Once the measurement was completed, the hematite suspension was placed back in the vessel and titration was continued. 2.4. Potentiometric Titrations. Potentiometric acid−base titrations of hematite suspensions (4.32 g/L) containing 50 mM NaCl, NaBr, and NaI were carried out at 25 °C under a humidified N2(g) atmosphere and continuous stirring using a overhead propeller. As in the zeta potential experiments, suspension pH was lowered to 4 at the beginning of a titration in order to remove dissolved CO2. Then titration was conducted using the acid (10 mM HCl, HBr, or HI in 40 mM of the corresponding sodium salts) and base (CO2-free 50 mM NaOH) titrants. Automated additions of acid/base were carried out with a 665 Dosimat (Metrohm) when the drift of the electromotive force of the pH electrodes was less than 0.6 mV/h (0.01 pH/h). Electromotive force was measured using a pH glass electrode

2. MATERIALS AND METHODS 2.1. Chemicals and Materials. All chemicals were used as received. Sodium iodide, sodium fluoride, hydrofluoric acid, hydrobromic acid, and hydroiodic acid were obtained from Merck (pro analysi, Darmstadt, Germany). Sodium bromide was obtained from KEBO Lab (Huddinge, Sweden), and sodium chloride (analar normapur) and hydrochloric acid (analytical grade) were obtained from VWR (Leuven, Belgium). An aqueous solution of sodium hydroxide was prepared by dissolving NaOH pellets from Eka Chemical AB (pro analysi, Bohus, Sweden) in preboiled ultrapure water, and then standardized against a hydrochloric acid solution of known concentration. Concentrations of all other diluted acids were checked by carrying out a colorimetric titration against the sodium hydroxide standard using phenolphthalein as an indicator. Ultrapure water used throughout this work was filtered through a Milli-Q plus 185 QPAK2 purification pack (lot# FIKA80297). N2 gas (99.99%) was humidified and decarbonated by passing through traps containing ultrapure water and a 0.1 M NaOH solution. Full descriptions of the procedures for the hematite synthesis and characterization were already presented in our previous paper23 and are therefore only briefly described here. Hematite was synthesized by forced hydrolysis by adding a 0.72 M FeCl3 solution dropwise to a vigorously stirred 3 mM HCl solution that was preheated to 100 °C. The resulting mixture was matured at 100 °C for 7 days and then dialyzed until the supernatant resistivity exceeded 106 Ω·cm. The concentration of hematite in the resulting suspension was 4.32 g/L. Powder X-ray diffraction measurement (Figure 1) confirms hematite as the only crystallographic phase in the sample. The

Figure 1. Powder XRD diffractogram of hematite. Peak positions for the synthetic α-Fe2O3 are identified in the figure. Miller index is assigned based on ref 31.

particles are spheroidal in shape with an average diameter of 36 nm, as observed with transmission electron microscopy imaging (Figure S1 of the Supporting Information of ref 23.). N2(g) adsorption/desorption isotherms revealed a BET specific surface area of 39.8 m2/g, and a BHJ pore volume of 0.0104 cm3/g and diameter of 1.077 nm. XPS conducted on the dried hematite power indicated resilient traces of chloride ion from the synthetic procedure. This impurity was however 2624

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

(Metrohm 6.0133.100) coupled with a double junction Ag/AgCl reference electrode (Metrohm 6.0726.100) connected to HP 3478A or 34401A multimeter.

The C 1s, Fe 2p, and O 1s regions do not differ significantly from those presented in previous articles from this series.20−23 The C 1s spectra (not shown) contain peaks originating from various oxy-carbon species (−COH, −COOH, CO32‑) in addition to aliphatic hydrocarbons. Higher levels of these carbon-containing contaminants were observed under alkaline conditions due to the greater affinities of hydrophobic molecules for the ice surface.36 Fe 2p spectra were unaffected by variations of pH, electrolyte type, or the amount of carbon impurities, and were identical to those of Boily and Shchukarev.22 The characteristic dove-tail multiplet of hematite22,23,37 is present in all Fe 2p spectra, suggesting negligible phase transformation to other iron (oxy)hydroxides. The O 1s spectra are also nearly identical to those previously reported in the work by Shchukarev and Boily.38 Figure 3 illustrates the

3. RESULTS AND DISCUSSION 3.1. Survey (Wide) Spectra. Survey spectra were collected to assess the chemical purity of the samples and to estimate the peak binding energies for the elements of interest. Figure 2

Figure 2. Representative survey XPS spectra of colloidal hematite frozen wet pastes. The colloidal suspensions were equilibrated in 50 mM sodium halide solutions under N2(g). Identified elements are as indicated in the figure. Figure 3. Iron-normalized atomic ratios of H2O (△), −OH (◆), and −O2‑ (■) as a function of suspension pH. Open circles show the sum of the values for −OH and −O2‑. Oxygen species from all electrolyte systems are included in this figure without being differentiated.

shows representative survey spectra (0−1100 eV) of the fastfrozen hematite pastes equilibrated in aqueous media containing the four sodium halide salts. No impurities were detected other than carbon-based compounds that are inevitably present when using in vacuo methods of measurement. An electron energy loss feature on the high binding energy side of the O 1s line arose from photoelectrons of subsurface origin.35 This secondary spectral feature is generated because the kinetic energy of the photoexcited electrons of the subsurface elements is partially consumed by the surroundings before leaving the sample.35 Its appearance in the spectra of cryogenic samples may, therefore, indicate the occurrence of an additional chemical process and/or the formation of threedimensional arrays at the interface, as reported by Shchukarev et al.21 This feature is not observed for the electrolyte ions of this study, suggesting that they do not accumulate in multiple layers and that ions are not precipitated at the mineral surface. Some elements are more prominent in Figure 2 than others because of the higher surface concentrations and/or higher atomic sensitivity factors (asf). An example of the latter case is the I 3d line (asf = 10.343, Casa XPS), which appears more intense than the F 1s line (asf = 1.000) even though its atomic concentration is about a half of that of fluoride ion. As for the former case, the Na 1s photoemission line (asf = 1.685) at around 1068 eV is not visible for the NaCl and NaBr systems due to their lower surface loadings. 3.2. High-Resolution (Regional) Spectra. High-resolution spectra for individual elements were obtained immediately after the collection of each survey scan. These were used to identify chemical speciation of elements and to provide more accurate readings of signal intensities, making them necessary for quantification.

iron-normalized atomic ratios of oxygen species as a function of the suspension pH. The average value for the (O + OH)/Fe ratio, that is, 1.42 ± 0.07, is lower than the empirical O/Fe ratio of 1.50, consistent with our previous results pointing to surface aquo groups20,22,23,39 as well as possibly nonstoichiometric H2O groups of the (hydro)hematite bulk. The deviation in the (O + OH)/Fe ratio displayed in Figure 3 is 5% and is taken as the standard error for this experiment. Figure 3 also shows that the water loadings at pH 10 were, on average, about twice as much as those below pH 5. This can be interpreted as a pH dependent decrease in the surface charge density. Furthermore, it indicates a thickening of the interfacial layers under basic conditions, especially near the isoelectric point of the mineral (pH 9.5) where particle aggregation is favored. Conversely, the higher levels of interparticle electrostatic repulsion observed under acidic conditions favor the formation of well-dispersed suspensions.40 This colloidal behavior was confirmed by visual inspection of the centrifuged hematite suspensions immediately prior to XPS analysis. Sediments formed at acidic pH were compact, and the supernatant contains residues of dispersed particles. In contrast, sediments formed under alkaline conditions were more voluminous. An accurate quantification of interfacial water remains a challenging task that will necessitate further technical improvements to minimize sample-loading errors such as variation in the surface roughness of the frozen paste. The spectra of the F 1s, Na 1s, Cl 2p, Br 3d, and I 3d photoelectron lines shown in Figure 4 (see Figure 3 of ref 20 2625

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

Figure 4. Representative high-resolution XPS spectra of (a) F−, (b) Cl−, (c) Br−, and (d) I− in the fast-frozen hematite wet pastes. The corresponding peak binding energies are summarized in Table 1.

sense, be partially explained by its high charge-to-size ratio. We also note that fluoride is the only halide ion with kosmotropic properties30 and has the most favorable hydration energy.42 The behavior of the remaining three ions cannot, however, be simply explained on the basis of size-to-charge ratio given their similar hydrated ionic radii. Halide sorption edges are thus indicating additional contributions to the stabilization of these ions beyond hydration energetic contributions alone. For these reasons, the adsorptive properties of F− and of the three other halide ions will be discussed separately in the following two sections. 3.3.1. Fluoride Ion Adsorption. Analysis of the F 1s line (Figure 4a and Table 1) confirms that only one type of fluoride species is present at the hematite/water interface. Zeta potential measurements (Figure 6) underscore the unique pH dependence of the NaF system relative to the other halide-bearing systems. The lower zeta potentials below pH 8 do not correlate with the steep adsorption edge (Figure 5), suggesting that fluoride ion is an effective charge screener. Moreover, the shift in the isoelectric point of hematite could potentially be taken as evidence for the specific adsorption of electrolyte ion;43,44 that is, direct Fe−F coordination may have formed via ligand exchange with surface hydroxo and aquo groups. Such complexes have notably been invoked to explain fluoride adsorption at the goethite/water interface,45 although they have not previously been detected by spectroscopy. In an effort to test this idea further, a number of additional XPS analyses were performed. Our findings are presented along with some relevant literature data46,47 in Table 2. Fast-frozen droplets of NaF and HF solutions (both without hematite) were first investigated to determine the intrinsic

for Na 1s) are all representative of the singly charged ions. The lack of the secondary spectral feature for electron energy loss described in the previous section moreover ensured that the peak binding energies of Table 1 are not from electrolyte salts precipitated on the hematite surface. These results will now be discussed in further detail in the following sections. 3.3. Electrolyte Ion Adsorption. The results of the 31 adsorption experiments carried out for this work are presented in Figure 5. In this and other related20−23 studies, spectral baselines were affected by variations in the paste surface geometry, the thickness of the hydrocarbon contaminants, and the water content of the sample. Baseline variations were minimized by normalizing all atomic abundances by those of Fe, as seen in Figure 5, rather than being presented as absolute values. The pH dependence of the atomic ratios in the figure thus illustrates that ion adsorption was driven by the surface charge of hematite prior to freezing. Moreover, the results indicate that both anions and cations are present across the entire pH range, a phenomenon also observed for alkali-metal chloride salts.20 The halide adsorption edges shown in Figure 5 follow the order F− > I− ≈ Cl− > Br−, while an expected order according to the direct Hofmeister series would be F− > Cl− > Br− > I−.30 The Hofmeister series would only be expected if the ion loadings were controlled by ionic hydration energies alone.39 Our result could thus point to additional or different interfacial processes that are particularly important for Cl−, Br−, and I− adsorption. Our results are nonetheless in line with surface complexation modeling results39 and with the ions’ hydrated radii (F−, 3.58 Å; Cl−, 3.31 Å; Br−, 3.30 Å; I−, 3.31 Å).41 The considerably higher loadings observed for F− can, in this latter 2626

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

Table 1. Peak Binding Energies (eV) of Photoelectron Lines of Elements at the Fast-Frozen Hematite/Water Interface at Various pH Valuesa pH

Fe 2p3/2 1

Fe 2p3/2 2

O 1s O2‑

O 1s OH

O 1s H2O

C 1sb

Na 1s

F 1s

4.93 5.38 5.58 5.58c 6.29 7.94 9.67 10.45 pH

709.6 709.8 709.8 709.8 709.8 709.7 709.7 709.8 Fe 2p3/2 1

710.8 711.1 711.1 711.0 711.1 710.9 710.9 711.0 Fe 2p3/2 2

530.0 (1.1) 530.0 (1.1) 530.0 (1.1) 530.0 (0.9) 530.0 (0.9) 530.0 (0.9) 530.0 (1.1) 530.0 (0.9) O 1s O2‑

531.1 (1.6) 531.3 (1.6) 531.2 (1.6) 531.5 (1.1) 531.3 (1.5) 531.1 (1.6) 531.2 (1.6) 531.2 (1.5) O 1s OH

532.6 (1.8) 533.1 (1.9) 533.3 (1.9) N/D 533.3 (1.9) 533.0 (1.6) 533.4 (1.9) 533.4 (1.9) O 1s H2O

285.3 285.2 285.0 284.9 285.1 285.5 285.3 285.3 C 1sb

1070.5 (1.3) 1071.2 (1.4) 1071.3 (1.4) 1071.9 (1.7) 1071.4 (1.4) 1071.1 (1.6) 1070.9 (1.6) 1071.5 (1.4) Na 1s

684.0 (1.6) 684.2 (1.5) 684.2 (0.9) 684.2 (1.4) 684.3 (1.4) 684.1 (1.4) 684.1 (1.6) 684.3 (1.5) Cl 2p3/2

2.46 4.01 6.21 7.92 9.96 10.69 pH

710.1 709.8 709.8 709.7 709.8 709.4 Fe 2p3/2 1

711.4 711.0 711.1 711.1 711.1 710.7 Fe 2p3/2 2

530.0 (0.9) 530.0 (0.9) 530.0 (1.0) 530.0 (1.0) 530.0 (1.0) 530.0 (0.9) O 1s O2‑

531.1 (1.2) 531.1 (1.5) 531.4 (1.5) 531.3 (1.3) 531.3 (1.2) 531.1 (1.2) O 1s OH

533.1 (1.6) 533.3 (1.6) 533.4 (1.7) 533.3 (1.8) 533.3 (1.8) 533.1 (1.9) O 1s H2O

284.8 285.1 285.1 285.1 285.0 285.4 C 1sb

1071.3 (1.4) 1071.3 (1.2) 1071.5 (1.3) 1071.4 (1.3) 1071.5 (1.4) 1071.5 (1.4) Na 1s

198.1 (1.1) 198.2 (1.1) 198.4 (1.2) 198.4 (1.2) 198.3 (1.2) 198.4 (1.2) Br 3d5/2

2.72 5.61 8.06 9.19 10.03 10.32 pH

709.8 709.8 709.8 709.8 709.7 709.8 Fe 2p3/2 1

711.0 711.0 711.0 711.1 711.0 711.0 Fe 2p3/2 2

530.0 (1.0) 530.0 (1.0) 530.0 (1.0) 530.0 (1.0) 530.0 (1.0) 530.0 (1.0) O 1s O2‑

531.1 (1.5) 531.3 (1.4) 531.4 (1.6) 531.3 (1.3) 530.9 (1.4) 531.3 (1.6) O 1s OH

533.0 (1.7) 533.3 (1.6) 533.5 (1.9) 533.4 (1.8) 533.5 (1.9) 533.5 (1.9) O 1s H2O

284.8 284.8 285.1 285.1 285.3 285.3 C 1sb

2.43 3.20 3.82 5.50 8.48 10.49

709.8 709.8 709.8 709.8 709.8 709.8

711.0 711.0 711.0 711.0 711.0 711.0

530.0 530.0 530.0 530.0 530.0 530.0

531.2 531.2 531.2 531.2 531.3 531.3

533.2 533.0 533.3 533.3 533.2 533.8

284.7 285.0 285.1 284.9 285.1 285.3

(0.9) (0.9) (0.9) (1.0) (0.9) (1.0)

(1.4) (1.4) (1.4) (1.6) (1.4) (1.6)

(1.5) (1.5) (1.6) (1.7) (1.9) (1.9)

1071.3 (1.4) 1071.4 (1.5) 1071.4 (1.4) 1071.3 (1.5) 1071.4 (1.4) 1071.5 (1.5) Na 1s 1071.3 1071.2 1071.4 1071.4 1071.4 1071.5

(1.3) (1.3) (1.3) (1.5) (1.5) (1.4)

68.5 (1.1) 68.5 (1.0) 68.6 (1.0) 68.8 (1.0) 69.0 (1.3) 68.7 (1.0) I 3d5/2 619.2 619.1 619.2 619.2 619.2 619.3

(1.2) (1.3) (1.2) (1.6) (1.4) (1.4)

a

All listed values are referenced to the O 1s of the surface oxide at 530.0 eV. The full width at half maximum for each peak is shown in parentheses. N/D: Not detected. bThe C 1s line is of aliphatic carbon. cDried at room temperature in vacuo.

Figure 6. Zeta potentials of hematite suspensions in 50 mM of NaF (▽), NaCl (□), NaBr (○), and NaI (△) collected at 25 °C.

Figure 5. Iron-normalized atomic ratios for F− (▽), Cl− (□), Br− (○), and I− (△). The atomic ratios of the corresponding sodium ions are indicated by closed symbols. Error bars indicate 5% the standard error for these measurements.

noteworthy to mention that F 1s binding energies in dry FeF3 are greater than in NaF solutions (Table 2) due to a more extensive sharing of F− valence shell electrons with the Fe.46 Furthermore, when FeF3 is dissolved in water, the resulting acidic solutions yield F 1s binding energies that become even more comparable to those of HF solutions (Table 2). Therefore, as the binding energies of the F 1s line of the hematite/water interface (Table 1) are in close agreement with those of NaF solutions of circumneutral pH (Table 2) and are independent of pH (684.0−684.3 eV), fluoride is more likely to

binding energies of individual aqueous fluoride species. The F 1s spectrum of the HF solution (Figure 7 and Table 2) reveals the HF and F− species with a peak separation of 1.5 eV. A comparable difference also occurs between F− peaks in HF and NaF (1.4 eV, Table 2). These results therefore highlight a relationship between the peak binding energies of the F 1s line and the degree of interaction of the anion with protons. It is 2627

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

The evidence presented in this work thereby points to a close association a (partially) hydrated F− with the hematite surface, one that is likely to be stabilized through hydrogen bonding with surface hydroxo groups (e.g., Fe−OH2···F−, Fe− OH···F−). While such interactions may be sufficiently strong to exist in an aqueous environment, they are too weak to be detected by XPS. It also explains why thermodynamic adsorption models have invoked inner-sphere complexes to predict large proton/fluoride adsorption ratios.45 We therefore propose that the charge of F− should be considered in the same adsorption plane as those of the potential-determining ions when predicting ionic adsorption, and/or that electric double layer models be better adapted for such configurations. In addition, the corresponding reactions should be reformulated to emphasize the formation of hydrogen-bonded species (e.g., Fe−OH + H+ + F− ⇌ Fe−OH2+···F−). Finally, we emphasize that enhanced sodium ion adsorption below the isoelectric point is an additional charge-neutralizing mechanism that mirrors fluoride loadings achieved in this system. Again, it may be necessary that thermodynamic adsorption models account for enhanced sodium loadings to provide a more complete description of the elemental composition of hematite surfaces contacted with the aqueous solutions of NaF. 3.3.2. Adsorption of Cl−, Br−, I−. Anion loadings of the remaining three halide ions considered in this work decreased in the order of I− ≈ Cl− > Br−, while those for sodium decreased as Na(I) > Na(Br) > Na(Cl). Zeta potentials of the three systems are nearly identical, with isoelectric points in the 9.3−9.7 range. The potentiometric titration data (Figure 8)

Table 2. Peak Binding Energies (eV) of Photoelectron Lines of Various Reference Compoundsa Mz

F−

HF

ref

686.8

1071.4 1071.4 1071.0 715.1c 715.3 714.6

685.3 683.9 684.6 683.9 685.3c 685.4 684.9

this study this study this study ref 46 ref 47 this study this study

b

HF NaFb NaF dry salt NaF dry salt FeF3 FeF3 dry salt FeF3b

686.6

a

Binding energies (eV) are referenced to the C 1s electrons at 285.0 eV. bFast frozen solution. cBinding energies are referred to C 1s at 284.8 eV.

Figure 7. High resolution XPS spectrum of the F− ion in a fast-frozen droplet of a dilute HF solution. Peaks representing different fluoride species are indicated, and the corresponding peak binding energies are listed in Table 2.

exist as a singly charged (partially) hydrated surface-bound species rather than a Fe−F complex. In an effort to test this idea further, water was removed from the samples by warming to room temperature in vacuo. Drying generally shifts peak binding energies of electrolyte ions through the loss of the hydration shell and a greater sharing of electrons. Table 2 shows that F 1s binding energies of NaF and FeF3 solutions are 0.7 and 0.5 eV larger, respectively, than their corresponding dry salts. The collapse of the electrical double layer48,49 is another contributing factor to the shift in binding energies upon drying. Photoelectrons emitted from the hematite bulk or surface experience an electric double layer potential drop (1 eV/V) as they travel through the frozen interfacial water (see Figure 7 of ref 48.), while ions located near the outer edge of the frozen water layer would not be impacted as much. However, because our elemental binding energies reported here are calibrated against the O 1s line of the bulk oxide at 530.0 eV (and not the C 1s 285.0 eV line of hydrocarbons), the effect is actually reversed. Specifically, binding energies of elements at or near the hematite surface are least affected by changes in interfacial potential ensuing water removal. Conversely, those located near the outer edge of the frozen water layer will experience the greatest shift. Thus, in accordance with this concept, our drying experiment did not affect the F 1s region, the 530.0 eV oxide peak, or the Fe 2p line (Table 1), while it shifted the Na 1s region shifted by 0.6 eV (Table 1). This experiment thereby added further confirmation that fluoride is closely associated to the hematite surface.

Figure 8. Potentiometric titration (alkalimetric) of hematite suspensions in 50 mM NaCl (□), NaBr (○), and NaI (△) collected at 25 °C.

however suggest slightly lower proton loadings in NaI. This lower uptake may be correlated to larger sodium loadings and/ or ion pairing. The latter process may be relevant in all systems, but this has yet to be experimentally demonstrated. The lower Br− loadings, in comparison to those of Cl−, appear to be inconsistent with the fact that both ions generate the same surface charge and that sodium loadings in the NaBr system are greater than those in NaCl. To investigate this issue further, we carried out the following additional tests. Reproducibility was first tested by repeated sets of measurements using freshly prepared solutions. XPS-derived atomic ratios were cross-validated using new sets of reagentgrade sodium halide salts. Ionic strength effects were also addressed through additional measurements in 100 mM NaBr. Finally, potential photocatalytic effects were examined by comparing results of samples exposed to light and those kept in the dark. All of these tests however gave rise to identical results to those presented in Table 1 and Figure 4c. Additional 2628

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir



experiments were also conducted with synthetic gibbsite (αAl(OH)3) nanosized platelets to test effects of mineral surface structure on electrolyte ion loadings. These experiments, carried out at pH 5, revealed nearly indistinguishable Cl−, Br− and I− loadings while cation adsorption edges followed the order Na(Br) > Na(I) > Na(Cl). These results (to be communicated separately) therefore confirmed that our approach was sufficiently sensitive for reproducing and distinguishing mineral-specific ionic loadings. We therefore suspect that the unique adsorption behavior for Br− on hematite may indicate specific attributes of the hematite/water interface, such as water structure, double layer thickness, and other related colligative properties. Molecular dynamics simulations and other experimental approaches, including electrochemical impedance spectroscopy, are currently being used to investigate these issues in greater detail. The results presented herein provide compelling evidence for the adsorption of single ionic Cl−, Br−, and I− species at the hematite/water interface. Halide ions are stabilized at the interface via electrostatic interaction and hydrogen bonding with surface hydroxo and aquo groups. The ions must also be present in their hydrated forms, as indicated by their peak binding energies. Ongoing molecular dynamics simulations of this system are in fact adding further support to this claim. The remarkable anion loadings above the pH of the isoelectric point moreover indicate that these ions are stabilized by local regions of positive charge, even when the net electrostatic charge is negative. We therefore suggest that thermodynamic models be developed to predict the coexistence of sodium and halide ions below and above the isoelectric point.

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +46 90 786 5361. Fax: +46 90 136310. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the Swedish Research Council (2009-3110; 2012-2976), as well as the Kempe, Wallenberg, and Carl-Tryggers Foundations.



REFERENCES

(1) Schwertmann, U.; Cornell, R. M. The Iron Oxides; Wiley-VCH: New York, 2003. (2) Vargas, M.; Kashefi, K.; Blunt-Harris, E.; Lovley, D. R. Microbiological evidence for Fe(III) reduction on early Earth. Nature 1998, 395, 65−67. (3) Lovley, D. R.; Phillips, E. J. P. Organic Matter Mineralization with Reduction of Ferric Iron in Anaerobic Sediments. Appl. Environ. Microbiol. 1986, 51, 683−689. (4) Zachara, J. M.; Fredrickson, J. K.; Smith, S. C.; Gassman, P. L. Solubilization of Fe(III) Oxide-Bound Trace Metals by a Dissimilatory Fe(III) Reducin Bacterium. Geochim. Cosmochim. Acta 2001, 65, 75− 93. (5) Wilkings, M. J.; Livens, F. R.; Vaughan, D. J.; Lloyd, J. R. The Impact of Fe(III)-Reducing Bacteria on Uranium Mobility. Biogeochemistry 2006, 78, 125−150. (6) Yanina, S. V.; Rosso, K. M. Linked Reactivity at Mineral-Water Interfaces Through Bulk Crystal Conduction. Science 2008, 320, 218− 222. (7) Wang, G.; Gou, X.; Horvat, J.; Park, J. Facile Synthesis and Characterization of Iron Oxide Semiconductor Nanowires for Gas Sensing Application. J. Phys. Chem. C 2008, 112, 15220−15225. (8) de Souza Ramos, M.; de Santana Santos, M.; Gomes, L. P.; Albornoz, A.; do Carmo Rangel, M. The influence of Dopants on the Catalytic Activity of Hematite in the Ethylbenzene dehydrogenation. Appl. Catal., A 2008, 341, 12−17. (9) de Carvalho, V. A. N.; de S. Luz, R. A.; Lima, B. H.; Crespilho, F. N.; Leite, E. R.; Souza, F. L. Highly Oriented Hematite Nanorods Arrays for Photoelectrochemical Water Splitting. J. Power Sources 2012, 205, 525−529. (10) Bockris, J. O. M.; Khan, S. U. M. Surface Electrochemistry-A Molecular Level Approach; Plenum Press: New York, 1993. (11) Isrealachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1992. (12) Park, C.; Fenter, P. A.; Nagy, K. L.; Sturchio, N. C. Hydration and Distribution of Ions at the Mica-Water Interface. Phys. Rev. Lett. 2006, 97, 016101. (13) Zang, Z.; Fenter, P.; Cheng, L.; Sturchio, N. C.; Bedzyk, M. J.; Machesky, M. L.; Anovitz, L. M.; Wesolowski, D. J. Zn2+ and Sr2+ Adsorption at the TiO2(110)-Electrolyte Interface: Influence of Ionic Strength, Coverage, and Anions. J. Colloid Interface Sci. 2006, 295, 50. (14) Notley, S. M. Adsorption of Nonionic Surfactants with Ethylene Oxide Headgroup Chemistry at the Titania-Water Interface. J. Phys. Chem. B 2012, 116, 6059−6065. (15) Mäkie, P.; Westin, G.; Persson, P.; Ö sterlund, L. Adsroption of Trimethyl Phosphate on Maghemite, Hematite, and Goethite Nanoparticles. J. Phys. Chem. A 2011, 115, 8948−8959. (16) Olsson, R.; Giesler, R.; Loring, J. S.; Persson, P. Adsorption, Desorption, and Surface-Promoted Hydrolysis of Glucose-1-Phosphate in Aqueous Goethite (α-FeOOH) Suspensions. Langmuir 2010, 26, 18760−18770. (17) Shimizu, K.; Lasia, A.; Boily, J.-F. Electrochemical Impedance Study of the Hematite/Water Interface. Langmuir 2012, 28, 7914− 7920.

4. CONCLUSIONS This work describes the codependence of electrolyte ion identity on adsorption at the surfaces of hematite nanoparticles equilibrated in aqueous solutions of sodium halides. Results suggest that elevated sodium loading resulted in lowered proton loadings in NaI compared to NaCl and NaBr. While it was expected that halide adsorption would be primarily determined by ionic size-to-charge ratio and nucleophilicity, Br− proved to be a clear exception. The unexpectedly low loadings for this anion may result from specific colligative properties of the hematite/water interface that are not apparent in analogous experiments using gibbsite in lieu of hematite. Of all halide ions tested, fluoride ion has the greatest loadings. The shift in the isoelectric point of hematite caused by fluoride adsorption resulted from the strong association of this ion with the hematite surface and from its effective charge-screening capability. Our results imply a partial loss of the hydration sheath of fluoride form direct hydrogen bonds with hydroxo and aquo groups of the hematite surface. To properly understand and model the phenomena observed in this work, it will be necessary to refine existing thermodynamic adsorption models predicting ionic loadings at the hematite/water interface. This should be done by developing models that permit the coexistence of electrolyte anions and cations below and above the isoelectric point and by including the charges of the adsorbed fluoride ions near or at the adsorption plane of the potential-determining ions. Such changes could be affected by modifying the models’ electric double layer structures based on experimental and theoretical information regarding the composition and structure of the hematite/water interface. 2629

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630

Langmuir

Article

(18) Shchukarev, A.; Sjöberg, S. XPS with Fast-Frozen Samples: A Renewed Approach to Study the Real Mineral/Solution Interface. Surf. Sci. 2005, 584, 106. (19) Ramstedt, M.; Andersson, B. M.; Shchukarev, A.; Sjöberg, S. Surface Properties of Hydrous Manganite (γ-MnOOH). A Potentiometric, Electroacoustic, and X-ray Photoelectron Spectroscopy Study. Langmuir 2004, 20, 8224. (20) Shimizu, K.; Shchukarev, A.; Kozin, Ph. A.; Boily, J.-F. X-ray Photoelectron Spectroscopy of Fast-Frozen Hematite Colloids in Aqueous Solutions. 4. Coexistence of alkali metal (Na+, K+, Rb+, Cs+) and Chloride Ions. Surf. Sci. 2012, 606, 1005−1009. (21) Shchukarev, A.; Boily, J.-F.; Felmy, A. R. XPS of Fast Frozen Hematite Colloids in NaCl Aqueous Solutions. I. Evidence for the Formation of Multiple Layers of Hydrated Sodium and Chloride Ions Induced by the {001} Basal Plane. J. Phys. Chem. C 2007, 111, 18307− 18316. (22) Boily, J.-F.; Shchukarev, A. X-ray Photoelectron Spectroscopy of Fast-Frozen Hematite Colloids in Aqueous Solutions. 2. Tracing the Relationship between Surface Charge and Electrolyte Adsorption. J. Phys. Chem. C 2010, 114, 2613−2616. (23) Shimizu, K.; Shchukarev, A.; Boily, J.-F. X-ray Photoelectron Spectroscopy of Fast-Frozen Hematite Colloids in Aqueous Solutions. 3. Stabilization of Ammonium Species by Surface (Hydr)oxo Groups. J. Phys. Chem. C 2011, 115, 6796−6801. (24) Wander, M. C.; Shuford, K. L. Alkali Halide Interfacial Behavior in a Sequence of Charged Slit Pores. J. Phys. Chem. C 2011, 115, 23610−23619. (25) Sajidu, S. M. I.; Masamba, W. R. L.; Thole, B.; Mwatseteza, J. F. Groundwater Fluoride Levels in Villages of Southern Malawi and Removal Studeis using Bauxite. Int. J. Phys. Chem. 2008, 3, 001−011. (26) Kaplan, D. I.; Serne, R. J.; Parker, K. E.; Kutnyakov, I. V. Iodide Sorptio to Subsurface Sediments and Illitic Minerals. Environ. Sci. Thech. 2000, 34, 399−405. (27) Teutli-Sequeira, A.; Solache-Ríos, M.; Balderas-Hernández, P. Modification Effects of Hematite with Aluminum Hydroxide on the Removal of Fluoride Ions from Water. Water Air Soil Poll. 2012, 223, 319−327. (28) Kerisit, S.; Ilton, E. S.; Parker, S. C. Molecular Dynamics Simulations of Electrolyte Solutions at the (100) Goethite Surface. J. Phys. Chem. B 2006, 110, 20491−20501. (29) Kerisit, S. Water Structure at Hematite-Water Interfaces. Geochim. Cosmochim. Acta 2011, 75, 2043−2061. (30) dos Santos, A. P.; Diehl, A.; Levin, Y. Surface Tensions, Surface Potentials, and the Hofmeister Series of Electrolyte Solutions. Langmuir 2010, 26, 10778−10783. (31) NuLi, Y.; Zhang, P.; Guo, Z.; Munroe, P.; Liu, H. Preparation of α-Fe2O3 submicro-flowers by a hydrothermal approach and their electrochemical performance in lithium-ion batteries. Electrochim. Acta 2008, 53, 4213−4218. (32) The XPS of Polymer Database; Beamson, G.,Briggs, D., Eds.; Surface Spectra Ltd: Manchester, 2001. (33) Shchukarev, A. V.; Korolkov, D. V. XPS Study of Group IA Carbonates. Cent. Eur. J. Chem. 2004, 2, 347−362. (34) Sze, A.; Erickson, D.; Ren, L.; Li, D. Zeta-Potential Measurement using the Smoluchowski Equation and the Slope of the Current-Time Relationship in Electroosmotic Flow. J. Colloid Interface Sci. 2003, 261, 402−410. (35) Ratner, B. D.; Castner, D. G. Electron Spectroscopy for Chemical Analysis. In Surface Analysis-The Principal Techniques, 2nd ed; Vickerman, J. C., Gilmore, I. S., Eds.; John Wiley & Sons Ltd.: West Sussex, 2009. (36) Creux, P.; Lachaise, J.; Graciaa, A.; Beattie, J. K.; Djerdjev, A. M. Strong Specific Hydroxide Ion Binding at the Pristine Oil/Water and Air/Water Interfaces. J. Chem. Phys. B 2009, 113, 14146−14150. (37) Droubay, T.; Chambers, S. A. Surface-Sensitive Fe 2p Photoemission spectra for α-Fe2O3 (0001): The Influence of Symmetry and Crystal-Field Strength. Phys. Rev. B 2001, 64, 205414. (38) Shchukarev, A.; Boily, J.-F. XPS Study of the Hematite-Aqueous Solution Interface. Surf. Interface Anal. 2008, 40, 349−353.

(39) Sverjensky, D. A. Prediction of Surface Charge on Oxide in Salt Solutions: Revisions for 1:1 (M+L−) Electrolytes. Geochim. Cosmochim. Acta 2005, 69, 225−257. (40) Schudel, M.; Behrens, S. H.; Holthoff, H.; Kretzschmar, R.; Borkovec, M. Absrolute Aggregation Rate Constants of Hematite Particles in Aqueous Suspensions: A Comparison of Two Different Surface Morphologies. J. Colloid Interface Sci. 1997, 196, 241−253. (41) Nightingale, E. R., Jr. Phenomenological Theory of Ion Solvation. Effective Radii of Hydrated Ions. J. Phys. Chem. 1959, 63, 1381−1387. (42) Kalyuzhnyi, Y. V.; Vlachy, V.; Dill, K. A. Hydration of Simple Ions. Effect of the Charge Density. Acta Chim. Slov. 2001, 48, 309− 316. (43) Sędłak, A.; Janusz, W. Specific Adsorption of Carbonate Ions at the Zinc Oxide/Electrolyte Solution Interface. Physicochem. Probl. Miner. Process. 2008, 42, 57−66. (44) Kosmulski, M.; Saneluta, C.; Mączka, E. Electrokinetic Study of Specific Adsorption of Cations on Synthetic Goethite. Colloids Surf., A 2003, 222, 119−124. (45) Hiemstra, T.; Van Riemsdijk, W. H. Fluoride Adsorption on Goethite in Relation to Different Types of Surface Sites. J. Colloid Interface Sci. 2000, 225, 94−104. (46) Morgan, W. E.; Van Wazer, J. R.; Stec, W. J. Inner-Orbital Photoelectron Spectroscopy of the Alkali Metal Halides, Perchlorates, Phosphates, and Pyrophosphates. J. Am. Chem. Soc. 1973, 95, 751− 755. (47) Grosvenor, A. P.; Kobe, B. A.; Biesinger, M. C.; McIntyre, N. S. Investigation of Multiplet Splitting of Fe 2p XPS Spectra and Bonding in Iron Compounds. Surf. Interface Anal. 2004, 36, 1564−1574. (48) Shchukarev, A. XPS at Solid-Aqueous Solution Interface. Adv. Colloid Interface Sci. 2006, 122, 149−157. (49) Shchukarev, A. Electrical Double Layer at the Mineral-Aqueous Solution Interface as Probed by XPS with Fast-Frozen Samples. J. Electron Spectrosc. Relat. Phenom. 2010, 176, 13−17.

2630

dx.doi.org/10.1021/la3039973 | Langmuir 2013, 29, 2623−2630