Environ. Sci. Technol. 2008, 42, 2451–2456
Immobilization of Selenite on Fe3O4 and Fe/Fe3C Ultrasmall Particles RAQUEL LÓPEZ DE ARROYABE LOYO,† S E R G E I I . N I K I T E N K O , * ,† ANDREAS C. SCHEINOST,‡ AND MONIQUE SIMONOFF† Chimie Nucléaire Analytique et Bioenvironnementale, Université de Bordeaux I,II - CNRS, BP 120 Le Haut Vigneau, 33175 Gradignan, France, and Institute of Radiochemistry and Rossendorf Beamline at ESRF, FZD, 01314 Dresden, Germany
Received October 11, 2007. Revised manuscript received December 19, 2007. Accepted December 20, 2007.
The sorption of selenite ions onto Fe3O4 and Fe/Fe3C nanoparticles (NPs) was studied in aqueous solutions under anoxic conditions using γ spectrometry and X-ray absorption spectrometry (XAS) techniques. This is the first study related to the remedial applications of Fe/Fe3C NPs. Fe3O4 NPs have been prepared by conventional coprecipitation of Fe(II) and Fe(III) in basic solutions. Stable Fe/Fe3C NPs have been prepared by Fe(CO)5 sonicating in diphenylmethane solutions and subsequently annealing the as-prepared product. Kinetic study demonstrated that Se(IV) sorption is extremely rapid: the equilibrium is reached in approximately 10 and 30 min for Fe3O4 and Fe/Fe3C NPs, respectively, at pH ) 4.9-5.1 in solutions of 0.1 M NaCl. The distribution coefficients are also very high for both kinds of NPs (Kd > 3000). Increasing the pH to 10.3 or adsorption of organic ligands, like L-lysine or dodecanoate, at the surface of NPs causes the decrease in Kd values. However, even in these cases Kd values exceed 150. Magnetic NPs loaded with selenium can be easily and completely removed from solution with a 0.4 T permanent magnet. XAS study revealed the absence of Se(IV) reduction during the sorption onto Fe3O4 NPs in the pH range of 4.8–8.0. By contrast, the removal of Se(IV) with Fe/Fe3C NPs in anaerobic conditions occurs via Se(IV) reduction to Se(-II) and subsequent formation of iron selenide at the particle surface. Thus, the Fe/Fe3C NPs are superior to Fe3O4 NPs due to their ability to immobilize rapidly and irreversibly Se(IV) via reductive mechanism. Presumably these particles could be also effective for the removal of other contaminants such as hexavalent chromium, actinides, technetium, and toxic organic compounds.
Introduction Although selenium is an essential trace element it is toxic if taken in excess. Selenium poisoning of water systems may result when new agricultural runoff courses through normally dry undeveloped lands enriched in selenium. This process leaches natural soluble selenium compounds, such as selenite (SeO32-) or selenate (SeO42-), into the water, which may then be concentrated in new “wetlands” as it evaporates (1). Oil refinery and mining wastewater often contain high levels of selenium (2). * Corresponding author phone: (33)557120905; fax: (33)557120900; e-mail:
[email protected]. † CNAB, Université de Bordeaux I,II - CNRS. ‡ Institute of Radiochemistry and Rossendorf Beamline at ESRF. 10.1021/es702579w CCC: $40.75
Published on Web 02/20/2008
2008 American Chemical Society
Moreover, isotope 79Se is one of the long-lived radioactive fission products (β emitter, t1/2 ) 4.8 × 105 yr). Release of this radionuclide from waste depositories to the biosphere through transport in underground water could contribute to the total cumulative dose of radioactivity (3). The treatment of selenium in waters is an important challenge because the high mobility of selenium anionic species. According to the U.S. EPA report (FRL-5649-7) the acute toxicity of Se(IV) is almost 10 times superior to that of Se(VI) and both species exist simultaneously in aerobic surface water often in comparable concentrations. Iron oxides, such as hematite (4, 5), goethite (6–8), and magnetite (9, 10), are known to be capable of binding SeO32- anions in groundwater with high distribution coefficients. In general, the sorption of selenium species onto iron oxides proceeds according to the surface complexation mechanism (4–9). However, in the case of magnetite, the presence of Fe(II) can contribute to reduce the selenium to a lower oxidation state. Recently it was shown that Fe3O4 reduces Tc(VII), Np(V) (11), and Pu(V) (12) in the pH range of 3–8. Preliminary X-ray absorption spectroscopy (XAS) study of selenite adsorption onto magnetite in oxic conditions indicated that the oxidation state of selenium did not change during the sorption at pH ) 4.45-7.46 (10). Thus, the possibility of Se(IV) reduction by magnetite is still unclear. In contrast to Fe3O4, the immobilization of selenium with zerovalent iron (ZVI) occurs via reductive mechanism. X-ray photoelectron spectroscopy (XPS) studies revealed that selenocyanate (13) and selenate (14) anions are removed by ZVI filings from wastewater through the formation of Se(0)/ Se(-II) or Se(IV)/Se(0), respectively, even in the presence of oxygen. The drawback of macro- or micrometer size range iron oxides and ZVI as a sorbent for selenium is related to the slow rate of sorption. The time needed for completion of the sorption of selenium species onto magnetite and ZVI reported by several authors was about 10 h (8, 9, 13). It is well-known that for the same mechanism and the similar experimental conditions the rate of sorption is directly proportional to the specific surface area of the sorbent particles. Consequently, reducing particle size is expected to greatly enhance the reaction rate. Recently it was reported that a steady-state for Pd(II), Rh(III), Pt(IV) (15), and Co(II) (16) adsorption onto magnetite nanoparticles (NPs) with a particle size of ∼10 nm is reached in less than 20 min. ZVI nanoparticles stabilized with carboxymethyl cellulose were shown to be effective for in situ immobilization of Cr(VI) in contaminated soils (17). However, about 90% of chromium can be removed only after ∼36 h of reaction. The slow reaction rate was attributed to the surface coating of ZVI with organic stabilizing agents. To date, there is no published data on selenium sorption onto iron or iron oxide NPs. The objectives of this work were the following: (i) to study the kinetics of selenite sorption onto Fe3O4 and Fe/Fe3C NPs, and (ii) to determine the chemical forms of selenium adsorbed at the particles surface. First, the paper is focused on a comparative study of Fe3O4 and recently developed airstable Fe/Fe3C NPs (18, 19) in the removal of selenite anions from aqueous solutions. Second, the possibility of selenium reduction with Fe3O4 and Fe/Fe3C NPs was examined using XAS. To our knowledge this is the first study related to the toxic element’s immobilization with Fe/Fe3C NPs. In principal, these NPs can be more effective than polymer protected ZVI particles since the iron carbide layer at the surface does not prevent the sorption of ionic species. In this study several experiments on selenium sorption at Fe3O4 NPs were VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2451
TABLE 1. Particle Size and Magnetic Properties of Fe3O4 and Fe/Fe3C NPs Measured at Room Temperaturea particles b
Fe3O4 Fe/Fe3Cc
particle size, nm
Ms, emu g-1
Hc, Oe
6(2 50 ( 30
39 (F ) 4.6 T) 202
8 40
a Ms ) saturation magnetization. Hc ) coercivity. 97.7 m2 g-1; data from ref 16. c Data from ref 18.
b
S )
performed in the presence of sodium dodecanoate (DC) and L-lysine (Lys) to evaluate the effect of organic ligands adsorption on selenium immobilization with magnetic NPs. Finally, the magnetically assisted chemical separation (MACS) was tested to remove selenium from aqueous solutions using magnetic nanoparticles. The MACS process has the advantages of less complex equipment and more efficient separation compared to conventional column-based techniques (20).
Experimental Section Magnetic Nanoparticles Synthesis and Characterization. All chemicals were used as received and of highest purity. Solutions were prepared using deionized water (Milli-Q 18 MF). Magnetite NPs were obtained by coprecipitation of Fe(II) and Fe(III) in basic solutions in argon as described elsewhere (21). The only difference of our method was the application of ultrasonic irradiation at 20 kHz instead of mechanical agitation (for details, see the Supporting Information). Precipitation under power ultrasound allows the preparation of ultrasmall particles with narrow size distribution (22, 23). Magnetite NPs coated with DC or Lys were prepared by stirring the Fe3O4 with a 0.1 M solution of corresponding ligands for 1 h at pH ) 9 in argon. Ligand excess was then removed by washing with 0.1 M NaCl. FT-IR spectra collected with a Spectrum One spectrometer (Perkin-Elmer) in KBr pellets revealed the presence of absorption bands at 2923 cm-1 (CH2 as), 2853 cm-1 (CH2 s), 788, 888 cm-1 (C-C-C δ), and 1478 cm-1 (CdO ν) for DC and 2923 cm-1 (CH2 as), 2853 cm-1 (CH2 s), 1490 cm-1 (CdO ν), and 1070, 1118 cm-1 (C-N ν) for Lys, respectively, indicating the adsorption of these ligands at the particle surface (the FT-IR spectra are presented in Supporting Information). Nanoparticles Fe/Fe3C have been prepared by sonicating Fe(CO)5 in diphenylmethane solutions in argon and subsequently annealing the amorphous asprepared product in argon at 700 °C as described elsewhere (18, 19) (for details, see the Supporting Information). Table 1 summarizes the data on particle size and magnetic properties of prepared materials. TEM images obtained with a Philips SM120 electron microscope are shown in Supporting Information. Magnetic measurements were performed at room temperature with a QPMS SQUID magnetometer. TEM measurements reveal that both kinds of particles are nanocrystalline. The magnetite particles have a mean size of 6 nm with a narrow size distribution. The Fe/Fe3C particles are bigger than magnetite NPs since their synthesis was performed by annealing at relatively high temperature. Their mean size is about 50 nm, and the maximal size does not exceed 100 nm. Very small coercivity of Fe3O4 NPs compared to the bulk magnetite (Hc ) 120–400 Oe) (24) indicates the superparamagnetic behavior of these particles. The Fe/Fe3C particles are ferromagnetic with high magnetization (Ms), close to that of bulk iron, and relatively low coercivity typical for the soft magnetic materials. It is worth noting that prepared Fe/Fe3C particles are stable at least for several days when in contact with air or aerated water. Sorption of Selenite Ions on Fe3O4 and Fe/Fe3C NPs. Sorption of selenite was studied in batch experiments performed in a nitrogen filled glovebox at 25 °C. A total of 2452
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
40 mg of solid was mixed and then mechanically agitated with 8 mL of 0.1 M NaCl solution (V/M ) 200) containing 1 × 10-6 M SeO32- labeled with 75Se. Solutions were purged with argon prior experiment. At various time intervals, a reaction vial was sacrificed for sampling. The solid phase was centrifuged at 6000 rpm. Selenium concentration in solution was then determined by means of conventional γ spectrometry (Eγ ) 264.5 keV). The values of pH were adjusted with NaOH or HCl solutions and controlled with a pH-meter after the sorption. In several experiments the solid phase was removed with the NdFeB permanent magnet (0.4 T) to investigate the feasibility of the MACS process with studied particles. The observed distribution coefficient values, Kd, were determined by eq 1: Kd ) (Ci/Cf - 1)(V/M)
(1)
where V is the volume of solution (mL), M is the mass of solid (g), Ci is the initial stock solution counts before sorption, and Cf is the final solution counts after sorption. XAS Speciation of Selenium Adsorbed on NPs. Samples were prepared in a nitrogen filled glovebox by mixing 100 mg of synthesized materials and natural selenite solution ([Se(IV)]i ) 5 × 10-5 M, 0.1 M NaCl, V ) 25 mL, V/M ) 250) to obtain about 1000 ppm (1 mg Se/1 g sorbent) of selenium adsorbed at the particle surface. The solids were separated with permanent magnet after ∼10 h of stirring at room temperature, and then the slurries were placed in sealed holders for XAS measurements. Preliminary experiments with radioactive selenite showed that about 90% of selenium is removed from solutions at these conditions. Great care was taken to exclude O2 during sample transport, storage, and measurement. X-ray absorption near-edge (XANES) and extended X-ray absorption fine-structure (EXAFS) spectra were collected at the Rossendorf Beamline at ESRF (Grenoble, France). The energy of the X-ray beam was tuned by a double crystal monochromator operating in channel-cut mode using a Si(111) crystal pair. Two platinum-coated Si mirrors before and after the monochromator were used to collimate the beam into the monochromator and to reject higher harmonics. A 13 element high-purity germanium detector (Canberra) together with a digital signal processing unit (XIA) was used to measure fluorescence spectra. Spectra were collected at 15 K using a closed-cycle He cryostat with a large fluorescence exit window and a low vibration level (CryoVac). For energy calibration, a gold foil (K-edge at 11919 eV) was chosen because of its greater inertness and hence reliability in comparison to Se. With this approach, we determined an edge energy of 12656 eV for trigonal Se, instead of the tabulated value of 12658 eV for zerovalent Se. Data in the XANES region were collected in steps of 0.5 eV, i.e., with better precision than the resolution of the Si(111) crystal at the given vertical divergence (1.7 eV) and the broadening due to the core-hole lifetime (2.3 eV). A comparison of single scans of the same sample showed an accuracy of better than 0.5 eV. Dead time correction of the fluorescence signal, energy calibration, and the averaging of single scans were performed with the software package SixPack (http://www.stanford.edu/ ∼swebb). Normalization, transformation from energy into k space, and subtraction of a spline background was performed with WinXAS using routine procedures (25). The EXAFS data were fit using theoretical backscattering amplitudes and phase shifts calculated with FEFF 8.2 (26). Shell fitting was done in WinXAS. A synthetic FeSe sample (Alfa Aesar), which consisted predominately of tetragonal form, and a trigonal gray Se(0) obtained from the Paris School of Mines Mineralogy Museum were used as a XAS references. As Se(IV) reference, an aqueous solution of Na2SeO3 · 5H2O (Merck) at pH ) 6.0 was used.
TABLE 3. Selenium Leaching from Fe3O4 and Fe/Fe3C NPs with 0.1 M NaCl at pH = 10.1, 25 °C, Ar Atmosphere Se leached, %a
a
FIGURE 1. Uptake percentage of selenium as a function of contact time. [Se(IV)]i ) 1 × 10-6 M, 0.1 M NaCl, (V/M) ) 200, 25 °C. (() Fe3O4, pH ) 4.9. (0) Fe/Fe3C, pH ) 5.1. (purple () Fe3O4/DC, pH ) 5.6. (b) Fe3O4/Lys, pH ) 5.2. (2) Fe3O4, pH ) 10.3.
TABLE 2. Kinetic and Thermodynamic Data for the Se(IV) Sorption onto Fe3O4 and Fe/Fe3C NPsa Kd, mL g-1
pH
Fe3O4 Fe3O4 (MS)b 1 h of contact 5 h of contact Fe3O4 (d < 5 µm)c Fe/Fe3C Fe/Fe3C (MS)b 1 h of contact 5 h of contact Fe3O4 Fe3O4/Lys Fe3O4/DC
4.9
10
0.14
3670 ( 280
4.9 4.9 4.2 5.1
1200 30
0.0015 0.046
3580 ( 310 3490 ( 320 1000 3810 ( 420
20 10 10
0.07 0.14 0.14
3700 ( 310 3640 ( 380 170 ( 12 388 ( 32 410 ( 39
5.1 5.1 10.3 5.2 5.6
teq, min
ks, min-1
particles
a [Se(IV)]i ) 1 × 10-6 M, 0.1 M NaCl, (V/M) ) 200, 25 °C. MS ) magnetic separation. c Data from ref 9; S ) 0.89 m2 g-1, [Se(IV)]i ) 3 × 10-6 M, 0.1 M NaCl, (V/M) ) 200, 25 °C. b
The XAS spectrum of ferroselite, orthorhombic FeSe2, was provided by Dan Strawn, University of Idaho.
Results and Discussion Selenium Sorption Kinetics. Figure 1 demonstrates the kinetics of Se(IV) sorption onto Fe3O4 and Fe/Fe3C NPs. These data indicate that the sorption is very rapid. The equilibrium is reached in approximately 10 and 30 min for Fe3O4 and Fe/Fe3C NPs, respectively, at pH ) 4.9-5.1. Table 2 summarizes the Kd, equilibrium time, teq, and estimated pseudofirst-order sorption rate constant, ks, values obtained at different experimental conditions. The values of ks were calculated using eq 2 presuming first-order reaction kinetics: ks ) 0.693/t1/2
(2)
where t1/2 is the sorption half-time. A very high rate of sorption does not allow the treatment of the kinetic curves in the semilogarithmic plots. The equilibrium time values for Se(IV) sorption are similar to those for Pd(II), Rh(III), Pt(IV) (15), and Co(II) (16) sorption onto Fe3O4 NPs with the particle size in the range of 5–20 nm. Somewhat less rapid sorption at Fe/Fe3C NPs if compared with Fe3O4 NPs is, probably, related to the difference in particle size. In contrast to the sorption rate, the Kd values are very close to each other for Fe3O4 and Fe/Fe3C NPs. Comparison of the data obtained at similar conditions for Fe3O4 NPs and micrometer size range magnetite particles (MPs) reveals a dramatic acceleration of Se(IV) sorption and
time, min
Fe3O4 NPs
Fe/Fe3C NPs
15 30 60
32 65 67
4 7 7
Statistic error is about 5%.
also a very significant increase of its distribution coefficients in the systems with ultrasmall particles. Undoubtedly, this is attributed to the large difference in particle size. It is worth noting that the values of specific surface area (S) obtained for solid powders with the BET technique can be used only as a rough approximation for quantitative comparison of kinetic data for NPs and MPs. In the absence of an external magnetic field, noncoated magnetic particles are known to be aggregated in colloidal solutions due to the magnetic dipole–dipole interactions (24). The number of single particles in aggregates increases with the increase of particle size. Consequently, the ratio of “specific surface areas” for NPs and MPs of magnetite should be higher in colloids than that obtained for powdered sorbents. Lysine and dodecanoate adsorbed at Fe3O4 NPs have no visible effect on the sorption kinetics but cause the decrease of the distribution coefficient for Se(IV). This result is of practical interest since it demonstrates the efficiency of Fe3O4 NPs in selenium removal even in the presence of organic ligands often present in the environment or in waste streams, like amino acids and detergents. The Kd and ks values also decrease with increasing pH. It should be emphasized that the effect of pH on the kinetic and thermodynamic parameters of Se(IV) sorption is more noticeable than what is observed in the case of ligand adsorption. The strong effect of pH on Se(IV) sorption at the NPs is in concordance with the general behavior of Se(IV) in the presence of iron oxide MPs (9). However, Se(IV) sorption on Fe3O4 NPs is still very rapid (teq ) 20 min) even at pH ) 10.3. As it can be seen from Table 2, the Kd values obtained by means of NPs magnetic separation are very close to those obtained with centrifugation as a phase separation technique. That confirms the high efficiency of MACS process for selenium removal from aqueous waste solutions. The low particle-mass-to-solution ratio and the excellent kinetics of Se(IV) sorption allow this process to compete with more traditional decontamination techniques, such as solvent extraction and ion exchange. It is important to mention that no significant loss of magnetization or selenium affinity of Fe3O4 and Fe/Fe3C NPs can be observed in the studied experimental conditions. Data in Table 2 show that Kd values obtained with MACS technique remain stable even after 5 h of phases contact. The results from selenium leaching experiments with basic 0.1 M NaCl solutions are shown in Table 3. At pH ) 10.1 about 67% of adsorbed selenium can be leached from Fe3O4 NPs for 30 min. In contrast, adsorbed selenium is not stripped from Fe/Fe3C NPs at the same conditions. The question arising then is whether this difference takes place due to the difference in stability of inner sphere Se(IV) complexes at the surface of NPs or whether there is a difference in the chemical forms of selenium adsorbed onto Fe3O4 and Fe/ Fe3C NPs. This was studied by XAS technique. XAS Speciation of Selenium Adsorbed on NPs. Figure 2 shows the XANES spectra at selenium K-edge for Se(IV) adsorbed onto Fe3O4 NPs at different pH values. The strong similarity of the edges in these spectra with that of aqueous VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2453
FIGURE 3. Se K-edge XANES of Se(IV) adsorbed to Fe/Fe3C NPs in comparison to Se(IV), Se(0), and Se(-II) references. The XANES suggests reduction to Se(-II).
FIGURE 2. Se K-edge XANES of Se(IV) sorbed to Fe3O4 NPs at three different pH values. The white line energy is in line with Se(IV) but not with elemental Se, indicating that no reduction of selenite occurred after sorption to magnetite. selenite indicates that the oxidation state of Se(IV) did not change due to the sorption in the pH range of 4.8–8.0. The selenium k3-weighted EXAFS signal of the sorption sample at pH 7.9, presented in the Supporting Information, shows a single sine-like oscillation, corresponding to a backscattering shell at a distance of 1.4 Å before correcting for the phase shift. This shell could be fitted by three O atoms at a distance of 1.70 Å, in line with the structure of Se(IV) compounds. After reduction to Se(0) or Se(-II), one would observe a significant shift of the coordination shell to longer distances, since Se and Fe coordination shells in a variety of solids vary between 2.3 to 2.6 Å. Since no significant backscattering contribution in this range could be found, this is clear evidence that no reduction of Se(IV) took place. The same conclusion applies for the EXAFS spectra at pH 5.6 and 4.8 (data not shown). In contrast to Fe3O4, sorption of Se(IV) onto Fe/Fe3C NPs causes significant changing in the selenium oxidation state (Figure 3). The XANES absorption edge (12654.8 eV) is 5.5 eV lower than that of aqueous selenite, 1.3 eV lower than that of trigonal Se(0), and similar to that of FeSe (12655.2 eV). At the same time, the white line is reduced in intensity in comparison to Se(IV), suggesting an increased population of the valence 4p levels. Both features are in line with a complete reduction to Se(-II). Thus it can be concluded that the removal of selenite ions with Fe/Fe3C NPs at anaerobic conditions occurs via Se(IV) reduction to Se(-II). According to the Fe-Se phase diagram (27), at room temperature there 2454
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
are two homogeneous and stable phases, FeSe and FeSe2. In principal, both kinds of iron selenides can be formed after selenite ion reduction with Fe/Fe3C NPs. However, formation of the diatomic anionic species Se22- appears to be less probable in our experiments due to the low selenium concentration. This is further supported by EXAFS analysis, presented in the Supporting Information. The EXAFS spectrum and the corresponding Fourier transform of the Fe/ Fe3C NPs sample are more similar to those of FeSe than to those of Se(0) and FeSe2. A multishell fit of the spectrum achieved the following structural data: 3.3 Fe @ 2.42 Å, 1.8 Se @ 3.46 Å, and 3.6 Se @ 3.98 Å, which are different from those of tetragonal FeSe: 4 Fe @ 2.37 Å, 8 Se @ 3.77 Å, and 4 Se @ 3.91 Å. The relatively small coordination numbers suggest formation of ultrasmall FeSe particles with a structure somewhat different from that of bulk crystalline FeSe. In conclusion, the XANES and EXAFS analyses suggest that the strong resistance of adsorbed selenium to the leaching in the case of Fe/Fe3C NPs can be undoubtedly attributed to selenium reduction and formation of iron selenide at the particle surface. Sorption Mechanisms for Fe3O4 and Fe/Fe3C NPs. The rate of Se(IV) removal with the studied NPs appears to be very fast. This observation indicates that at the first stage the selenium is removed by Fe3O4 as well as by Fe/Fe3C NPs largely through the surface adsorption, which are limited by selenite species diffusion and thus exhibit fast reaction kinetics. The modeling of the thermodynamic data on Se(IV) sorption onto Fe3O4 MPs have led to the conclusions that the sorption edge coincides with the predominance of HSeO3and that Se(IV) forms two inner-sphere complexes at the magnetite surface (9): dFe-OH + HSeO3- T (dFeOHSeO3)2- + H+ -
dFe-OH + HSeO3 T (dFeHSeO3) + H2O
(3) (4)
It seems to be reasonable to assume that the mechanisms of Se(IV) sorption at NPs and MPs are quite similar. The much higher Kd values for NPs can be attributed to the higher specific surface area in the case of nanosized sorbent. As was mentioned above, increasing of pH or adsorption of organic ligands causes the drop in Kd for Se(IV). At the same time the
adsorption kinetics is not influenced by pH. The effects of pH or organic acids can be interpreted by the inner-sphere model presuming that the sorption of selenite anions occurs at the coordination sites which are not occupied by OH-, Lys-, or DC- anions. The replacement of these species at the magnetite surface by Se(IV) is less probable since the complexes of selenite ions with iron cations are much weaker than those with OH- or organic acid anions (28). It is interesting to note that the DC- anion with a C12 aliphatic chain hinders selenium sorption approximately to the same extent as the relatively small Lys- anion. This observation is in a good agreement with the observation that fatty acids form nonuniform self-assembled monolayers at the surface of iron oxides (29). Consequently, adsorption of dodecanoate is not able to isolate the available coordination sites at the surface of Fe3O4 NPs. XANES measurements clearly show the absence of Se(IV) reduction during the sorption onto Fe3O4 NPs in the pH range of 4.8–8.0. A table available in the Supporting Information summarizes the redox potentials for some selenium and iron reactions. From these data it follows that magnetite cannot reduce Se(IV) in basic solutions since its Eh potential is more positive than that of selenite. By contrast, reduction of Se(IV) with Fe3O4 or Fe(II) cations is thermodynamically possible in acid medium. Presumably, reduction of Se(IV) with magnetite is very slow even in acidic solutions. It is known that the reduction of selenate and selenite ions with Fe(II) is kinetically inhibited in the homogeneous solutions, but selenate reduction may occur with interlayered green rust, a Fe(II)-Fe(III) hydroxide sulfate (30, 31). Green rust has an unusual crystal structure, with positively charged oxide layers that create interlayer sites where anions such as SeO42- are loosely bonded. Its high redox reactivity is related to the presence of ordered Fe(II) and Fe(III) cations in the hydroxide layers, which allows rapid electron transfer. Structure of magnetite (inverse spinel) is very different from that of green rust. Presumable Se(IV) reduction can take place only at the interface. Probably, in this case electron transfer is inhibited like for homogeneous reaction. The standard electrode potentials for HSeO3-/Se(0) and Se(0)/HSe- redox couples (see Supporting Information) are more positive than that for the Fe(II)/Fe(0) couple, indicating that Se(IV) can be reduced to Se(-II) by Fe(0) in the studied pH range. The effective reduction of Se(IV) with Fe/Fe3C and the stability of Kd values and magnetization in time indicate that these nanoparticles are rather stable at the studied conditions. However, the surface oxidation of metallic iron cannot be excluded. It is well-known that iron corrosion in water in an anoxic environment is initiated by the following process (32): Fe(0) + 2H2O f Fe(II) + H2 + 2OH-
(5)
Then Fe(II) hydrolysis and oxidation causes formation of magnetite and iron oxyhydroxides at the metal surface (33). Nevertheless, the electrochemical study (34) revealed that the corrosion potential of Fe(0) NPs are more negative than that of iron oxides or microsized Fe(0). Our data on Se(IV) reduction confirm that the surface corrosion does not inhibit the reductive properties of Fe(0) NPs. Finally, batch experiments demonstrated that Se(IV) was rapidly and effectively removed from the aqueous solutions with Fe3O4 or Fe/Fe3C ultrasmall particles in the pH range of 4-10 even in the presence of organic ligands adsorbed at the particle surface. Both kinds of NPs coupled with the MACS process have proven to be a promising method for the wastewater remediation containing selenite ions. The Fe/ Fe3C NPs are of particular interest due to their ability to immobilize rapidly selenium via reductive mechanism. Presumably these particles could be also effective for the removal of other contaminants such as hexavalent chromium,
actinides, technetium, and nitro- and chloro- aromatic compounds by the similarity with the effect of other Fe(0) NPs and ZVI surface (14, 34). It is worth noting that commercially available Fe(0) NPs are usually synthesized by the high temperature reduction of iron oxides with H2 or by the reductive precipitation with NaBH4 (34). The cost of Fe/ Fe3C NPs obtained by sonochemistry seems to be competitive with these conventional processes. Moreover, the sonochemical equipment has been considerably improved during the past decade. Modern industrial scale sonoreactors allow the treatment of 10 m3 of liquids per hour with an output power of about 16 kW (35). This sonoreactor can be used for the industrial production of nanomaterials.
Acknowledgments This research was supported by the ACTINET European Research Program and by the PARIS French Research Program. R.L.d.A.L. thanks the European ERASMUS Student Program for her fellowship. The authors thank Dr. Bernard Chevalier from ICSM, Pessac, France, for his help in SQUID measurements and Dr. Joséphine Lai Kee Him from the University of Bordeaux-I, France, for the assitance in TEM measurements. We acknowledge the help of Dr. Harald Funke during XAS data collection.
Supporting Information Available Details of Fe3O4 and Fe/Fe3C NPs syntheses, TEM images of prepared particles, FT-IR spectra of Fe3O4 NPs with adsorbed dodecanoate and L-lysine, the redox potentials for some selenium and iron reactions, and Se K-edge EXAFS spectrum and the corresponding Fourier transform of Se(IV) sorbed to magnetite and Fe/Fe3C NPs. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Zingaro, R. A.; Cooper, W. C. Selenium; Van Nostrand Reinhold: New York, 1974. (2) Meng, X.; Bang, S.; Korfiatis, G. P. Removal of selenocyanate from water using elemental iron. Water Res. 2002, 36, 3867– 3873. (3) Bouillis, B. Retraitement et séparation des radionucléides à vie longue. In Les déchets nucléaires; Turkey, R., Ed.; Lesititions de Physique: Paris, 1997; pp 69–91. (4) Duc, M.; Lefèvre, G.; Fédoroff, M. Sorption of selenite ions on hematite. J. Colloid Interface Sci. 2006, 298, 556–563. (5) Rovira, M.; Giménez, J.; Martinez, M.; Martinez-Llado, X.; de Pablo, J.; Marti, V.; Duro, L. Sorption of selenium(IV) and selenium(VI) onto natural iron oxides: goethite and hematite. J. Hazard. Mater. 2007, 10.1016/j.jhazmat.2007.04.098. (6) Zhang, P.; Sparks, D. L. Kinetics of selenate and selenite adsorption/desorption at the goethite/water interface. Environ. Sci. Technol. 1990, 24, 1848–1856. (7) Su, C.; Suarez, D. L. Selenate and selenite sorption on iron oxides; an infrared and electrophoretic study. Soil. Sci. Soc. Am. J. 2000, 64, 101–111. (8) Duc, M.; Lefevre, G.; Fedoroff, M.; Jeanjean, J.; Rouchaud, J. C.; Monteil-Rivera, F.; Dumonceau, J.; Milonjic, S. Sorption of selenium anionic species on apatites and iron oxides from aqueous solutions. J. Environ. Radioact. 2003, 70, 61–72. (9) Martinez, M.; Giménez, J.; de Pablo, J.; Rovira, M.; Duro, L. Sorption of selenium(IV) and selenium(VI) onto magnetite. Appl. Surf. Sci. 2006, 252, 3767–3773. (10) Missana, T.; Alonso, U.; Scheinost, A. C.; Granizo, N.; GarciaGutiérrez, M. Selenite retention by iron oxy-hydroxides: role of adsorption, reduction and dissolution/co-precipitation. MIGRATION’07, München, Germany, August 26–31, 2007; Abstracts, p 71. (11) Cui, D.; Eriksen, T. E. Reduction of Tc(VII) and Np(V) in solution by ferrous ion. A laboratory study of homogeneous and heterogeneous redox processes; SKB Technical Report TR-9603: Stockholm, Sweden, 1996. (12) Powell, B. A.; Fjeld, R. A.; Kaplan, D. I.; Coates, J. T.; Serkiz, S. M. Pu(V)O2+ adsorption and reduction by synthetic magnetite (Fe3O4). Environ. Sci. Technol. 2004, 38, 6016–6024. VOL. 42, NO. 7, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2455
(13) Meng, X.; Bang, S.; Korfiatis, G. P. Removal of selenocyanate from water using elemental iron. Water Res. 2002, 36, 3867– 3873. (14) Qiu, S. R.; Lai, H.-F.; Roberson, M. J.; Hunt, M. L.; Amrhein, C.; Giancarlo, L. C.; Flynn, G. W.; Yarmoff, J. A. Removal of contaminants from aqueous solution by reaction with iron surfaces. Langmuir 2000, 16, 2230–2236. (15) Uheida, A.; Iglesias, M.; Fontas, C.; Hidalgo, M.; Salvado, V.; Zhang, Yu.; Muhammed, M. Sorption of palladium(II), rhodium (III), and platinum(IV) on Fe3O4 nanoparticles. J. Colloid Interface Sci. 2006, 301, 402–408. (16) Uheida, A.; Salazar-Alvarez, G.; Björkman, E.; Yu, Z.; Muhammed, M. Fe3O4 and γ-Fe2O3 nanoparticles for the adsorption of Co(II) from aqueous solution. J. Colloid Interface Sci. 2006, 298, 501– 507. (17) Xu, Y.; Zhao, D. Reductive immobilization of chromate in water and soil using stabilized iron nanoparticles. Water Res. 2007, 41, 2101–2108. (18) Nikitenko, S. I.; Koltypin, Yu.; Palchik, O.; Felner, I.; Xu, X. N.; Gedanken, A. Synthesis of highly magnetic, air-stable iron-iron carbide nanocrystalline particles by using power ultrasound. Angew. Chem., Int. Ed. 2001, 40, 4447–4449. (19) Nikitenko, S. I.; Koltypin, Yu.; Felner, I.; Yeshurun, I.; Shames, A. I.; Jiang, J. Z.; Markovich, V.; Gorodetsky, G.; Gedanken, A. Tailoring the properties of Fe-Fe3C nanocrystalline particles prepared by sonochemistry. J. Phys. Chem. B 2004, 108, 7620– 7626. (20) Kaminski, M. D.; Nunez, L. Extractant-coated magnetic particles for cobalt and nickel recovery from acidic solution. J. Magn. Magn. Mater. 1999, 194, 31–36. (21) Kim, D. K.; Mikhaylova, M.; Zhang, Yu.; Muhammed, M. Protective coating of superparamagnetic iron oxide nanoparticles. Chem. Mater. 2003, 15, 1617–1627. (22) Suslick, K. S.; Price, G. J. Application of ultrasound to materials chemistry. Annu. Rev. Mater. Sci. 1999, 29, 295–326. (23) Gedanken, A. Using sonochemistry for the fabrication of nanomaterials. Ultrason. Sonochem. 2004, 11, 47–55.
2456
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 7, 2008
(24) Cullity, B. D. Introduction to magnetic materials; AddisonWesley Publishing: Reading, MA, 1972. (25) George, G. N.; Pickering, I. J. EXAFSPAK, a suite of computer programs for analysis of X-ray absorption spectra; Stanford Synchrotron Radiation Laboratory: Stanford, CA, 2000. (26) Ankudinov, A. L.; Ravel, B.; Rehr, J. J.; Conradson, S. D. Realspace multiple-scattering calculation and interpretation of x-ray absorption near-edge structure. Phys. Rev. B. 1998, 58, 7565– 7576. (27) Okamoto, H. Iron-Selenium Binary Diagram. ASM Phase Diagrams; 1991; http://www.asminternational.org/. (28) NIST. Critically Selected Stability Constants of Metal Complexes: Version 8. 2007; http://www.nist.gov/srd/nist46.htm. (29) Ulman, A. Formation and structure of self-assembled monolayers. Chem. Rev. 1996, 96, 1533–1554. (30) Myneni, S. C. B.; Tokunaga, T. K.; Brown, G. E., Jr. Abiotic selenium redox transformations in the presence of Fe(II, III) oxides. Science 1997, 278, 1106–1109. (31) Johnson, T. M.; Bullen, T. D. Selenium isotope fractionation during reduction by Fe(II)-Fe(III) hydroxide-sulfate (green rust). Geochim. Cosmochim. Acta 2003, 67, 413–419. (32) Ponder, S. M.; Darab, J. G.; Bucher, J.; Caulder, D.; Craig, I.; Davis, L.; Edelstein, N.; Lukens, W.; Nitsche, H.; Rao, L.; Shuh, D. K.; Mallouk, T. E. Surface chemistry and electrochemistry of supported zerovalent iron nanoparticles in the remediation of aqueous metal contaminants. Chem. Mater. 2001, 13, 479–486. (33) Furukawa, Y.; Kim, J.-W.; Watkins, J.; Wilkin, R. T. Formation of ferrihydrite and associated iron corrosion products in permeable reactive barriers of zero-valent iron. Environ. Sci. Technol. 2002, 36, 5469–5475. (34) Nurmi, J. T.; Tratnyek, P. G.; Sarathy, V.; Baer, D. R.; Amonette, J. E.; Pecher, K.; Wang, C.; 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. (35) Hielscher Ultrasound Technology. http://www.hielscher. com/ultrasonics/industry.htm.
ES702579W
