Second Harmonic Generation Studies of Fe(II) Interactions with

Jan 31, 2013 - Iron oxides are a ubiquitous class of compounds that are involved in many biological, geological, and technological processes, and the ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCC

Second Harmonic Generation Studies of Fe(II) Interactions with Hematite (α-Fe2O3) David S. Jordan,† Christopher J. Hull,† Julianne M. Troiano,† Shannon C. Riha,‡ Alex B. F. Martinson,‡ Kevin M. Rosso,§ and Franz M. Geiger*,† †

Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States Argonne-Northwestern Solar Energy Research (ANSER) Center and Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439, United States § Chemical & Materials Sciences Division, Pacific Northwest National Laboratory, Battelle Boulevard, K1-83, Richland, Washington 99354, United States ‡

S Supporting Information *

ABSTRACT: Iron oxides are a ubiquitous class of compounds that are involved in many biological, geological, and technological processes, and the Fe(III)/Fe(II) redox couple is a fundamental transformation pathway; however, the study of iron oxide surfaces in aqueous solution by powerful spectroscopic techniques has been limited due to “strong absorber problem”. In this work, atomic layer deposition (ALD) thin films of polycrystalline α-Fe2O3 were analyzed using the Eisenthal χ(3) technique, a variant of second harmonic generation that reports on interfacial potentials. By determining the surface charge densities at multiple pH values, the point of zero charge was found to be 5.5 ± 0.3. The interaction of aqueous Fe(II) at pH 4 and in 1 mM NaCl with ALD-prepared hematite was found to be fully reversible and to lead to about 4 times more ferrous iron ions adsorbed per square centimeter than on fused-silica surfaces under the same conditions. The data are consistent with a recently proposed conceptual model for net Fe(II) uptake or release that is underlain by a dynamic equilibrium between Fe(II) adsorbed onto hematite, electron transfer into favorable surface sites with attendant Fe(III) deposition, and electron conduction to favorable remote sites that release and replenish aqueous Fe(II).

I. INTRODUCTION Iron oxides are widespread in nature. In addition to their relevance in the corrosion of iron,1 they play an important role in the environment influencing the transport and transformation of ionic metal pollutants, eutrophication compounds, and xenobiotics.2 Iron oxides are intimately involved in biogeochemical iron cycling,3 a process that is linked to several important nutrient cycles, such as the carbon4 and oxygen5 cycles. Iron cycling in aqueous environments frequently involves the interaction of Fe(II) with Fe(III), and environmental conditions in which these redox species interact typically dictate that Fe(II) is an aqueous ion but Fe(III) occurs as a solid oxide. In addition to the relevance to the corrosion of materials, understanding the Fe(II)aqueous/Fe(III)oxide interfacial system is also critical for understanding soil and groundwater chemistry because it has been known to affect the mobility and speciation of contaminants,6,7 influence microbial activity,8 and fractionate iron isotopes.9 To investigate the interaction of aqueous Fe(II) with Fe(III) oxides at a fundamental level, techniques such as batch chemical extractions,10,11 TEM,12 AFM,13 Mössbauer spectroscopy,14,15 and X-ray reflectivity16 have been used. These studies report that Fe(II) forms stable complexes with ligand binding sites on Fe(III) oxide surfaces17 and that electron and atom exchange between the aqueous Fe(II) species and the Fe(III) atoms of the bulk oxide, which has long been suspected,18,19 © 2013 American Chemical Society

can occur, as recently shown by Scherer and Rosso and coworkers.15,20−22 Despite the importance of this system, and even though many studies regarding hematite/water interfaces exist,14,15,23−54 a surface-specific, in situ (i.e., under aqueous solution) study of Fe(II) adsorption to an Fe(III) oxide has yet to be achieved because of the difficulty in distinguishing small amounts of Fe(II) at the interface of a bulk iron oxide. The issue of surface specificity is an important one, as the interfacial speciation of several heavy metal ions can drastically differ from predictions derived from bulk aqueous properties.55,56 However, the surface-specific interactions between aqueous Fe(II) and solid Fe(III) oxides are very difficult to investigate under aqueous flow conditions. This issue is mainly due to the fact that most spectroscopies that have been applied to study hematite/water interfaces produce signals that are dominated by the response from the species located in the aqueous bulk, requiring careful experimental design and limiting such an approach to a few select scientific questions. Although nonlinear optical methods, such as second harmonic or vibrational sum frequency generation (SHG and SFG), are interface-selective, so far, those approaches have been limited to iron nanoparticles or iron-doped materials57−59 because they Received: November 15, 2012 Revised: January 29, 2013 Published: January 31, 2013 4040

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

also suffer from the strong absorber problem. Simply put, when it comes to iron oxides, the obstacle is that the samples either burn when nano-, pico-, or femtosecond laser pulses at optical frequencies are directed onto iron oxides, or, if the incident light fields are attenuated, insufficient signal is produced to study the system of interest efficiently. We note that zerovalent iron does not pose the strong absorber problem of iron oxides. In fact, Baldelli’s pioneering nonlinear optical work60−64 on Zn, Cu, and steel surfaces has been very successful. Here, we overcome this classical “strong absorber problem” by following the approach reported by Martinson et al.65 to deposit extremely thin films (10 nm thickness) of polycrystalline α-Fe2O3 onto fused-silica substrates using atomic layer deposition (ALD). We show that the optical density of these samples is low enough that their interface with aqueous solutions can be studied using SHG. Demonstrating that ALDprepared and natural hematite behave similarly in terms of point of zero charge and interaction with Fe(II) species would ensure that the results obtained from ALD-prepared hematite have environmental significance in terms of the geochemical science targets (transport, fate, remediation). The results shown below demonstrate that this is, indeed, the case. The aims of this work then are to (1) assess the applicability of the hematite thin films as a model analogue to natural hematite and (2) investigate the surface interactions between aqueous Fe(II) ions and the hematite thin films and compare these results to previous studies.

II. EXPERIMENTAL SECTION A. Hematite Deposition. Fused-silica windows and hemispheres (ISP Optics) were sonicated in acetone for 10 min, followed by a 10 min sonication in isopropyl alcohol, and, finally, blown dry with a stream of N2. Prior to deposition, the fused-silica substrates were loaded into the ALD chamber and ozone-cleaned for 3 min. Silicon substrates for ellipsometric measurements were also cleaned using the procedure above. Fe2O3 deposition was carried out in a Cambridge Nanotech, Inc. Savannah 200 ALD reactor. The ALD reaction timing followed the sequence t1−t2−t3−t4, where t1 and t2 are the exposure and purge times for ferrocene, and t3 and t4 are the exposure and purge times for ozone. Ferrocene (Fe(Cp)2, 98%, Aldrich) was used without further purification. Ozone was generated by a DelOzone generator (5 wt % in 500 sccm ultrahigh-purity oxygen). Fe2O3 was deposited on the fusedsilica and silicon substrates using the sequence 60 s/35 s/90 s/ 20 s. We note that the 60 s Fe(Cp)2 exposure was split into two consecutive 30 s static exposures separated by a 5 s purge, and the 90 s ozone exposure was two consecutive 45 s pulses separated by a 1 s purge. Following deposition, the Fe2O3 thin films were annealed at 500 °C for 30 min in UHP O2 (250 sccm) using a Lindberg Blue three-zone furnace using a ramp rate of 10 °C/min. Fe2O3 film thicknesses were estimated using a J.A. Woolam Co. M2000 variable angle spectroscopic ellipsometer (VASE) from ellipsometric measurements on witness silicon chips. B. Raman Spectroscopy and X-ray Diffraction (XRD). Raman data were taken on a Renishaw Ramascope. One spectrum was recorded with 633 nm excitation, and a second spectrum was recorded with 785 nm excitation wavelength. Data were collected using a 50× objective, 100% laser power with a 100−1000 cm−1 range, and an acquisition time of 30 s. Figure 1A and B shows the Raman spectra for each excitation wavelength used. The vertical lines in each spectrum

Figure 1. ALD-deposited 10 nm hematite on a fused-silica substrate analyzed using Raman spectroscopy taken with an excitation wavelength of 785 nm (A) and 633 nm (B), and analyzed using GIAXRD (C). The vertical lines in A and B correspond to the Raman modes that are specific to the alpha phase of Fe2O3. The peaks in C are referenced to α-Fe2O3 (PDF 00-033-0664).

correspond to the Raman modes that are specific to the alpha phase of Fe2O3. Exact peak positions, as well as peak positions for other phases of Fe2O3 and other types of iron oxides, can be found in Supporting Information Table S2, along with the corresponding references. Figure 1A shows two unidentified peaks at ∼343 and ∼556 cm−1. These peaks are not present in Figure 1B and are very broad, leading to the conclusion that they are artifacts of the 785 nm excitation wavelength measurement. Grazing incidence angle X-ray diffraction (GIAXRD) experiments were performed on the 10 nm thin iron oxide films using a Rigaku ATX-G Thin Film Diffraction Workstation using Cu Kα radiation coupled to a multilayer mirror. These data are shown in Figure 1C. The peaks are referenced to α-Fe2O3 (PDF 00-033-0664). The absence of peak intensity at 2θ values corresponding to γ-Fe2O3 and Fe3O4 (see Table S3 in the Supporting Information) and the absence of peaks in the Raman spectra that correspond to γ-Fe2O3 and Fe3O4 suggests the absence of γ-Fe2O3 and Fe3O4 in the film. Instead, the Raman results and the presence of peak intensity at most of the 2θ values corresponding to α-Fe2O3 suggests that the thin films contain α-Fe2O3. The very low peak intensity corresponding to the 110 index of α-Fe2O3 suggests that there are preferred orientations in the crystallites. 4041

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

Figure 2. XPS spectrum of ALD-prepared hematite (A), a multiplet fit of the Fe 2p3/2 peak region (B), and SHG response at 300 nm as a function of input power using the p-in/all-out polarization combination (C) and a function of polarizer analyzer angle in the presence of pH 4 water (D) of ALD-prepared hematite.

E. Aqueous Solution Preparation. Ferrous iron solutions were prepared using a Fe(II)Cl2·4H2O (Aldrich, 99.99%) salt in Millipore water (18.2 MΩ) with a background NaCl (VWR, 99.0%) concentration of 1 mM. All solutions were maintained at pH 4 using dilute solutions of HCl (E.M.D., ACS grade) and NaOH (E.M.D., pellets). The speciation of the aqueous iron solutions was assessed using the ChemEQL program.69 These speciation calculations indicate that 99.5% of total dissolved Fe(II) is in the form of a free hexaquo ion under the experimental conditions used here. F. SHG, Laser, and Flow System. In this work, we use a nonresonant variant of SHG, the χ(3) technique, to monitor changes in the interfacial potential of the hematite surface as it is exposed to a charged species.70 SHG occurs only where centrosymmetry is broken, in this case at the interface between a bulk oxide and an aqueous phase.71 The SHG electric field (ESHG) is proportional to the induced second-order polarization (P2ω) as follows:

C. X-ray Photoelectron Spectroscopy (XPS). XPS experiments were performed using a Thermo Scientific ESCALAB 250Xi spectrometer utilizing Κα radiation from a monochromatic aluminum source. The spectrometer was calibrated to give a binding energy of 83.96 eV for the Au 4f7/2 line. Charge compensation was achieved using a flood gun, which ejected electrons and low-energy Ar+ ions. Narrow scans were taken with a pass energy of 10 eV and a step size of 0.05 eV. Charge correction was made using the peak for adventitious carbon, set to 284.8 eV. Fitting of the narrow scans was done utilizing the program Advantage v.5.33 with Gaussian− Lorentzian peaks with an L/G mix of 30%. The software’s smart background, a modified Shirley background, was subtracted from each of the narrow scans. Initial XPS measurements were performed to determine the purity of the hematite thin films. A survey scan of the polycrystalline thin film shows prominent iron and oxygen peaks with an adjusted Fe signal/O signal ratio of between 0.64 and 0.73, typical for hematite surfaces, and an adventitious carbon peak as a result of sample exposure to the ambient atmosphere (Figure 2A). A high-resolution scan of the Fe 2p region shows a small step in the low binding energy side of the 2p3/2 peak at 711.1 eV, a feature that is characteristic of αFe2O3 (Supporting Information Figure S3).66 A multiplet fit (Figure 2B), which explicitly shows the multiple peaks that result from the interaction between the core hole and the spin unpaired electrons in the valence shell, was performed on the Fe 2p3/2 peak and yielded fit parameters (Supporting Information Table S1) that conform well with previously published values for pure α-Fe2O3.67,68 D. Atomic Force Microscopy (AFM). AFM images of the ALD-hematite substrates were taken using a Bruker Icon atomic force microscope using tapping mode under ambient conditions in air. Images were taken before and after sample exposure to a solution of 1 mM Fe(II) at pH 4 and 1 mM NaCl. This solution was flowed across the ALD-hematite surface for 1 h at a flow rate of 1 mL/s.

ESHG ∝ P2ω = χ (2) EωEω + χ (3) EωEω Φ0

(1)

Here, Eω is the incident electric field at frequency ω, Φ0 is the interfacial potential, and χ(2) and χ(3) are the second- and thirdorder nonlinear susceptibility tensors, respectively.72,73 The physical interpretation of each tensor has been described previously.74 The presence of charged species, such as Fe(II) ions, in a system that possesses an interfacial potential, such as the hematite surface in contact with an aqueous phase, will interact with and modulate that interfacial potential, and a corresponding change in the SHG signal intensity will then be observed.75 This change in signal intensity is proportional to the concentration of charged particles in the system and can be modeled using mean field theories, such as the Gouy− Chapman model.76,77 Under the experimental conditions employed here, Eω, χ(2), and χ(3) are assumed to be constant,78 which allows us to follow the approach put forth by Salafsky and Eisenthal,70 who expressed the SHG signal intensity as a function of analyte concentration according to 4042

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C ⎡ ⎛ ⎛ K [M] ⎞⎞ ads ESHG = A + B′arcsinh⎢σ0 + ⎜⎜σm⎜ ⎟⎟ ⎢⎣ ⎝ ⎝ 1 + K ads[M] ⎠⎠ ⎤ π ⎥ 2εkBT (Celec + [M]) ⎥⎦

Article

5.5(3). Here, the digit in parentheses indicates the uncertainty associated with the least significant digit of the significand. We then determined the surface charge density, σ0, of the hematite/water interface at pH 4, 5.5, and 7 by following our previously published procedure (see Supporting Information Figure S6 for the SHG salt screening isotherms).79,80 As shown in Figure 3A, the surface charge was found to be

(2)

Here, A and B′ are fit parameters related to χ(2) and χ(3), kB is the Boltzmann constant, T is the system temperature, z is the valency of the background electrolyte in solution, ε is the dielectric constant for water at 25 °C, σ0 is the initial surface charge density, σm is the surface charge density at maximum metal ion surface coverage, Kads is the binding constant, M is the analyte concentration, and Celec is the background electrolyte concentration. The SHG experiments were carried out using a regeneratively amplified Ti:sapphire laser system (Hurricane, SpectraPhysics, 1 kHz repetition rate, 120 fs, 1 mJ), which pumps an optical parametric amplifier (OPA-CF, Spectra-Physics) tuned to a fundamental frequency of ω = 600 ± 10 nm. The laser beam was directed through a variable density filter where the peak power was attenuated to 0.30 ± 0.05 μJ per 120 fs pulse, which is far below the power level where optical breakdown of the surface occurs (Figure 2C). The beam was then focused onto the substrate/water interface at an angle just below that of total internal reflection. The generated second harmonic signal was isolated from the fundamental beam using a UV-grade Schott filter and directed into a monochromator set to 2ω. The second harmonic signal, which is well polarized along the plane of incidence when probing with p-polarized input light (Figure 2D) was then sent into a photomultiplier tube, amplified, and collected using a gated single-photon detection system. The SHG measurements were carried out under flow conditions using our previously published flow setup. A hematite-coated hemisphere substrate was placed on a Viton O-ring and clamped atop a custom-built Teflon flow cell. The flow cell was fed by two peristaltic pumps which draw from two different reservoirs, one containing Millipore water adjusted to pH 4 and 1 mM NaCl, and the other containing a Fe(II) solution at pH 4 and 1 mM NaCl. The flow rate was monitored using a flow meter and adjusted to ∼1 mL/s. Aliquots of solution were collected for every Fe(II) concentration employed. These aliquots were analyzed with inductively coupled plasma atomic emission spectroscopy (Varian) to determine the exact metal ion concentration. UV−vis spectra taken before and after flowing aqueous solutions relevant for this work over the ALD-prepared hematite indicate that the material is stable under those conditions (Supporting Information Figure 1S).

Figure 3. Surface charge density of ALD-prepared hematite as a function of bulk solution pH (A) and histogram of reported PZC values for natural (solid), commercial (hashed) and synthetic hematite (empty) and ALD result (vertical black line) (B).

0.003(1) C/m2 at pH 4, −0.005(1) C/m2 at pH 7, and zero at pH 5.5. These surface charge density values are within an order of magnitude of surface charge density values previously measured for hematite,81 which provides confirmation that the hematite substrates used in this work possess electrostatic properties similar to those of natural hematite surfaces found in the environment. As shown in Figure 3B, the PZC determined here for the hematite thin films is consistent with other PZC measurements for “natural” hematite samples,43,82−85 (PCZ = 5−7), but is below those reported for “commercial” (PCZ = 6− 9) or “synthetic” (PCZ = 7−10) hematite samples.43,82−87 The lower measured PZC value for the ALD-synthesized samples used in this work could be due to the polycrystalline nature of the material as well as the presence of adventitious carbon at the hematite surface. The classification used in Figure 3B follows the work of Kosmulski,86 which includes the higher PZC values of single-face PZC values from hematite single crystals and the lower PZC values obtained from bulk powders, with one of the significant reported impurities for natural hematite in this list being 6% silica by weight. B. Fe(II) Interaction With the Hematite−Solution Interface. The interaction of Fe(II) with the ALD-prepared hematite was assessed under dynamic flow conditions at pH 4 and in a 1 mM NaCl background electrolyte. Adsorption/

III. RESULTS AND DISCUSSION A. Point of Zero Charge at the Hematite−Solution Interface. To characterize the hematite thin-film substrates, we applied SHG to determine the point of zero charge (PZC), which is defined as the pH at which a given surface possesses a net neutral charge,76 that is, when the net potential at an interface is zero. The PZC is an important parameter because it imparts key information about the electrostatic behavior of a solid surface in an aqueous system. According to eq 1, there are no net charges to screen out when the pH equals the PZC, and the SHG response becomes invariant with the NaCl concentration. Figure S4 in the Supporting Information shows that this situation occurs at a bulk solution pH of 4043

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

repel Fe(II). This, then, is the nominal free energy landscape for site-specific Fe(II) interaction with the hematite surface. To assess the free energy of Fe(II)-hematite interaction, we collected equilibrium adsorption isotherms at pH 4 and 1 mM constant NaCl concentration, followed by analysis using the Langmuir adsorption model.89 Use of this model seemed to be justified given the full reversibility of Fe(II)-hematite interactions shown in Figure 4 and given the self-limiting surface coverages observed as the Fe(II) concentration was increased. Details of how these SHG χ(3) adsorption studies are performed can be found in our previous publications.79,80 Briefly, the SHG signal intensity was measured at multiple aqueous Fe(II) concentrations. The SHG E-field was then calculated for each analyte concentration by taking the square root of the raw SHG signal intensity. This value was normalized to a baseline water SHG E-field calculated from the SHG signal intensity in the absence of added Fe(II)aq. The normalized SHG E-field was then plotted against the metal concentration to yield an adsorption isotherm (Figure 5). This isotherm was

desorption traces in which we track the SHG signal intensity before, during, and after flowing aqueous Fe(II) across the interface at the same pH and NaCl concentration are shown in Figure 4 for two different Fe(II) concentrations. These traces

Figure 4. SHG E-field as a function of time before, during, and after flowing 0.034 mM (top) and 0.21 mM (bottom) Fe(II) across ALDprepared hematite at pH 4 and 1 mM NaCl at 298 K. The two traces are offset by 70 counts.

show that the SHG E-field decreases upon exposing the interface to Fe(II), and then the SHG signal intensity returns to the baseline level upon replacing the solution containing ferrous iron with background solution. This observation indicates that the interaction between Fe(II) and hematite is fully reversible. Full reversibility was observed at every metal concentration studied here. This finding is in agreement with results obtained by Jeon at al.,88 who reported that the interaction between Fe(II) and hematite particles at pH 4 and in 0.01 mM NaCl was fully reversible over a reaction time of several days. These observations of Fe(II) adsorption/desorption are unique from all previous studies in that they are specific to the interface. Our pH conditions are well below the generally accepted Fe(II) adsorption edge for hematite based on aqueous solution analytics (e.g., by colorimetric assay or mass spectrometry, etc.).10,88 Our surface-specific observations show that despite no expected net consumption of Fe(II) from bulk solution, there are significant interactions with Fe(II) at the interface nonetheless. As discussed previously, the hematite surface possesses a net positive surface charge of 0.003(1) C/m2 at pH 4, corresponding to 1−3 × 1012 positive charges per cm2. This number of positive charges is the sum of the negatively and positively charged surface sites located within a 1 cm2 area of the hematite/water interface and represents roughly 1−3% of the total number of surface sites within that area, assuming standard densities. Given that this approximation is probably reliable to within 1 order of magnitude, we conclude that the total number of positively charged sites at the hematite surface at pH 4 is at least 1 order of magnitude smaller than the total number of neutral sites, such as those present in the form of Fe2OH groups. Given this situation, there exist a large number of neutral and also some negatively charged adsorption sites that would allow for electrostatically favorable Fe(II) interactions, whereas the relatively small number of positively charged sites would

Figure 5. Adsorption isotherm for Fe(II) adsorbing to the hematite interface in a pH 4 and 1 mM NaCl solution at 298 K. The blue line represents the fit of the double layer model to the experimental data.

fit using eq 2, which was constrained by the experimentally determined value of σ0, the surface charge density, of the hematite/water interface at pH 4 (Figure 3A). From the isotherm fit, the observed binding constant for Fe(II) interacting with the hematite surface was found to be 7000 ± 2000 M−1, which corresponds to an adsorption free energy of −31.9(8) kJ/mol when referenced to 55.5 M water, and the surface charge density at maximum metal ion surface coverage was found to be 0.08(6) C/m2. The ΔGads value of −31.9(8) kJ/mol is very similar to adsorption free energies of other divalent ions interacting electrostatically with the silica surface at pH 790 and with trivalent ions adsorbing to silica at pH 4,55,80 whereas the upper bound of the maximum metal ion surface coverage corresponds to a total number of up to 4 × 1013 adsorbed Fe(II) ions, assuming they are, indeed, doubly charged. As mentioned in the Introduction, evidence for electron transfer from adsorbed Fe(II) ions to the bulk Fe(III) oxide has been observed in previous studies of the Fe(II)aqueous− Fe(III)oxide system.14−16,21 In the absence of complexing ligands, this process has been hypothesized to occur through a highly reversible formation of a bridged Fe(III)oxide−OH− Fe(II)ads surface complex.91,92 In the event of electron transfer, Scherer and co-workers and Rosso and co-workers reported that the newly formed Fe(III) surface species forms an additional layer of the Fe(III) oxide material.14,20 Although 4044

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

this second process is likely not to be reversible, it has been shown that the electron injected into the bulk Fe(III) oxide can reduce a structural Fe(III) atom elsewhere on the solid, which is subsequently released from the surface into the bulk aqueous solution as aqueous Fe(II).16,21 Conceptually, this process occurs even if there is no net reduction or oxidation of iron and, given our sensitivity to changes in interfacial potential, would appear to be macroscopically reversible using the χ (3) technique, consistent with the proposed reversible formation of the bridged surface complex. Whatever process occurs at the interface, it does not appear to produce precipitates under our experimental conditions: AFM images collected before and after aqueous flow of ferrous iron over the hematite sample showed (Figure 6) microtopography characteristic of a

Figure 6. AFM images of ALD-prepared hematite before (left) and after (right) exposure to 1 mM Fe(II) at 1 mM NaCl and pH 4 for 1 h. The average surface roughness is calculated to be between 0.8 and 1.0 nm.

Figure 7. Surface coverage of Fe(II) for ALD-prepared hematite (A) and for fused silica (B) as a function of aqueous Fe(II) concentration at pH 4 and 1 mM NaCl calculated using eq 3 and a site density of 5 sites/nm2 for A and 1 site/nm2 for B.

polycrystalline surface,93 but no changes in surface roughness (0.8 and 1.0 nm for before vs after ferrous iron exposure), which is consistent with other AFM studies of polycrystalline hematite thin films.94 Further insights into the interactions of Fe(II) ions with hematite surfaces can be obtained by plotting the SHG data as a function of the absolute number of adsorbed ions on the surface, Nads. This was achieved for Fe(II) adsorbing to the hematite surface as well as separately to the fused-silica surface to represent a redox-inert substrate, maintained at the same pH and electrolyte concentration (the SHG isotherm for Fe(II) adsorbing to fused silica is shown in Supporting Information Figure S6). To this end, we rearranged eq 4 according to

described as representing a change in surface speciation with increasing bulk aqueous concentration.96 The presence of two slopes (namely, of dNads/d[Fe(II)] = 1.8(8) × 1018 cm−2/M for Fe(II) concentrations below 100 μM and 2.0(2) × 1017 cm−2/ M for Fe(II) concentrations between 100 μM and 1 mM) shown in Figure 7A is remarkably reminiscent of the adsorption isotherm reported by Larese-Casanova and Scherer,14 who determined Fe(II) uptake to hematite nanoparticles at pH 7.4 in 25 mM background KBr using a batch extraction technique. The authors also used Mössbauer spectroscopy to determine that the steep slope at low Fe(II) concentrations represents an electron transfer process from aqueous Fe(II) to the Fe(III) oxide, whereas at higher Fe(II) concentrations, the shallow slope represents the formation of a stable Fe(II) surface species at surface site saturation. Because of the error associated with the σm parameter, we are not able to accurately assess surface site saturation in Figure 5; however, the similarity in the shape of our measured isotherm and the published isotherm of Larese-Casanova and Scherer leads to the conclusion that a similar process may be occurring at discrete surface sites of our polycrystalline thin films.

Nads

⎛ ⎞ ESHG − A ⎜ sinh ⎟ ρ B′ =⎜ − σ0⎟ π ⎜ ⎟ σm ⎝ 2εkBT(Celec + [M]) ⎠

(

)

(5)

Here, ρ is the density of the surface sites, which we took from the literature2,95 to be 5 sites/nm2 for hematite and 1 site/nm2 for fused silica. Figure 7A and B shows the density of the adsorbed ferrous ions, assuming that they are, indeed, divalent, as a function of aqueous Fe(II) concentration for the hematite/ solution and the fused-silica/solution interface, respectively, at pH 4 and 1 mM NaCl concentration. The plot for the fusedsilica surface appears to be Langmuirian, which is typical for an electrostatic adsorption process. However, the plot for the hematite surface shows not only roughly four times larger Fe(II) surface coverages than for the Fe(II)−fused-silica system, but also a steeper slope at low Fe(II) concentrations (up to 0.2 mM) and a shallow slope at high Fe(II) concentrations. This specific isotherm pattern has been

IV. CONCLUSIONS We have shown that the combination of second harmonic generation and sufficiently thin precision film deposition overcomes the classical “strong absorber problem” posed by most iron-bearing oxide materials, enabling detailed surfacespecific information in situ. We have demonstrated that ALDprepared and natural hematite samples behave similarly in terms of point of zero charge and that the interaction of their surfaces with aqueous Fe(II) ions can be readily quantified and analyzed in the context of the recently proposed bulk 4045

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

Pacific Northwest National Laboratory. A portion of the research was performed at Argonne National Laboratory, a U.S. Department of Energy, Office of Science, Laboratory operated under Contract No. DE-AC02-06CH11357 by UChicago Argonne, LLC.

conduction mechanism for coupling forward and reverse Fe(II)aqueous−Fe(III)oxide interfacial electron exchange in these fascinating materials. The data analysis is consistent with the notion that ALD-prepared hematite surfaces adsorb about four times more ferrous iron ions per square centimeter than fusedsilica surfaces at pH 4 and 1 mM NaCl. The interaction of Fe(II) with hematite at the interface is found to be fully reversible. On the basis of the recent work by Yanina and Rosso,20 the aqueous ferrous iron concentration is the result of a dynamic equilibrium that is due to its constant consumption and replenishment at two different sites on the hematite surface, according to the simplified set of coupled equilibria: Fe(II)aq + Si ⇔ Fe(II)adsSi and Fe(II)adsSi ⇔ Fe(III)adsSi + e−, where Si indicates a surface site on hematite with i = 1 for the forward and i = 2 for the backward reaction. Although the equilibrium constants for both processes could, in principle, be calculated from the two slopes, additional information is necessary to connect those results to electron transfer at the Fe(II)/hematite interface. Specifically, future work will focus on how the two slopes in the two concentration regimes depend on pH, ionic strength, and the thickness of the ALD-prepared hematite, with the goal of pursuing a full thermodynamic analysis of the Fe(II)/hematite system. By opening the surfaces of hematite to study by nonlinear optics, we envision that it will now be possible to pursue a wide variety of experiments, ranging from spectroscopic to imaging to time-resolved and ultrafast measurements on hematite, goethite, iron oxyhydroxide, and other important and abundant iron oxide forms that are of relevance for understanding, controlling, and predicting their behavior in the environment, industrial systems, and in technological applications.

■

■

ASSOCIATED CONTENT

S Supporting Information *

SHG salt screening data and a table of XPS fit parameters can be found in the Supporting Information file. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Whitman, W. G. Chem. Rev. 1926, 2, 419−35. (2) Cornell, R. M.; Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrences, and Uses; WILEY-VCH Verlag GmbH & Co. KGaA: Weinheim, 2003. (3) Stumm, W.; Sulzberger, B. Geochim. Cosmochim. Acta 1992, 56, 3233−57. (4) Kappler, A.; Straub, K. L. Rev. Mineral. Geochem. 2005, 59, 85− 108. (5) Weber, K. A.; Achenbach, L. A.; Coates, J. Nat. Rev. Microbiol. 2006, 4, 752−64. (6) Charlet, L.; Liger, E.; Gerasimo, P. J. Environ. Eng. 1998, 124, 25− 30. (7) Liger, E.; Charlet, L.; Van Cappellen, P. Geochim. Cosmochim. Acta 1999, 63, 2939−55. (8) Frierdich, A. J.; Luo, Y.; Catalano, J. G. Geology 2011, 39, 1083− 6. (9) Crosby, H. A.; Roden, E. E.; Johnson, C. M.; Beard, B. L. Environ. Sci. Technol. 2005, 39, 6698−704. (10) Jeon, B.; Dempsey, B.; Burgos, W. Environ. Sci. Technol. 2003, 37, 3309−15. (11) Coughlin, B. R.; Stone, A. T. Environ. Sci. Technol. 1995, 29, 2445−55. (12) Hansel, C. M.; Benner, S. G.; Fendorf, S. Environ. Sci. Technol. 2005, 39, 7147−53. (13) Rosso, K. M.; Zachara, J. M.; Fredrickson, J. K.; Gorby, Y. A.; Smith, S. C. Geochim. Cosmochim. Acta 2003, 67, 1081−7. (14) Larese-Casanova, P.; Scherer, M. M. Environ. Sci. Technol. 2007, 41, 471−7. (15) Williams, A.; Scherer, M. Environ. Sci. Technol. 2004, 38, 4782− 90. (16) Catalano, J. G.; Fenter, P.; Park, C.; Zhang, Z.; Rosso, K. M. Geochim. Cosmochim. Acta 2010, 74, 1498−512. (17) Charlet, L.; Silvester, E.; Liger, E. Chem. Geol. 1998, 151, 85−93. (18) Zinder, B.; Furrer, G.; Stumm, W. Geochem. Cosmochim. Acta 1986, 50, 1861−9. (19) Jolivet, J. P.; Tronc, E. J. Colloid Interface Sci. 1988, 125, 688− 701. (20) Yanina, S. V.; Rosso, K. M. Science 2008, 320, 218−22. (21) Handler, R. M.; Beard, B. L.; Johnson, C. M.; Scherer, M. M. Environ. Sci. Technol. 2009, 43, 1102−7. (22) Rosso, K. M.; Yanina, S. V.; Gorski, C. A.; Larese-Casanova, P.; Scherer, M. M. Environ. Sci. Technol. 2010, 44, 61−7. (23) Ho, C. H.; Miller, N. H. J. Colloid Interface Sci. 1986, 110, 165− 71. (24) Waite, T. D.; Torikov, A.; Smith, D. D. J. Colloid Interface Sci. 1986, 112, 412−20. (25) Waite, T. D.; Torikov, A. J. Colloid Interface Sci. 1987, 119, 228− 35. (26) Barron, V.; Herruzo, M. G.; Torrent, J. Soil Sci. Soc. Am. J. 1988, 52, 647−51. (27) Barron, V.; Torrent, J. J. Colloid Interface Sci. 1996, 177, 407−10. (28) Becker, U.; Hochella, M. F., Jr.; Apra, E. Am. Mineral. 1996, 81, 1301−14. (29) Jones, F.; Farrow, J. B.; van Bronswijk, W. Langmuir 1998, 14, 6512−7. (30) Samson, S. D.; Eggleston, C. M. Environ. Sci. Technol. 1998, 32, 2871−5. (31) Wang, X. G.; Weiss, W.; Shaikhutdinov, S. K.; Ritter, M.; Petersen, M.; et al. Phys. Rev. Let. 1998, 81, 1038−41. (32) Grolimund, D.; Trainor, T. P.; Fitts, J. P.; Kendelewicz, T.; Liu, P.; et al. J. Synchrotron Radiat. 1999, 6, 612−4.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Special thanks to Jonathan W. Hennek for the XRD data collection and interpretation. This work was supported by the National Science Foundation Environmental Chemical Sciences program under Grant No. CHE-0950433. We also acknowledge the International Institute for Nanotechnology (IIN) at Northwestern University for capital equipment support and an Irving M. Klotz professorship to F.M.G. We acknowledge Spectra-Physics Lasers, a division of Newport Corporation, for equipment support. The ICP-AES analysis was completed at the Northwestern University Integrated Molecular Structure Education and Research Center (IMSERC). Part of this work was supported as part of the ANSER Center, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences (OBES) under Award No. DE-SC0001059. K.M.R. acknowledges support from the OBES Division of Chemical Sciences, Geosciences, and Biosciences through 4046

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047

The Journal of Physical Chemistry C

Article

(33) Kendelewicz, T.; Liu, P.; Doyle, C. S.; Brown, G. E.; Nelson, E. J.; Chambers, S. A. Surf. Sci. 1999, 424, 219−31. (34) Lenhart, J. J.; Honeyman, B. D. Geochem. Cosmochim. Acta 1999, 63, 2891−901. (35) Jones, F.; Rohl, A. L.; Farrow, J. B.; van Bronswijk, W. Phys. Chem. Chem. Phys. 2000, 2, 3209−16. (36) Namjesnik-Dejanovic, K.; Maurice, P. A. Geochim. Cosmochim. Acta 2000, 65, 1047−57. (37) Duckworth, O. W.; Martin, S. T. Geochim. Cosmochim. Acta 2001, 65, 4289−301. (38) Lenhart, J. J.; Bargar, J. R.; Davis, J. A. J. Colloid Interface Sci. 2001, 234, 448−52. (39) Martin, S. T.; Han, J. H.; Hung, H. M. Geophys. Res. Lett. 2001, 28, 2601−4. (40) Redman, A. D.; Macalady, D. L.; Ahmann, D. Environ. Sci. Technol. 2002, 36, 2889−96. (41) Warschkow, O.; Ellis, D. E.; Hwang, J. H.; Mansourian-Hadavi, N.; Mason, T. O. J. Am. Ceram. Soc. 2002, 85, 213−20. (42) Eggleston, C. M.; Stack, A. G.; Rosso, K. M.; Higgins, S. R.; Bice, A. M.; et al. Geochem. Cosmochim. Acta 2003, 67, 985. (43) Das, M. R.; Bordoloi, D.; Borthakur, P. C.; Mahiuddin, S. Colloids Surf. A 2005, 254, 49−55. (44) Baltrusaitis, J.; Cwiertny, D. M.; Grassian, V. H. Phys. Chem. Chem. Phys. 2007, 9, 5542−54. (45) Jang, J-H; Dempsey, B. A.; Burgos, W. D. Environ. Sci. Technol. 2007, 41, 4305−10. (46) Ha, J.; Yoon, T. H.; Wang, Y.; Musgrave, C. B.; Brown, G. E. Langmuir 2008, 24, 6683−92. (47) Zeng, H.; Sing, A.; Basak, S.; Ulrich, K-U; Sahu, M.; et al. Environ. Sci. Technol. 2009, 43, 1373−8. (48) Jubb, A. M.; Allen, H. C. ACS Appl. Mater. Interfaces 2010, 2, 2804−12. (49) Skomurski, F. N.; Rosso, K. M.; Krupka, K. M.; McGrail, B. P. Environ. Sci. Technol. 2010, 44, 5855−61. (50) Adamescu, A.; Hamilton, I. P.; Al-Abadleh, H. A. Environ. Sci. Technol. 2011, 45, 10438−44. (51) Adamescu, A.; Mitchell, W.; Hamilton, I. P.; Al-Abadleh, H. A. Environ. Sci. Technol. 2010, 44, 7802−7. (52) Chabot, M.; Hoang, T.; Al-Abadleh, H. A. Environ. Sci. Technol. 2009, 43, 3142−7. (53) Mitchell, W.; Goldberg, S.; Al-Abadleh, H. A. J. Colloid Interface Sci. 2011, 358, 534−40. (54) Tofan-Lazar, J.; Al-Abadleh, H. A. J. Phys. Chem. A 2012, 116, 1596−604. (55) Jordan, D. S.; Saslow, S. A.; Geiger, F. M. J. Phys. Chem. A 2011, 115, 14438−45. (56) Hayes, P. L.; Malin, J. N.; Konek, C. T.; Geiger, F. M. Geochem. Cosmochim. Acta 2009, 73, A506. (57) Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P. Langmuir 2011, 27, 10007−18. (58) Huang, Z.; Guo, Y. Chem. J. Chin. Univ. (Chinese) 2012, 33, 1517−22. (59) Vandendriessche, S.; Valev, V. K.; Verbiest, T. Appl. Opt. 2012, 51, 209−13. (60) Baldelli, S.; Cimatu, K. Abstr. Papers Am. Chem. Soc. 2008, 235. (61) Baldelli, S.; Romero, C. R.; Zhang, H.; Cimatu, K. Abstr. Papers Am. Chem. Soc. 2006, 232. (62) Cimatu, K.; Baldelli, S. J. Phys. Chem. C 2007, 111, 7137−43. (63) Romero, C. R.; Zhang, H. P.; Baldelli, S. Abstr. Papers Am. Chem. Soc. 2005, 230, U1089−U90. (64) Zhang, H. P.; Romero, C. R.; Baldelli, S. Abstr. Papers Am. Chem. Soc. 2005, 230, U1134−U5. (65) Martinson, A. B. F.; DeVries, M. J.; Libera, J. A.; Christensen, S. T.; Hupp, J. T.; et al. J. Phys. Chem. C 2011, 115, 4333−9. (66) Aronniemi, M.; Lahtinen, J.; Hautojarvi, P. Surf. Interface Anal. 2004, 36, 1004−6. (67) Junta-Rosso, J.; Hochella, M. Geochim. Cosmochim. Acta 1996, 60, 305−14.

(68) Biesinger, M. C.; Payne, B. P.; Grosvenor, A. P.; Lau, L. W. M.; Gerson, A. R.; Smart, R. S. C. Appl. Surf. Sci. 2011, 257, 2717−30. (69) Muller, B. ChemEQL; EAWAG: Kastanienbaum, Switzerland, 1996. (70) Salafsky, J.; Eisenthal, K. J. Phys. Chem. B 2000, 104, 7752−5. (71) Eisenthal, K. Chem. Rev. 1996, 96, 1343−60. (72) Xiao, X.; Vogel, V.; Shen, Y. Chem. Phys. Lett. 1989, 163, 555−9. (73) Zhao, X.; Ong, S.; Eisenthal, K. Chem. Phys. Lett. 1993, 202, 513−20. (74) Yan, E.; Liu, Y.; Eisenthal, K. J. Phys. Chem. B 1998, 102, 6331− 6. (75) Hayes, P. L.; Malin, J. N.; Jordan, D. S.; Geiger, F. M. Chem. Phys. Lett. 2011, 499, 183−92. (76) Langmuir, D. Aqueous Environmental Geochemistry; PrenticeHall, Inc: Upper Saddle River, NJ, 1997. (77) Stumm, W; Morgan, J. J. Aquatic Chemistry; John Wiley & Sons, Inc.: New York, 1996. (78) Jena, K. C.; Covert, P. A.; Hore, D. K. J. Phys. Chem. Lett. 2011, 2, 1056−61. (79) Hayes, P. L.; Malin, J. N.; Konek, C. T.; Geiger, F. M. J. Phys. Chem. A 2008, 112, 660−8. (80) Jordan, D. S.; Malin, J. N.; Geiger, F. M. Environ. Sci. Technol. 2010, 44, 5862−7. (81) Rustad, J. R.; Wasserman, E.; Felmy, A. R. Surf. Sci. 1999, 424, 28−35. (82) Ahmed, S. M.; Maksimov, D. Can. J. Chem. 1968, 46, 3841−6. (83) Song, Q. Y.; Xu, F.; Tsai, S. C. Int. J. Miner. Process. 1992, 34, 219−29. (84) Wang, Y.; Pugh, R. J.; Forssberg, E. Colloids Surf. A 1994, 90, 117−33. (85) Yuhua, W.; Jianwei, R. Int. J. Miner. Process 2005, 77, 116−22. (86) Kosmulski, M. Properties of Material Surfaces; CRC Press and Marcel Dekker: 270 Madison Ave., New York, NY, 2001. (87) Joy, A. S.; Watson, D.; Cropton, R. W. G. Trans. AIME 1964, 229, 5−7. (88) Jeon, B. H.; Dempsey, B. A.; Burgos, W. D.; Royer, R. A. Colloids Surf. A 2001, 191, 41−55. (89) Masel, R. I. Principles of Adsorption and Reaction on Solid Surfaces; John Wiley & Sons: New York, 1996. (90) Malin, J. N.; Hayes, P. L.; Geiger, F. M. J. Phys. Chem. C 2009, 113, 2041. (91) Silvester, E.; Charlet, L.; Tournassat, C.; Gehin, A.; Greneche, JM; Liger, E. Geochim. Cosmochim. Acta 2005, 69, 4801−15. (92) Wehrli, B.; Sulzberger, B.; Stumm, W. Chem. Geol. 1989, 78, 167−79. (93) Nilsen, O.; Lie, M.; Foss, S.; Fjellvag, H.; Kjekshus, A. Appl. Surf. Sci. 2004, 227, 40−7. (94) Glasscock, J. A.; Barnes, P. R. F.; Plumb, I. C.; Bendavid, A.; Martin, P. J. Thin Solid Films 2008, 516, 1716−24. (95) Christl, I.; Kretzschmar, R. Geochim. Cosmochim. Acta 1999, 63, 2929−38. (96) Katz, L. E.; Hayes, K. F. J. Colloid Interface Sci. 1995, 170, 477− 90.

4047

dx.doi.org/10.1021/jp3113057 | J. Phys. Chem. C 2013, 117, 4040−4047