Anaerobic Reaction of Nanoscale Zerovalent Iron with Water

Jun 2, 2014 - Nanoscale zerovalent iron (nZVI) is commonly used in advanced groundwater remediation processes. Here, we present a combined experimenta...
0 downloads 15 Views 5MB Size
Article pubs.acs.org/JPCC

Anaerobic Reaction of Nanoscale Zerovalent Iron with Water: Mechanism and Kinetics Jan Filip,† František Karlický,† Zdeněk Marušaḱ ,† Petr Lazar,† Miroslav Č erník,‡ Michal Otyepka,*,† and Radek Zbořil*,† †

Regional Centre of Advanced Technologies and Materials, Departments of Physical Chemistry and Experimental Physics, Faculty of Science, Palacký University, 17. listopadu 1192/12, 771 46 Olomouc, Czech Republic ‡ Centre for Nanomaterials, Advanced Technologies and Innovations, Technical University of Liberec, Studentská 2, 46 117 Liberec, Czech Republic S Supporting Information *

ABSTRACT: Nanoscale zerovalent iron (nZVI) is commonly used in advanced groundwater remediation processes. Here, we present a combined experimental and computational approach to elucidate the mechanism and kinetics of the reaction of nZVI with water under anaerobic conditions, which represents the basic reaction controlling the stability of nZVI in groundwater. The reaction kinetics was monitored at temperatures of 25 and 80 °C by 57Fe Mössbauer spectroscopy on frozen dispersion samples. The experimentally determined rate constant for reaction of nZVI with water at 25 °C was 1.14 × 10−3 h−1; the activation barrier measured for 60 nm sized nanoparticles (ΔG⧧298K(aq) = 26.3 kcal/mol) fits the range delineated by two limiting theoretical models from advanced quantum chemical calculations: rate-limiting activation barriers of 31.6 and 18.0 kcal/mol depending on the computational model, i.e., an iron atom and an infinite iron surface, respectively. The computations indicated a two-step reaction mechanism involving two one-electron transfer processes: the first can be described by the reaction Fe + H2O → HFeOH, which represents the rate-limiting step, and the second by HFeOH + H2O → Fe(OH)2 + H2. At 25 °C, the reaction product was identified experimentally as Fe(OH)2, which forms flat layered sheets extensively overgrowing nZVI particles. At 80 °C, ferrous hydroxide undergoes secondary anaerobic transformation to magnetite (Fe3O4).



INTRODUCTION Reductive technologies for the decontamination of ground- and wastewater using zerovalent iron (ZVI), and especially nanoscale zerovalent iron (nZVI), have become increasingly popular in the past two decades.1−4 The high reductive capacity of ZVI has been known for a long time, but recently developed efficient methods for the preparation of micro- and nanoscale ZVI particles have resulted in the widespread use of this powerful reductant. nZVI performs well in the decontamination of various organic,5−11 inorganic,12−15 and microbial (e.g., bacteria, such as cyanobacteria)16,17 pollutants. It is particularly useful for in situ applications because the products generated by the oxidation of metallic iron nanoparticles are nontoxic iron oxides commonly occurring in sediments and soils.18 In water, and particularly in groundwater where anaerobic conditions prevail, nZVI particles mainly act as (i) a reductant, which is typically used for dehalogenation of chlorinated hydrocarbons, (ii) a sorbent with adsorption of pollutants either on nZVI itself or on subsequently formed iron oxides (this property is used mainly for removal of heavy metals), and (iii) a coagulant, where dissolution followed by immediate precipitation of iron is a critical parameter.1,3,19 Typically, combined reductive-sorption and/or reductive-(co)precipitation processes occur, leading to efficient degradation and removal of a broad range of toxic compounds.18 In all the © 2014 American Chemical Society

above-mentioned modes of nZVI action, the primary reaction of nZVI with water represents the key aspect due to (i) production of the primary radical and ionic species able to chemically degrade pollutants,20 (ii) formation of iron oxide/ oxyhydroxide nanoparticles that act as efficient sorbents (coprecipitates),18 and (iii) competitive reactions in which the nZVI surface acts as a reductant.3,19 Undoubtedly, an understanding of the mechanism and kinetics of the nZVI reaction with water under anaerobic conditions is of fundamental importance to enable the various modes of nZVI action against particular pollutants to be distinguished and to predict nZVI stability in groundwater. Surprisingly, with respect to the fundamental importance of nZVI anaerobic reaction with water,21,22 controversial data have appeared in a few published studies to date. Contrary to macroscopic/microscopic iron,23 in the case of iron nanoparticles, the main analytical problems lie in the ambiguous identification of the final reaction products (i.e., magnetite vs Fe(OH)2), discordance in the derived rate constants of the nZVI reaction with various substances, and clear definition of the reaction mechanism. Therefore, unambiguous identification Received: December 6, 2013 Revised: May 11, 2014 Published: June 2, 2014 13817

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

of the primary reductive species and definition of individual reaction steps represent the main gaps in knowledge of nZVI anaerobic reaction with water, being tightly related to the kinetics of hydrogen evolution, phase equilibria, reactivity changes upon aging, etc.12,24−26 We believe that the principal reason for the contradictory data is the experimental and analytical approaches used, particularly since previously the contact of nZVI with air has not been completely prevented during (i) sample pretreatment, (ii) reaction with water, or (iii) subsequent analyses after the reaction. Moreover, the significance of hydrogen evolution and hydroxyl radical formation is generally underestimated. Considering the complexity and possible misinterpretation of the experimental observations on the nZVI/water system under anaerobic conditions, it would be highly beneficial to compare exact experimental data with theoretical calculations. Therefore, the aim of this study is to thoroughly elucidate the mechanism and kinetics of the reaction of nZVI (60 nm particles) with water under anaerobic conditions at 25 and 80 °C. To simulate the anoxic conditions of typical groundwater, all experiments were performed under an oxygen-free atmosphere of nitrogen in a glovebox. 57Fe Mössbauer spectroscopy on frozen dispersion samples was used for the identification and quantification of the solid reaction species. The combination of experiments performed in a glovebox and Mössbauer measurements of solid phases in frozen samples (i.e., for elimination of subsequent iron oxidation) with the ability to unambiguously identify and quantify all solid species in the reaction systems predict the acquired results for further generalization. Moreover, the combination of experimental results with computational chemistry enabled the primary reaction steps to be elucidated, including the first unambiguous identification of the primary reductive intermediates and ratelimiting step.

atmosphere in a Jacomex P(BOX) glovebox. During all experiments, the monitored concentration of oxygen in the glovebox was below 20 ppm and the N2 gas pressure was maintained at 98 Pa above atmospheric pressure. Samples for qualitative and quantitative 57Fe Mössbauer analyses were extracted at different reaction times in the glovebox, immediately frozen in liquid N2, and transferred to the cryostat of the Mössbauer spectrometer for measurement. Simultaneously, the pH of the first sample of each experimental sequence was measured outside the glovebox by a calibrated conventional laboratory pH meter (Multi 350i instrument, WTW GmbH, Germany, equipped with an SenTix 41-3 electrode) in glass vials placed on a magnetic plate for 5 min to separate out the nZVI particles. Instrumentation Employed for Characterization of nZVI Particles and Reaction Products. When dealing with reaction systems containing iron-based solid reactants and products, 57Fe Mö ssbauer spectroscopy offers a unique analytical tool for qualitative and quantitative monitoring of both the reaction kinetics and reaction mechanism. This is because 57Fe Mössbauer spectroscopy is entirely selective for iron atoms and allows their valence and spin states to be unambiguously determined and distinguished among various iron oxides/(oxy)hydroxides. Transmission 57Fe Mössbauer spectra were collected in the constant acceleration mode with a 57 Co (Rh) source (1.85 GBq). The absorbers were prepared in a glovebox under a protective atmosphere of nitrogen as concentrated dispersions (∼5 mg of Fe cm−2) sealed between two inner protective Parafilm M plates with an additional outer sealing Parafilm M layer. They were immediately fast-frozen in a liquid nitrogen bath and kept at 200 K using a closed He cycle cryostat during collection of the spectra. Isomer shift values were calibrated against an α-Fe foil at room temperature. Spectra were fitted with Lorentzian line shapes using the leastsquares method in the MossWinn computer program. The contents of different Fe species were quantified from the corresponding relative subspectral areas. The effects of nonideal absorber thickness and variable recoil-free fractions for iron atoms in nonequivalent structural sites of different phases were expected to be within experimental errors (hyperfine parameters, ±0.02 mm s−1; relative spectral area, ±3%). X-ray powder diffraction (XRD) measurement was performed on a PANalytical X’Pert PRO diffractometer (Bragg− Brentano geometry, equipped with a fast X’Celerator detector and iron-filtered CoKα radiation, 40 kV, 30 mA). The magnetically preconcentrated nZVI suspension was quickly inserted into a conventional front-loading shallow cavity sample holder and repeatedly scanned over the 2θ range of 5−105° under ambient conditions (12 fast scans per hour). The acquired patterns were processed using X’Pert HighScore Plus software and the PDF-4+ database. Detailed morphological investigation of the prepared and reacted nZVI samples (i.e., solid products of the reaction of nZVI with water) was performed on a JEOL JEM-2010 transmission electron microscope equipped with a LaB6 cathode (accelerating voltage of 160 kV, point-to-point resolution of 0.194 nm). A drop of high-purity distilled water containing the ultrasonically dispersed nZVI particles was placed onto a holey carbon film supported on a copper mesh transmission electron microscopy (TEM) grid (SPI Supplies, United States) and dried at room temperature; all procedures were performed in a glovebox. Scanning electron microscopy (SEM) images were collected on a field emission scanning



MATERIALS AND METHODS Source and Characteristics of nZVI. The commercially available nZVI particles “NANOFER 25P” manufactured by NANO IRON, s.r.o. (thermally synthesized from iron oxide powder in hydrogen) were additionally thermally reduced in hydrogen for a prolonged time to get nZVI particles containing 100 wt % of metallic α-Fe for subsequent experiments. The average particle size was ∼60 nm according to transmission electron microscopy, the specific surface area was ∼36 m2 g−1 on the basis of BET measurements, and the nanoparticles were shown to be exclusively composed of metallic iron (α-Fe, bcc structure) with no other detectable phases on the basis of 57Fe Mössbauer spectroscopy (see Figure S1 in the Supporting Information). The prepared nZVI particles were stored in a dry state in sealable metal containers under a protective nitrogen atmosphere at room temperature. Batch Experiments. Dispersion samples for kinetic measurements contained 200 mg of nZVI in 50 mL of distilled water (previously degassed using a N2 flow). The experiments were performed in 60 mL glass screw-cap vials, allowing the samples to be depressurized following hydrogen evolution (namely, the vials were quickly opened in a glovebox upon unscrewing the top cover immediately after shaking; see below). All dispersion samples were prepared with deionized water (18 MΩ cm−1, Millipore). The reactions of nZVI with water were carried out at 25 and 80 °C with periodic shaking in a glovebox (hand-shaking for approximately 5 s every day). All batch experiments were performed under an anaerobic N2 13818

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

electron microscope (Hitachi SU6600) with ultrahigh point-topoint resolution (1−2 nm), operating at 5 kV, and equipped with an energy-dispersive X-ray (EDX) spectrometer, NORAN (Thermo Scientific). Quantum Chemical Calculations. Theoretical study of nZVI (and generally any kind of nanoparticle) is challenging because of difficulties in choosing an appropriate model system. The nZVI particles used in this study contained hundreds of thousands of iron atoms (surrounded by millions of water molecules); such large systems are computationally intractable. As such, it is preferable to model the nZVI particle either as a cluster of a few iron atoms or as a solid-phase surface.26 Unfortunately, both approaches suffer from drawbacks and limitations; in particular, the sizes of systems involved, the need for periodic boundary conditions, and simulation of electron correlation mean that the use of density functional theory (DFT) methods is required. This is a potentially serious issue because DFT functionals do not always provide systematic results for transition-metal (TM) compounds, and the accuracy of their results is highly system dependent.27,28 Moreover, because of the exponential dependence of the rate constant (measured) on the free energy barrier (calculated) in the Eyring equation, for reasonable comparison, it is necessary to obtain calculated energy differences with accuracy on the order of 1 kcal/mol, which can only be achieved by very advanced quantum chemical (QC) methods beyond DFT. Therefore, we chose to consider two limiting models: an atomic model, which represents the simplest model for the nZVI reaction with water but allows QC calculations with high thermochemical accuracy, i.e., an error of less than 3 kcal/mol for TMs,29 and an infinite iron surface, which was treated using random phase approximation (RPA).30 Very recently, we showed31 that the RPA, which is formulated beyond DFT, provided reaction kinetics and thermodynamics in good agreement with the reference CCSD(T) data.31 The atomic approach adopted in this work has previously been used to study the reaction of iron with CCl4 and crossvalidated with experimental data.32,33 It should be noted that the reactions of a water molecule with Fe(100) and Fe(111) surfaces are in many respects analogous to the water molecule reaction with a single Fe atom.31,34 To calculate the required energy barriers according to transition-state theory, the key minima and transition points on the potential energy surface of the model system have to be identified. Geometry optimization was performed using the DFT method (B97-1 functional35) with a cc-pVTZ basis set.36 Zero-point energies and ΔG corrections using harmonic approximation were obtained at the same level of theory. Single-point unrestricted CCSD(T) energy calculations using a complete basis set (CBS) extrapolation scheme on the cc-pVTZ and cc-pVQZ basis sets were carried out. Alternatively, scalar relativistic (using the Douglas−Kroll−Hess method of second order, DKH) CCSD(T)-DKH and valence plus outer-core (3s3p3d4s) electronic correlation CCSD(T)-3s3p-DKH calculations were also performed using the CBS limit and cc-pVnZ-DK and ccpwCVnZ-DK basis sets,36 respectively. We also included implicit solvent effects (εr = 78.3553 for water) at the DFT level using a self-consistent isodensity polarizable continuum model (SCIPCM).37 All calculations were performed using the Gaussian suite of programs.38 The reliability of the method used for electronic structure calculations has been extensively tested in our previous works.27,39

The calculations on an infinite iron surface were performed using the projector-augmented wave (PAW) method in the plane-wave VASP code.40,41 The Fe(111) surface was modeled by a periodic slab with the (2 × 2) surface geometry. The slab contained six atomic layers (i.e., 24 iron atoms in total) separated by 10 Å of vacuum. The energy cutoff of 400 eV and 4 × 4 × 1 k-point sampling were used to ensure sufficiently accurate total energies and forces. Force relaxation was stopped when the forces acting on the atoms were converged to 10−3 eV/Å. The exchange-correlation effects were described by the Perdew−Burke−Ernzerhof (PBE) functional.42 Spin polarization was allowed in all calculations. ΔG corrections for the reaction barrier of the surface model were obtained from calculations on the Fe7 cluster model43 using Gaussian 09.38 The adsorption energies were defined as the total energy difference of the Fe slab with adsorbed species relative to the sum of the energies of isolated H2O molecules and the isolated Fe slab. The initial geometries of the adsorbed species were chosen according to our previous study.34 The configurations corresponding to increased coverage were created by adding further species into empty surface sites and by relaxation of the forces within the conjugate-gradient algorithm. Several configurations were inspected for each surface coverage to justify that thermodynamically the most stable one was found.



RESULTS AND DISCUSSION Mechanism and Kinetics at 25 °C: Experimental Observations. The transmission 57Fe Mössbauer spectra of the dispersion samples collected at various reaction times at 25 °C exhibited one sextet with hyperfine parameters typical for metallic α-Fe (for a data-collection temperature of 200 K, the isomer shift δ = 0.05 mm s−1, quadrupole shift εQ = 0.00 mm s−1, and hyperfine magnetic field Bhf = 33.1 T)44 and one paramagnetic doublet of Fe2+ in a high-spin state (for a datacollection temperature of 200 K, δ = 1.20 mm s−1 and the quadrupole splitting ΔEQ = 2.92 mm s−1; see Table S1 in the Supporting Information), assigned to Fe(OH)2.45 Representative spectra are shown in Figure 1. With increasing reaction time, the central doublet of Fe(OH)2 became increasingly dominant in the spectra, whereas the spectral area of the sextet ascribed to metallic iron decreased exponentially to ∼30% after 133 days (Figure 2). No other Fe-bearing species were detected, even after 133 days, and the only detectable solid product of the reaction of nZVI with water was thus Fe(OH)2. The Fe(OH)2 phase was also confirmed by XRD measurement (Figure 3). This suggests that reaction 1 could represent a general formula for the reaction between nZVI and water under anaerobic conditions at 25 °C: Fe + 2H 2O → Fe(OH)2 + H 2

(1)

To quantify the kinetics of the nZVI reaction with water at 25 °C (i.e., on the basis of the data in Figure 2), a pseudo-firstorder reaction model with a correction for passivation (through an additive term; see the Supporting Information) was applied (R2 = 0.999), yielding a rate constant for reaction 1 of 1.14 × 10−3 h−1 (±3.6 × 10−5) and a calculated reaction half-life of 608 h. The corresponding activation Gibbs energy barrier ΔG⧧ calculated from the Eyring equation was 26 kcal/mol. It should be noted that the Eyring equation was derived for homogeneous reactions, and therefore, the estimated activation Gibbs barrier for the heterogeneous reaction of nZVI with water should be used only for semiquantitative comparisons, keeping in mind all potential limitations. The experimental measure13819

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

Figure 3. X-ray diffraction patterns demonstrating fast transformation of Fe(OH)2 at ambient conditions (sample of nZVI particles collected after 240 h of reaction with water at 25 °C under anaerobic conditions). In the first scan (t = 0 min), the Fe(OH)2 phase is the only reaction product; the last scan (t = 60 min) shows magnetite (Fe3O4) is present instead of Fe(OH)2. Respective PDF cards shown below the XRD patterns: Fe(OH)2, 00-013-0089; Fe3O4, 01-089-0691; α-Fe, 03-065-4899.

morphology is characteristic of iron(II) hydroxide, and the observed crystals were typically hundreds of nanometers long with a thickness of a few atomic layers up to several tens of nanometers. The results of microscopic observation were thus in accordance with those from 57Fe Mössbauer spectroscopy and XRD. One of the most important observations resulting from the experimental monitoring of the nZVI interaction with water is that Fe(OH)2 is the only solid reaction product formed at 25 °C. Though it is in full accordance with the known mechanism of anaerobic surface corrosion of macroscopic iron and steel,46−48 and also consistent with predictions made for nZVI,49 this finding is in contrast to several previously published studies focused on anaerobic corrosion of nZVI.21,22,50 In all of those reports, magnetite was identified as the only reaction product formed at the same experimental temperature as in our study. In both our and Reardon’s21 study, nZVI particles of approximately the same surface area (i.e., ∼30 m2 g−1) were used and all experiments were conducted under anaerobic conditions with similar nZVI/water weight ratios and a pH value around 11 shortly after reaction initiation. The apparent rate constants are of the same order, and saturation of the reaction kinetics was clearly observed in both cases. However, contrary to Reardon et al.,21 and other previously published studies,22,50 we did not detect any traces of magnetite even after 3000 h of nZVI reaction with water. In this study, we used 57Fe Mö ssbauer spectroscopy on frozen dispersion samples to monitor the reaction mechanism (i.e., to identify the reaction products) and kinetics in detail, whereas Reardon et al.21 utilized a method based on hydrogen evolution measurements combined with SEM and X-ray powder diffraction. To shed light on the above-presented discrepancies between our experimental data and previously published results, we have performed a complementary experiment using X-ray powder diffraction: the nZVI sample in the form of a dense slurry was repeatedly scanned under ambient conditions (Figure 3). From the resulting XRD patterns we can confirm that magnetite may be formed as a product of fast oxygenation of metastable iron(II) hydroxide by air oxygen during sample pretreatment or

Figure 1. Representative 57Fe Mössbauer spectra of solid reaction species collected at various reaction times during the reaction of nZVI with water at 25 °C under anaerobic conditions. Mössbauer spectra were measured at 200 K in zero applied magnetic field. Key: green, sextet of α-Fe; blue, doublet of Fe(OH)2.

Figure 2. Time evolution of Mössbauer spectral areas corresponding to nZVI and Fe(OH)2 during anaerobic reaction of nZVI with water at 25 °C. Lines represent the fitted reaction kinetics model.

ments also showed that reaction 1 saturated after 2000 h, with ∼30% of elemental iron remaining as a solid fraction (Figure 2). We suggest that saturation of the reaction kinetics is due to nZVI surface passivation and the extensive coverage of the nZVI surface by Fe(OH)2 (see Figure 4a,b and text below). Detailed microscopic investigation (i.e., using a combination of SEM + EDX and TEM) of the reacted samples (Figure 4a− c) revealed the formation of flat layered crystals extensively overgrowing nanoparticles of metallic iron. Such sheetlike 13820

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

Figure 4. Representative scanning electron microscopy image (a), transmission electron microscopy image (b), and energy-dispersive X-ray spectrum (the spectrum was taken from sheets using SEM) of zerovalent iron nanoparticles reacted with water at 25 °C for 3192 h under anaerobic conditions. The inset in the TEM image shows the layered structure of Fe(OH)2 sheets. (d) Representative transmission electron microscopy image of zerovalent iron nanoparticles reacted with water at 80 °C for 211.8 h under anaerobic conditions.

The quantitative data derived from detailed Mössbauer spectroscopy monitoring of the nZVI reaction with water under anaerobic conditions at 80 °C are plotted in Figure 6. The rate constant determined for nZVI corrosion to Fe(OH)2 at 80 °C was 6.8 × 10−3 h−1 (±1.4 × 10−3), and the calculated reaction half-life was 101 h. Evidently, the primary reaction 1 is considerably accelerated and exhibited a nearly 1 order of magnitude higher rate constant when the temperature was increased from 25 to 80 °C. However, subsequent to the formation of Fe(OH) 2 according to reaction 1, transformation of the metastable iron(II) hydroxide to more stable magnetite (Fe3O4) clearly takes place (see the increased content of magnetite sextets in Figure 5 at the expense of the Fe(OH)2 doublet) with increasing time at 80 °C according to the so-called Schikorr reaction:46,47

even during XRD/SEM measurements (see below) at less than 1 h. Therefore, this measurement not only uniquely confirmed the validity of our experimental approach but also clearly demonstrated that extremely careful experimental control has to be applied during sample preparation and measurement when dealing with nZVI particles as primary reaction products could be quickly transformed into more stable secondary phases (i.e., Fe(OH)2 to Fe3O4 as in this case). Contrary to macroscopic iron or steel, Fe(OH)2 formed on the nZVI surface is in the form of “nanosheets” (see Figure 4), being more susceptible to fast spontaneous transformation. Mechanism and Kinetics at 80 °C: Experimental Observation. With the aim to decipher the evolution of reaction products, we conducted the nZVI reaction with water at a temperature of 80 °C under anaerobic conditions. Similar to the experiments performed at 25 °C, the frozen-state transmission 57Fe Mössbauer spectra of the samples reacted at 80 °C and collected at various reaction times (Figure 5) exhibited the same sextet of α-Fe (for a data-collection temperature of 200 K, δ = 0.05 mm s−1, εQ = 0.00 mm s−1, and Bhf = 33.1 T) and one paramagnetic doublet assigned to Fe(OH)245 (for a data-collection temperature of 200 K, δ = 1.20 mm s−1 and ΔEQ = 2.92 mm s−1; see Table S2 in the Supporting Information). With increasing reaction time, two new sextets with hyperfine parameters typical of tetrahedral iron(III) (Bhf = 48.2 T) and octahedral iron(II)/iron(III) atoms (Bhf = 45.1 T) in a magnetite spinel structure became more pronounced (see the gray subspectra in Figure 5 and Table S2).

3Fe(OH)2 → Fe3O4 + 2H 2O + H 2

(2)

The kinetic curves in Figure 6 suggest that both reactions 1 and 2 proceed consecutively under anaerobic conditions at 80 °C, where reaction 2 represents the rate-limiting step. As a result of reaction 2, the isometric globular-shaped magnetite nanoparticles surround nanoparticles of metallic iron (Figure 4d) together with remnants of larger Fe(OH)2 lamellae. Theoretical Analysis of the Reaction Mechanism at 25 °C: Identification of the Rate-Limiting Step. In this section, we compare the experimental data for the nZVI reaction with water at 25 °C under anaerobic conditions with theoretical calculations. The experimental results indicated that a solid Fe(OH)2 phase was the primary product of the reaction 13821

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

The energy differences among species localized along the reaction profile were calculated at the CCSD(T)-3s3p-DKH/ CBS level, which included almost full electron correlation, eliminated a large portion of spin contamination, predicted the correct ground-state spin multiplicity, involved scalar relativistic effects, and was free of basis set superposition error as the complete basis set was used. This method was carefully benchmarked against reference experimental data obtained in our previous studies, which showed that it provided results that were in good agreement with experiments.27,39 The free energy profiles indicated two separate reaction steps (Figure 7 and Table S3 in the Supporting Information), i.e. Fe0 + H 2O → HFe IOH

(3)

followed by HFe IOH + H 2O → Fe II(OH)2 + H 2

Figure 5. Representative 57Fe Mössbauer spectra of solid reaction species collected at various reaction times during the reaction of nZVI with water at 80 °C under anaerobic conditions. Mössbauer spectra were measured at 200 K in zero applied magnetic field. Key: green, sextet of α-Fe; blue, doublet of Fe(OH)2; gray, two sextets of Fe3O4 (i.e., tetrahedral A and octahedral B positions of iron in the magnetite structure).

(4)

Figure 7. Reaction scheme showing the quintet (multiplicity M = 5) reaction path of the Fe + 2H2O reaction in the gas phase (black) and including solvation effects (red). Calculations at the highest level (CCSD(T)-3s3p-DKH/CBS) were used. Values for the gas-phase reaction were taken from the work by Karlicky et al.31

Both steps represent one-electron transfers, as the oxidation number of iron is increased by 1. Analysis of the orbitals provided for quintet Fe(OH)2 yielded the effective electronic atomic configuration Fe(4s0.33d6.34p0.2)−2O(2s1.82p5.3)−2H(1s0.5) compared to the quintet ground state of the Fe(4s23d6) atom. The effective electronic configuration of quintet HFeOH, H(1s1.5)−Fe(4s0.53d6.24p0.2)−O(2s1.82p5.3)−H(1s0.5), is very similar on both Fe and O to that of Fe(OH)2. The formation of HFeOH involves the transfer of 4s electrons from iron to the 1s orbital of the hydrogen atom bound to the Fe and O 2p shell donated by the Fe and the H in the OH group. Thus, the charge on the two hydrogen atoms differs by nearly one electron. The first reaction step, i.e., splitting of the H−OH bond of water and formation of a HFeOH molecule, was predicted to be the rate-limiting step with a corresponding activation barrier ΔG⧧298K(aq) of 31.6 kcal/mol. The activation barrier of the second step, which generates Fe(OH)2 and hydrogen, was calculated as 16.2 kcal/mol. Thus, for the first time, we identified HFeOH as an important primary reaction intermediate in the reaction of zerovalent iron with water. It is worth noting that the existence of the HFeOH molecule has previously been demonstrated experimentally for other systems by argon matrix isolation FTIR spectroscopy.51,52

Figure 6. Time evolution of Mössbauer spectral areas corresponding to nZVI, Fe(OH)2, and Fe3O4 during anaerobic reaction of nZVI with water at 80 °C. Lines represent fitted reaction kinetics models according to the consecutive reaction model formalism.

between nZVI and water. To elucidate a detailed reaction mechanism and identify the elementary reaction steps, quantum chemical calculations were employed. We considered the reaction of a single iron atom (Fe0) with water as our first model system (see also the Materials and Methods) because such a model allows application of quantum chemical methods providing energy changes with high thermochemical accuracy. 13822

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

was unraveled as the reaction intermediate (eqs 3 and 4, Figure 7). The first step of eq 3 was proven as the rate-limiting step corresponding to the break of the H−O bond of the water molecule. As the second calculated barrier was significantly lower and the corresponding second reaction was significantly faster, HFeOH was not confirmed experimentally by Mössbauer spectroscopy (Figures 1 and 2). The theoretical computations also delimit the expected activation barrier for the reaction. Both activation barriers (ΔG⧧,atom = 31.6 kcal/mol for an iron atom, ΔG⧧,surf = 18.0 kcal/mol for an infinite periodic surface) may be considered as (upper/lower) limits for the real reaction barriers observed for nZVI; nZVI itself contains too large a number of atoms for modeling. General agreement of ΔG⧧ with the measurement and the order ΔG⧧,atom > ΔG⧧,nZVI > ΔG⧧,surf consistent with preliminary data on small iron clusters26 confirmed the suggested reaction mechanism. Such a joint experimental and theoretical study provided data on the reactivity of nZVI which may also help to identify shortcomings of current methods and models; hence, theoretical description of nZVI and its reactivity is still challenging.26

The reactions of a water molecule with infinite iron surfaces are in many respects analogous to the water molecule reaction with a single Fe atom.31,34 Previously, we identified that the HFeOH intermediate exists also on the Fe(100) and Fe(111) surfaces.34 However, the corresponding activation barrier (ΔE⧧,surf at the RPA level) of 14.6 kcal/mol31 was significantly lower than the value obtained for the atomic system (ΔE⧧ = 29.2 kcal/mol at the CCSD(T)-3s3p-DKH/CBS level).31 After ΔG corrections obtained from the Fe7 cluster model (Figure S2 in the Supporting Information), the final surface activation barrier ΔG⧧,surf298K(aq) was 18.0 kcal/mol. This implies that the experimental value ΔG⧧,nZVI298K(aq) = 26 kcal/mol lies between the values from both extreme computational models (i.e., an iron atom and an infinite iron surface). The computations on an infinite iron surface suggested that the primary reaction products of one water molecule were H and O atoms which stayed adsorbed on the iron surface.34 Such a picture of the reaction seems to contradict the experimentally observed evolution of gas-phase H2 during the reaction of nZVI with water.21 To resolve this issue, we studied the dependency of the adsorption energies of hydrogen and oxygen atoms on the coverage of the Fe(111) surface. The (111) surface is very open and corrugated, mimicking surface facets characteristic for nanoparticles. We compared the energies of two states, (i) O and H atoms adsorbed on the surface and (ii) O atoms on the surface and free H2 molecule, as a function of gradually increased surface coverage. The results indicate that while at low coverages the H atoms indeed prefer to stay on the iron surface, increased oxygen coverage makes the desorption of H atoms more likely (Figure 8).



CONCLUSIONS



ASSOCIATED CONTENT

The experimentally investigated complex mechanism of nZVI reaction with water under anaerobic conditions at 25 °C revealed saturation of pseudo-first-order reaction kinetics, which can be explained by surface passivation of nZVI and its extensive coverage by Fe(OH)2 sheets (the only solid reaction product at this temperature). For the first time, we identified HFeOH as the primary reaction intermediate in the reaction of zerovalent iron with water and thus probably a crucial reductant in the degradation of pollutants present in groundwater. The final formation of ferrous hydroxide (Fe(OH)2) is accompanied by hydrogen evolution (as confirmed by theoretical calculations), providing an additional reductant in the aqueous system. The experimentally measured rate constant of nZVI reaction with water was 1.14 × 10−3 h−1. The corresponding value of the activation Gibbs energy barrier ΔG⧧,nZVI298K(aq) = 26 kcal/mol lies between values from two computational models, i.e., between ΔG⧧298K(aq) = 31.6 kcal/mol (for reaction with an iron atom) and ΔG⧧,surf298K(aq) = 18.0 kcal/ mol (for reaction with an infinite iron surface). Finally, nZVI reaction with water under anaerobic conditions at 80 °C was considerably accelerated and exhibited a nearly 1 order of magnitude higher rate (6.8 × 10−3 h−1) and magnetite as the dominant reaction product. The effects of particular solid reaction intermediates/products on the overall reactivity of nZVI under anaerobic conditions will be the subject of ongoing research.

Figure 8. Adsorption energies (kcal/mol) of oxygen and hydrogen species on Fe(111) as a function of the oxygen coverage of the surface and the surface geometries of the most stable configuration at each coverage.

Oxygen atoms fill surface adsorption sites and push hydrogen to less favorable positions. At full surface coverage, the formation of a H2 molecule becomes energetically preferred and the surface is covered with an oxygen layer with a regular 1 × 1 pattern. Although the adsorption energies in Figure 8 decrease with increasing coverage, the oxygen layer is still thermodynamically stable (compared to the reference state, which is an isolated Fe surface and the respective number of H2O molecules). The theoretical calculations show that hydrogen molecules are released during anaerobic corrosion, in agreement with experimental observations.21 The computations provided detailed insight into the initial steps of the reaction of water with nZVI. Two one-electron steps were identified as the elementary reactions, and HFeOH

S Supporting Information *

Inclusive characterization of nZVI particles (X-ray powder diffraction pattern, 57Fe Mössbauer spectrum, scanning electron microscopy image, and particle size distribution), reaction scheme, 57Fe Mössbauer hyperfine parameters, values of energies, Gibbs energy differences from quantum chemical calculations, and complete ref 38. This material is available free of charge via the Internet at http://pubs.acs.org. 13823

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C



Article

supported zerovalent iron nanoparticles in the remediation of aqueous metal contaminants. Chem. Mater. 2001, 13, 479−486. (14) Ponder, S. M.; Darab, J. G.; Mallouk, T. E. Remediation of Cr(VI) and Pb(II) aqueous solutions using supported, nanoscale zerovalent iron. Environ. Sci. Technol. 2000, 34, 2564−2569. (15) Su, C. M.; Puls, R. W. Arsenate and arsenite removal by zerovalent iron: Kinetics, redox transformation, and implications for in situ groundwater remediation. Environ. Sci. Technol. 2001, 35, 1487− 1492. (16) Diao, M. H.; Yao, M. S. Use of zero-valent iron nanoparticles in inactivating microbes. Water Res. 2009, 43, 5243−5251. (17) Maršaĺ ek, B.; Jančula, D.; Maršaĺ ková, E.; Mashlan, M.; Šafárǒ vá, K.; Tuček, J.; Zbořil, R. Multimodal action and selective toxicity of zerovalent iron nanoparticles against cyanobacteria. Environ. Sci. Technol. 2012, 46, 2316−2323. (18) Klímková, S.; Č erník, M.; Lacinová, L.; Filip, J.; Jančík, D.; Zbořil, R. Zero-valent iron nanoparticles in treatment of acid mine water from in situ uranium leaching. Chemosphere 2011, 82, 1178− 1184. (19) Li, L.; Fan, M. H.; Brown, R. C.; Van Leeuwen, J. H.; Wang, J. J.; Wang, W. H.; Song, Y. H.; Zhang, P. Y. Synthesis, properties, and environmental applications of nanoscale iron-based materials: A review. Crit. Rev. Environ. Sci. Technol. 2006, 36, 405−431. (20) Keenan, C. R.; Sedlak, D. L. Factors affecting the yield of oxidants from the reaction of manoparticulate zero-valent iron and oxygen. Environ. Sci. Technol. 2008, 42, 1262−1267. (21) Reardon, E. J.; Fagan, R.; Vogan, J. L.; Przepiora, A. Anaerobic corrosion reaction kinetics of nanosized iron. Environ. Sci. Technol. 2008, 42, 2420−2425. (22) Sarathy, V.; Tratnyek, P. G.; Nurmi, J. T.; Baer, D. R.; Amonette, J. E.; Chun, C. L.; Penn, R. L.; Reardon, E. J. Aging of iron nanoparticles in aqueous solution: Effects on structure and reactivity. J. Phys. Chem. C 2008, 112, 2286−2293. (23) Linnenbom, V. J. The Reaction between Iron and Water in the Absence of Oxygen. J. Electrochem. Soc. 1958, 105, 322−324. (24) Liu, Y. Q.; Lowry, G. V. Effect of particle age (Fe-o content) and solution pH on NZVI reactivity: H-2 evolution and TCE dechlorination. Environ. Sci. Technol. 2006, 40, 6085−6090. (25) Nurmi, J. T.; Tratnyek, P. G.; Sarathy, V.; Baer, D. R.; Amonette, J. E.; Pecher, K.; Wang, C. M.; Linehan, J. C.; Matson, D. W.; Penn, R. L.; Driessen, M. D. Characterization and properties of metallic iron nanoparticles: Spectroscopy, electrochemistry, and kinetics. Environ. Sci. Technol. 2005, 39, 1221−1230. (26) Karlický, F.; Otyepka, M. Challenges in the Theoretical Description of Nanoparticle Reactivity: Nano Zero-Valent Iron. Int. J. Quantum Chem. 2014, DOI: 10.1002/qua.24627. (27) Karlický, F.; Otyepka, M. First step in the reaction of zerovalent iron with water. J. Chem. Theory Comput. 2011, 7, 2876−2885. (28) Yang, Y.; Weaver, M. N.; Merz, K. M. Assessment of the “631+G**+LANL2DZ” mixed basis set coupled with density functional theory methods and the effective core potential: Prediction of heats of formation and ionization potentials for first-row transition-metal complexes. J. Phys. Chem. A 2009, 113, 9843−9851. (29) DeYonker, N. J.; Peterson, K. A.; Steyl, G.; Wilson, A. K.; Cundari, T. R. Quantitative computational thermochemistry of transition metal species. J. Phys. Chem. A 2007, 111, 11269−11277. (30) Eshuis, H.; Bates, J. E.; Furche, F. Electron correlation methods based on the random phase approximation. Theor. Chem. Acc. 2012, 131. (31) Karlický, F.; Lazar, P.; Dubecký, M.; Otyepka, M. Random phase approximation in surface chemistry: Water splitting on iron. J. Chem. Theory Comput. 2013, 9, 3670−3676. (32) Ginovska-Pangovska, B.; Camaioni, D. M.; Dupuis, M. About the barriers to reaction of CCl4 with HFeOH and FeCl2. J. Phys. Chem. A 2011, 115, 8713−8720. (33) Parkinson, G. S.; Dohnalek, Z.; Smith, R. S.; Kay, B. D. Reactivity of Fe0 atoms, clusters, and nanoparticles with CCl4 multilayers on FeO(111). J. Phys. Chem. C 2009, 113, 1818−1829.

AUTHOR INFORMATION

Corresponding Authors

*Phone: +420 585634764. Fax: +420 585634761. E-mail: [email protected]. *Phone: +420 585634947. Fax: +420 585634761. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support from the Czech Science Foundation (Grant GACR P208/11/P463), the Technology Agency of the Czech Republic, Competence Centers (Project TE01020218), Operational Program Education for Competitiveness, European Social Fund (Projects CZ.1.07/2.3.00/20.0017 and CZ.1.07/ 2.3.00/20.0056 of the Ministry of Education, Youth and Sports of the Czech Republic), the Ministry of Industry and Trade (Project FR-TI3/622), and Operational Program Research and Development for Innovations, European Regional Development Fund (Project CZ.1.05/2.1.00/03.0058 of the Ministry of Education, Youth and Sports of the Czech Republic), is gratefully acknowledged. We thank Ivo Medřı ́k, Klára Šafárǒ vá, and Jiřı ́ Pechoušek (all from the Regional Centre of Advanced Technologies and Materials, Faculty of Science, Palacký University, Olomouc, Czech Republic) for technical assistance.



REFERENCES

(1) Tratnyek, P. G.; Johnson, R. L. Nanotechnologies for environmental cleanup. Nano Today 2006, 1, 44−48. (2) Zhang, W. X. Nanoscale iron particles for environmental remediation: An overview. J. Nanopart. Res. 2003, 5, 323−332. (3) Li, X. Q.; Elliott, D. W.; Zhang, W. X. Zero-valent iron nanoparticles for abatement of environmental pollutants: Materials and engineering aspects. Crit. Rev. Solid State 2006, 31, 111−122. (4) Cundy, A. B.; Hopkinson, L.; Whitby, R. L. D. Use of iron-based technologies in contaminated land and groundwater remediation: A review. Sci. Total Environ. 2008, 400, 42−51. (5) Amonette, J. E.; Sarathy, V.; Linehan, J. C.; Matson, D. W.; Wang, C.; Nurmi, J. T.; Pecher, K.; Penn, R. L.; Tratnyek, P. G.; Baer, D. R. Chemistry of metallic iron nanoparticles. Geochim. Cosmochim. Acta 2005, 69, A263−A263. (6) Arnold, W. A.; Roberts, A. L. Pathways and kinetics of chlorinated ethylene and chlorinated acetylene reaction with Fe(O) particles. Environ. Sci. Technol. 2000, 34, 1794−1805. (7) Gillham, R. W.; Ohannesin, S. F. Enhanced degradation of halogenated aliphatics by zero-valent iron. Ground Water 1994, 32, 958−967. (8) Kim, J. H.; Tratnyek, P. G.; Chang, Y. S. Rapid dechlorination of polychlorinated dibenzo-p-dioxins by bimetallic and nanosized zerovalent iron. Environ. Sci. Technol. 2008, 42, 4106−4112. (9) Nam, S.; Tratnyek, P. G. Reduction of azo dyes with zero-valent iron. Water. Res. 2000, 34, 1837−1845. (10) Sayles, G. D.; You, G. R.; Wang, M. X.; Kupferle, M. J. DDT, DDD, and DDE dechlorination by zero-valent iron. Environ. Sci. Technol. 1997, 31, 3448−3454. (11) Wang, C. B.; Zhang, W. X. Synthesizing nanoscale iron particles for rapid and complete dechlorination of TCE and PCBs. Environ. Sci. Technol. 1997, 31, 2154−2156. (12) Miehr, R.; Tratnyek, P. G.; Bandstra, J. Z.; Scherer, M. M.; Alowitz, M. J.; Bylaska, E. J. Diversity of contaminant reduction reactions by zerovalent iron: Role of the reductate. Environ. Sci. Technol. 2004, 38, 139−147. (13) Ponder, S. M.; Darab, J. G.; Bucher, J.; Caulder, D.; Craig, I.; Davis, L.; Edelstein, N.; Lukens, W.; Nitsche, H.; Rao, L. F.; Shuh, D. K.; Mallouk, T. E. Surface chemistry and electrochemistry of 13824

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825

The Journal of Physical Chemistry C

Article

(34) Lazar, P.; Otyepka, M. Dissociation of water on iron surfaces: generalized gradient functional and range-separated hybrid functional study. J. Phys. Chem. C 2012, 116, 25470−25477. (35) Hamprecht, F. A.; Cohen, A. J.; Tozer, D. J.; Handy, N. C. Development and assessment of new exchange-correlation functionals. J. Chem. Phys. 1998, 109, 6264−6271. (36) Balabanov, N. B.; Peterson, K. A. Systematically convergent basis sets for transition metals. I. All-electron correlation consistent basis sets for the 3d elements Sc-Zn. J. Chem. Phys. 2005, 123, 064107. (37) Foresman, J. B.; Keith, T. A.; Wiberg, K. B.; Snoonian, J.; Frisch, M. J. Solvent effects. 5. Influence of cavity shape, truncation of electrostatics, and electron correlation ab initio reaction field calculations. J. Phys. Chem. 1996, 100, 16098−16104. (38) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009. (39) Karlický, F.; Otyepka, M.; Schroeder, D. Ligand effects on single-electron transfer of isolated iron atoms in the gaseous complexes [(OC)mFe(OH2)n]+ (m, n = 0−2, m + n = 1, 2). Int. J. Mass Spectrom. 2012, 330−332, 95−99. (40) Blochl, P. E. Projector augmented-wave method. Phys. Rev. B 1994, 50, 17953−17979. (41) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 1999, 59, 1758−1775. (42) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1997, 78, 1396 (erratum). Original paper: Phys. Rev. Lett. 1996, 77, 3865−3868. (43) Lim, D. H.; Lastoskie, C. M. Density functional theory studies on the relative reactivity of chloroethenes on zerovalent iron. Environ. Sci. Technol. 2009, 43, 5443−5448. (44) Zbořil, R.; Andrle, M.; Opluštil, F.; Machala, L.; Tuček, J.; Filip, J.; Marušaḱ , Z.; Sharma, V. K. Treatment of chemical warfare agents by zero-valent iron nanoparticles and ferrate(VI)/(III) composite. J. Hazard. Mater. 2012, 211, 126−130. (45) Cornell, J. M.; Schwertmann, U. The Iron Oxides; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2003; p 156. (46) Daub, K.; Zhang, X.; Wang, L.; Qin, Z.; Noel, J. J.; Wren, J. C. Oxide growth and conversion on carbon steel as a function of temperature over 25 and 80 °C under ambient pressure. Electrochim. Acta 2011, 56, 6661−6672. (47) Ma, M.; Zhang, Y.; Guo, Z. R.; Gu, N. Facile synthesis of ultrathin magnetic iron oxide nanoplates by Schikorr reaction. Nanoscale Res. Lett. 2013, 8, 1−7. (48) Genin, J. M.; Bauer, P.; Olowe, A. A.; Rezel, D. Mossbauer Study of the Kinetics of Simulated Corrosion Process of Iron in Chlorinated Aqueous Solution around Room Temperature: The Hyperfine-Structure of Ferrous Hydroxides and Green Rust I. Hyperfine Interact. 1986, 29, 1355−1360. (49) Reardon, E. J. Zerovalent irons: Styles of corrosion and inorganic control on hydrogen pressure buildup. Environ. Sci. Technol. 2005, 39, 7311−7317. (50) Reinsch, B. C.; Forsberg, B.; Penn, R. L.; Kim, C. S.; Lowry, G. V. Chemical Transformations during Aging of Zerovalent Iron Nanoparticles in the Presence of Common Groundwater Dissolved Constituents. Environ. Sci. Technol. 2010, 44, 3455−3461. (51) Kauffman, J. W.; Hauge, R. H.; Margrave, J. L. Studies of reactions of atomic and diatomic Cr, Mn, Fe, Co, Ni, Cu, and Zn with molecular water at 15 K. J. Phys. Chem. 1985, 89, 3541−3547. (52) Zhang, L. N.; Zhou, M. F.; Shao, L. M.; Wang, W. N.; Fan, K. N.; Qin, Q. Z. Reactions of Fe with H2O and FeO with H2. A combined matrix isolation FTIR and theoretical study. J. Phys. Chem. A 2001, 105, 6998−7003.

13825

dx.doi.org/10.1021/jp501846f | J. Phys. Chem. C 2014, 118, 13817−13825