Introduction of Cobalt Ions in γ-Fe2O3 Nanoparticles by Direct

Article pubs.acs.org/JPCC

Introduction of Cobalt Ions in γ‑Fe2O3 Nanoparticles by Direct Coprecipitation or Postsynthesis Adsorption: Dopant Localization and Magnetic Anisotropy Charlotte Vichery,† Isabelle Maurin,*,† Olivier Proux,‡,§ Isabelle Kieffer,‡,§ Jean-Louis Hazemann,‡,∥ Robert Cortès,† Jean-Pierre Boilot,† and Thierry Gacoin† †

Physique de la Matière Condensée, UMR7643, CNRSEcole Polytechnique, 91128 Palaiseau, France BM30B/FAME beamline, European Radiation Synchrotron Facility, 38043 Grenoble cedex 9, France § Observatoire des Sciences de l’Univers de Grenoble, UMS 832, CNRSUniversité Joseph Fourier, 38041 Grenoble cedex 9, France ∥ Institut Néel, UPR 2940, CNRSUniversité Joseph Fourier, 25 avenue des Martyrs, BP 166, 38042 Grenoble cedex 9, France ‡

S Supporting Information *

ABSTRACT: The influence of cobalt doping on the magnetic anisotropy of γ-Fe2O3 nanoparticles has been investigated using two different approaches: (i) simultaneous precipitation of Fe2+, Fe3+, and Co2+ precursors in water and (ii) adsorption of Co2+ ions onto the surface of preformed iron oxide particles followed by diffusion in the solid phase upon heat treatment. The incorporation of small amounts of Co dopants, less than 1 at %, was monitored by magnetization measurements combined with X-ray absorption spectroscopy experiments at the Co K-edge. These latter measurements were carried out in fluorescence mode using a crystal analyzer spectrometer for an enhanced sensitivity. Analyses of the X-ray absorption fine structures allowed for unraveling the differences in local atomic structure and valence state of Co in the two series of samples. A thermally activated diffusion in the spinel lattice was observed in the 250−300 °C range, leading to a substantial increase in magnetocrystalline anisotropy. At higher annealing temperature, magnetic anisotropy was still found to increase due to an enhanced surface contribution associated with dehydroxylation of terminal Fe atoms. This study not only provides direct correlations between magnetic anisotropy and dopant localization in Co doped γ-Fe2O3 but also demonstrates for the first time that simultaneous coprecipitation of Fe2+, Fe3+, and Co2+ may actually lead to heterogeneous doping, with a significant part of the Co dopants adsorbed at the particle surface.

■

INTRODUCTION

because it controls the maximum achievable SLP but also modulates the influence of particle size distribution.10 From an experimental point of view, several investigations have been carried out to tune the SLP coefficient, at fixed chemical composition, by changing the particle size, d.5,6,11,12 They showed a sharp maximum of SLP for 15 nm γ-Fe2O3 particles at a field frequency of 700 kHz.6,11 Nevertheless, for such small sizes, the contribution of surface anisotropy to the effective anisotropy constant (K) is important,13 so both particle size and magnetic anisotropy were actually modified in these studies. An alternative approach, which separates the K and d variables, is to tune the magnetic anisotropy through doping at fixed particle size. For instance, it is well-known that insertion of cobalt into the spinel structure of maghemite (γ-Fe2O3) or magnetite (Fe3O4) leads to an increased magnetic anisotropy.14−16 The anisotropy constant of cobalt ferrite CoFe2O4 is

Iron oxide nanoparticles have been extensively studied for biomedical applications, as contrast agents for magnetic resonance imaging or new vectors for drug delivery, tissue engineering, or cancer therapy using hyperthermia.1−3 To reach optimum efficiency, each application requires particles with tailored saturation magnetization and magnetic anisotropy. These parameters are strongly influenced by the chemical composition, size, and shape distributions but also by the crystallinity4,5 of the particles. For instance, recent theoretical studies have shown that the specific loss power (SLP), which measures the heating produced by the relaxation of magnetic particles placed in a rf magnetic field, can present a marked dependence on particle size in the single-domain regime, in the case of dominant Néel contribution.6,7 Designing nanoparticles with a narrow size distribution is thus mandatory for hyperthermia therapeutics and optimized protocols have been developed such as seed-assisted growth and heating-up processes.8,9 Magnetic anisotropy is another central parameter for the optimization of particles for magnetic hyperthermia © 2013 American Chemical Society

Received: June 3, 2013 Revised: August 26, 2013 Published: August 26, 2013 19672

dx.doi.org/10.1021/jp405450p | J. Phys. Chem. C 2013, 117, 19672−19683

The Journal of Physical Chemistry C

Article

about 2 orders of magnitude larger than the one of γ-Fe2O3 and Fe3O4. Cobalt doping of Fe3O4 or γ-Fe2O3 nanoparticles can be achieved either by simultaneous reaction of Co and Fe precursors15−17 or by postsynthesis diffusion of the dopant into preformed iron oxide particles.18−20 To do so, one method consists of first synthesizing magnetite nanoparticles by coprecipitation in water. The corresponding colloid or its oxidized form (γ-Fe2O3) is then refluxed in a solution of Co2+ ions in order to promote the immobilization of the dopants at the surface of the iron oxide.18,19 This latter treatment was also shown to induce a full oxidation of the iron oxide.18 Although restricted to small doping ranges, this method ensures a fixed size for the final particles. The doping level can be tuned by the pH, contact time, or concentration of the Co2+ solution.21,22 Alternatively, a given amount of Co2+ ions can be adsorbed at the particle surface, and their progressive insertion into the spinel lattice can be controlled by annealing at increasing temperatures.18 To avoid any aggregation or sintering of the particles upon heating, particles have first to be dispersed into a refractory matrix such as sol−gel silica.23 A large (up to 3.5 times) and progressive increase of the magnetic anisotropy constant could be achieved, at very low Co doping level (less than 1 at %), upon annealing between 100 and 800 °C.18 These large changes presumably originate from the diffusion of Co into the γ-Fe2O3 spinel structure. Nevertheless, changes were not gradual and diffusion seemed to be activated in the 200− 300 °C temperature range. As modifications were still visible above 300 °C, open questions remain such as the valence state of Co and its possible changes upon heating. Besides, the diffusion process may imply a transfer via the less energetically favorable tetrahedral sites of the spinel structure, for which we expect depressed single-ion anisotropy.14,24 In this former study, no structural evidence for cobalt diffusion could be obtained from X-ray powder diffraction due to the very low dopant concentration.18 The aim of the present work is to assess the cobalt localization and its valence state in lightly doped γ-Fe2O3:Co nanoparticles prepared by surface treatment and monitor their evolution upon heating up to 800 °C. These issues have been addressed from X-ray absorption fine structure (XAFS) spectroscopy experiments performed at the Co K-edge. Cobalt doped maghemite particles, synthesized by direct coprecipitation of Co2+, Fe2+, and Fe3+ precursors and heat treated at similar temperatures, were also studied as references. The major difficulty for such investigations comes from the small amount of Co absorber, which requires detection in fluorescence mode, and the dilution of Co in a phase mainly formed of iron, as the Kβ fluorescence signal of Fe and the Kα fluorescence signal of Co are too close in energy to be separated by conventional solid state detectors.

another 30 min of stirring, the excess of HNO3 was discarded and 6 mL of a Co(NO3)2 solution at 1.5 M was added to the precipitate. The mixture was then heated to reflux for 30 min in order to oxidize magnetite into maghemite. The resulting particles were washed with a solution of HNO3 at 2 M and finally dispersed in HNO3 at pH 2. Four centrifugation steps at 11400g for 15 min were performed to lower the size polydispersity. Reference samples of Co doped maghemite (Fe2O3:Co, series 2) were synthesized by direct coprecipitation of Fe2+, Fe3+, and Co2+ ions in alkaline medium. Six milliliters of a KOH aqueous solution (13 M) was quickly poured into 21 mL of an acidic (HCl, 0.57 M) solution of ferrous, ferric, and cobalt chlorides ([Fe3+] = [Fe2+] + [Co2+] = 0.48 M, [Co2+]/[Fe2+] = 0.04). After 30 min of stirring, the black precipitate was magnetically decanted and washed with distilled water. A 2.45 mL volume of nitric acid (2 M) was added to the flocculate, and the mixture was stirred for 1 h. The excess of HNO3 was then discarded before the addition of 6 mL of a Fe(NO3)3 solution at 1.5 M. The mixture was heated to reflux for 30 min. As in series 1, the particles were washed with 2 M HNO3 and dispersed in HNO3 at pH 2. Two centrifugations at 11400g for 15 min were subsequently carried out. Undoped maghemite particles (γ-Fe2O3, series 3) were also used as a reference. Their synthesis is fully described in ref 25. The protocol involved coprecipitation of Fe2+ and Fe3+, similarly to what was detailed for series 1, except that the reflux was performed in the presence of Fe(NO3)3 instead of Co(NO3)2. Nanoparticles of series 1 and 2 were then dispersed into a sol−gel silica matrix to avoid their aggregation and growth during heat treatments. An acidic solution of tetraethoxysilane in ethanol was first hydrolyzed and condensed by heating at 60 °C for 1 h, as described in ref 25. Different amounts of magnetic colloids were added to the silica sol and the mixtures were dried at 100 °C, leading to powder materials with a large (ca. 0.08) or small (ca. 0.008) Fe/Si atomic ratio. The preparation of the Fe2O3/SiO2 composite samples (series 3) is detailed in ref 25. Powders were subsequently heated in air at different temperatures, ranging from 200 to 800 °C, for 1 h. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) measurements carried out on the concentrated composites yielded Co/Fe atomic ratios of 0.009 ± 0.002 for series 1 and 0.011 ± 0.002 for series 2. The distribution of particle sizes was determined by transmission electron microscopy (TEM) using a Philips CM30 microscope operating at 300 kV (analysis of ca. 400 particles). Note that all following experiments refer to composite samples, i.e., iron oxide nanoparticles dispersed within silica. Powder X-ray diffraction (XRD) patterns were recorded using a PANalytical X’Pert diffractometer equipped with Cu anticathode (λKα = 1.5418 Å). All measurements were carried out on concentrated composites (Fe/Si = 0.08) to have enough resolved diffraction profiles. An internal silicon calibrant was used for an accurate determination of the lattice parameter value. The structural coherence length was evaluated through peak profile analysis, as implemented in the Fullprof suite,26,27 after correction for the instrumental resolution function. Magnetization measurements were performed using Cryogenic SX600 and Quantum Design MPMS-5 SQUID magnetometers. Samples were mounted in polycarbonate capsules with the powder blocked by paraffin wax. Magnetic data were systematically corrected for the diamagnetic contributions of the silica matrix, wax, and polycarbonate

■

EXPERIMENTAL SECTION Cobalt doped maghemite nanoparticles (Fe2O3:Coads, series 1) were synthesized by a multistep protocol involving the precipitation of magnetite nanoparticles in alkaline medium, followed by a reflux treatment in the presence of cobalt nitrate. At first, 5.7 mL of ammonium hydroxide (13 M) was quickly added to 21 mL of an acidic (HCl, 0.57 M) solution of ferrous and ferric chlorides ([Fe3+] = [Fe2+] = 0.48 M) under vigorous stirring. After 30 min of stirring, the particles of magnetite were separated by magnetic decantation and washed with distilled water before addition of 2.45 mL of nitric acid at 2 M. After 19673



■

capsule. The fraction of magnetic material in each sample was derived from the iron content measured by ICP-AES. X-ray absorption spectroscopy experiments were performed at the Co K-edge on the FAME/BM30B beamline28 at the European Radiation Synchrotron Facility (ESRF), France. All measurements were collected at room temperature using a Si(220) pseudochannel-cut monochromator. A piece of Co foil was used for energy calibration (7709 eV, at the first inflection point of the absorption edge). For samples of series 1 and 2, about 30 mg of the silica composite (Fe/Si = 0.08) was ground with 10 mg of BN (Alfa Aesar, 99.5%) and then pressed into pellets. Data were collected in partial fluorescence mode for an improved sensitivity, because of the very low cobalt content in the composite samples (ca. 0.07 wt %). Fluorescence detection is classically employed to investigate dilute systems, by using solid state detectors, which have a typical energy resolution of 150−300 eV. However, in these conditions, it is impossible to discriminate in energy the Fe Kβ (7058 eV) and Co Kα1,2 (6930 and 6915 eV) fluorescence signals. As the samples under study contain 100 times more iron than cobalt, the Co fluorescence signal would be hidden by a 10 times higher background signal due to Fe fluorescence. Si(333) crystal analyzer spectrometers (CAS), with a theoretical energy resolution of 3 eV,29,30 were thus used to separate the iron Kβ signal from the Co Kα1 one, leading to a typical intensity ratio of 1.5 before and after the edge. A helium bag was inserted between sample and CAS to minimize absorption by air. However, higher resolution was reached at the expense of the signal-to-noise ratio, despite the use of four spherical bent crystals (0.5 m radius of curvature) in Johann geometry.30 Each spectrum was recorded using a 5 eV step between 7550 and 7700 eV, a 0.4 eV step up to 7760 eV, and a 0.05 Å−1 k-step for the EXAFS region (7760−8090 eV). Sixteen scans were averaged to get enough counting statistics for the EXAFS analyses, leading to a typical measurement time of 8 h with a flux of about 5 × 1011 photons/s/200 mA. XAFS spectra of reference samples in which the Co absorber is present in different valence states and local geometries were also recorded. The following standards were measured in transmission mode: CoFe2O4 (prepared by standard coprecipitation followed by annealing at 1200 °C), Co(NO3)2·6H2O (Fluka, ≥99%), Co3O4 (Strem Chemicals, 99.5%), and CoO.31,32 A spectrum of Co doped alumina (Al2O3:Co about 1 at %, prepared by annealing in air at 1000 °C a mixture of Al(OH)3, Fluka purum, and Co(NO3)2·6H2O) was also recorded in fluorescence mode, using a 30-element Ge solid state detector (Canberra). Data reduction was performed using the ATHENA 0.9 software package.33 The experimental XAFS spectra were corrected for pre-edge and post-edge absorption and subsequently normalized. Background was determined by a linear fit below the edge, and by a cubic spline above the edge. For EXAFS data, the spectra were windowed with a Hanning function over a k = 2.45−7.79 Å−1 range (dk = 5) and Fourier transformed with a k2-weighting. EXAFS analyses were performed using ARTEMIS.33 The k2χ(k) data were fitted in k-space with theoretical XAFS curves generated by the FEFF8 code.34 Note that the Fourier transformed functions shown in Results and Discussion were not corrected for phase shift, which leads to peak positions at shorter distances than real distances, typically by 0.5 Å.

Article

RESULTS AND DISCUSSION

Co Doping by Surface Treatment or Coprecipitation. The anisotropy constant of ca. 8 nm γ-Fe2O3 nanoparticles has been tuned through cobalt doping using two different approaches. The first strategy involves adsorption of Co2+ ions onto preformed magnetite particles, followed by further oxidation in maghemite and diffusion of the dopants in the bulk upon heating (series 1). The adsorption of metals and metalloids at the surface of iron oxides is well documented, as these nanomaterials are high-potential candidates in pollution control and waste management.35 Adsorption of Co2+ ions at the surface of magnetite and maghemite is controlled by different parameters, such as pH, contact time, concentration of Co2+ ions in solution, and temperature.21,22 For instance, Uheida et al. showed that the amount of Co ions adsorbed per unit surface area of maghemite and magnetite first increases when increasing the Co2+ concentration and then reaches a plateau for concentrations larger than 0.07 mol/L.21 Adsorption is also favored at high pH. For pH values lower than 5, no substantial adsorption occurs,21,22,36 whereas pH values larger than 8 lead to the precipitation of Co hydroxide.22 In the present study, the relatively high temperature and Co2+ excess used during the reflux treatment should have promoted a large Co uptake in spite of the reduced pH (5.5). However, the subsequent peptization step using concentrated HNO 3 solutions presumably induced a substantial leaching of the Co2+ ions.21 Co immobilization was nevertheless confirmed by elemental analysis performed on the silica composite, i.e., after the washing and peptization steps, with a Co/Fe atomic ratio of 0.009 ± 0.002. The second doping strategy used for the synthesis of Fe2O3:Co reference samples is based on a singlestep protocol by coprecipitation of Co2+, Fe2+, and Fe3+ precursors, followed by an oxidative treatment in solution (series 2). The amount of CoCl2 used in the coprecipitation step was adjusted to reach a final Co content similar to the one reported for series 1, Co/Fe = 0.011 ± 0.002. The shape and size distributions of the as-synthesized particles were characterized by TEM. Representative images of the particles are shown as Supporting Information (Figure S1). The corresponding histograms were fitted to log-normal functions, yielding a mean diameter (dm) of 8.3 nm and a standard deviation (σd) of 0.25 for the Fe2O3:Coads sample (series 1), or dm and σd values of 7.4 nm and 0.22 for coprecipitated Fe2O3:Co particles (series 2). These values compare well to the ones obtained for undoped particles (series 3), synthesized using a similar coprecipitation protocol: dm = 7.1 nm and σd = 0.28. The Bragg reflections for series 1 (Fe2O3:Coads/SiO2 composite samples annealed between 100 and 800 °C) and series 2 (Fe2O3:Co/SiO2, annealed in the same temperature range) were all consistent with the Fd3̅m space group of magnetite, maghemite, and cobalt ferrite. All Xray diffraction patterns are displayed in Supporting Information (Figure S2). Similarly to what was reported for undoped γFe2O3,25 the oxidative treatment carried out by reflux in slightly acidic conditions and in the presence of Co2+ or Fe3+ nitrates led to a full oxidation in maghemite. Indeed, the lattice parameter value of the various samples is in good agreement with the one of natural γ-Fe2O3 (a = 8.352 Å, JCPDS 39-1346) and departs from that of bulk magnetite (a = 8.396 Å, JCPDS 19-0629). According to the results shown in Table 1, the cell parameters remain almost constant upon annealing up to 800 °C in the two series. Nevertheless, one should only expect a Δa 19674



Article

dispersed in silica. Dilute composites (Fe/Si atomic ratio of 0.008) were selected, so magnetic dipole interactions and their change upon sintering of the SiO2 matrix could be neglected. This dilution corresponds to a volume fraction of ca. 0.3% in particles. In that case, the behavior of the particle assembly can be approximated by the sum of individual contributions instead of a collective behavior.38−40 Accordingly, the magnetic anisotropy constant (K) can be evaluated from the temperature of the maximum of the ZFC curve (Tpeak), which is characteristic of the transition between the blocked and superparamagnetic states. It is important to note that Tpeak is strictly equal to the blocking temperature (TB) only for an assembly of monodisperse particles. The larger the size polydispersity is, the more Tpeak will deviate from TB.41,42 The relation between Tpeak, TB, and σd has been previously determined in the case of 7 nm γ-Fe2O3 particles from simulations of the ZFC curves:25

Table 1. Structural Characterization of Series 1 and 2 temp (°C) 100 200 300 600 800 a

Fe2O3:Coads

Fe2O3:Co

aa (±0.003 Å) Lcb (±0.5 nm)

a (±0.003 Å) Lc (±0.5 nm)

8.347 8.349 8.349 8.347 8.348

7.4 7.5 7.3 7.9 8.2

8.352 8.350 8.344 8.344 8.348

5.4 5.8 5.5 6.9 7.1

Lattice parameter. bCoherence length.

= 0.0042x increase upon Co insertion in the CoxFe3‑xO4 solid solution, according to Sorescu et al.37 The cobalt content used in the present work (about 1 at %) is thus not large enough to induce any measurable change of the lattice parameter value. The structural coherence length values (Lc) are 7.4 ± 0.5 and 5.4 ± 0.5 nm, respectively, for the Fe2O3:Coads (series 1) and Fe2O3:Co (series 2) composites dried at 100 °C. The geometric sizes determined by TEM and Lc values are thus comparable, indicating that the particles are essentially monocrystalline. No significant changes in coherence length were reported for heat treatments below 600 °C (see Table 1). At higher annealing temperature, the slight increase in Lc could reflect an enhanced crystallinity or a very limited growth of the particles. This result confirms the efficiency of the confinement in the silica matrix to prevent grain coarsening through sintering and coalescence. The absence of transition into the αFe2O3 polymorph is also a clear indication of the good dispersion state of the particles in the matrix. Magnetic Study: Increased Anisotropy through Co Doping. The magnetic anisotropy constant of the Co doped samples of series 1 and 2 was determined from zero field cooled (ZFC) magnetization measurements. The corresponding plots are displayed in Figure 1. Magnetic measurements were all carried out on composite samples, with iron oxide nanoparticles

Tpeak TB

= 0.98 + 0.21e8.5σd

(1)

We also found from the simulations that eq 1 is not sensitive to small variations of the mean particle size, like the different dm values reported for series 1, 2, and 3 (undoped γ-Fe2O3 particles). In addition, the assumption of uniaxial anisotropy, which was used throughout the calculations, is probably still valid for low Co contents such as those investigated in the present work. For each sample, the value of TB was calculated from eq 1, by considering a standard deviation σd of 0.25 for series 1 and of 0.22 for series 2. The values of Tpeak, TB, and K = 25kBTB/V, where kB is the Boltzmann constant and V = πdm3/6 is the mean particle volume, are listed in Table 2. As shown in Figure 1, Tpeak and thus the effective anisotropy constant of the two Co doped series systematically increase upon thermal annealing. This behavior was expected for series 1, as a result of the insertion of Co2+ ions into the spinel structure, but not for the coprecipitated samples of series 2. This trend was nevertheless confirmed by the evolution of the coercive field values, measured at 10 K in the blocked state (Figure S3, Supporting Information). Indeed, μ0Hc increases from 140 to 2520 G for series 1 (Fe2O3:Coads) and from 1090 to 3380 G for series 2 (Fe2O3:Co), for heat treatments carried out between 100 and 800 °C. Note that Salazar-Alvarez et al. reported a 4-fold increase in coercivity upon Co2+ adsorption on γ-Fe2O3 nanoparticles after heat treatment at 90 °C,19 whereas little change of the μ0Hc value was observed before annealing in the present study. To further compare the evolution of the magnetic anisotropy upon annealing, the values of the blocking temperature were plotted on the same graph for the two Co doped series (Figure 2). TB values of undoped 7 nm γ-Fe2O3 particles (series 3) were also displayed as references.25 We chose to consider TB rather than K values in order to enhance the contribution of magnetocrystalline anisotropy. This is important as, for such small particles, a slight variation of diameter will give rise to large changes in anisotropy due to the dominant contribution of surface anisotropy. Indeed, in the case of iron oxides, surface anisotropy can be up to 2 orders of magnitude larger than the bulk contribution.13 To assess the localization of the Co dopants in series 1 after synthesis (Fe2O3:Coads, 100 °C), the magnetic properties of the corresponding sample were compared to those of undoped (γ-

Figure 1. Zero field cooled magnetization curves measured under 25 G for Fe2O3:Coads (series 1) and Fe2O3:Co (series 2) samples annealed between 100 and 800 °C. 19675



Article

Table 2. Magnetic Characterization of Series 1, 2, and 3 Fe2O3:Coads Tpeaka (±1 K) TBb (±0.5 K) Kc (±0.05 × 105 erg/cm3) μ0Hcd (±5 G) Fe2O3:Co Tpeak (±1 K) TB (±0.5 K) K (±0.05 × 105 erg/cm3) μ0Hc (±5 G) γ-Fe2O3 Tpeak (±1 K) TB (±0.5 K) K (±0.05 × 105 erg/cm3) μ0Hc (±5 G) a

100 °C 41 15.0 1.73 140 100 °C 120 51.3 8.34 1090

250 °C

275 °C

300 °C

90 32.8 3.78 670

112 40.9 4.71 1210

135 49.3 5.68 2020

200 °C

300 °C

132 56.4 9.17 1270

164 70.1 11.40 2580

600 °C 141 51.5 5.93 2400 600 °C 178 76.1 12.37 3260

90 °C

290 °C

540 °C

75 23 4.2 160

80 25 4.6 170

84 26 4.8 200

800 °C 145 52.9 6.10 2520 800 °C 183 78.2 12.72 3380 840 °C 102 31 5.7 260

Temperature of the maximum of the ZFC curve. bBlocking temperature. cAnisotropy constant. dCoercive field value at 10 K.

the very low Co content in the samples under study, one could anticipate a similar surface chemistry and thermal evolution for series 1 and 2. If the larger magnetic anisotropy reported for series 2 with respect to series 1 probably reflects a slightly higher dopant content (0.011 ± 0.002 compared to 0.009 ± 0.002), the large increase in magnetic anisotropy upon annealing was unexpected for these reference samples. Indeed, the Co2+ ions were simultaneously precipitated with the iron precursors and it is generally assumed that such a coprecipitation process leads to homogeneous doping. Two explanations could account for the large changes reported in the present study: (i) part of the cobalt ions were not incorporated in the particles but merely adsorbed at their surface at the end of synthesis or (ii) Co ions were fully inserted into the spinel structure after synthesis, but distributed over both tetrahedral and octahedral cationic sites. In that latter case, the increased anisotropy would result from structural rearrangements and to the displacement of some Co atoms from the less energetically favorable tetrahedral sites to the more anisotropic octahedral sites (KCo(octahedral) = +850 × 10−24 J and KCo(tetrahedral) = −79 × 10−24 J).24 An alternative explanation involves a change in the Co oxidation state, but would require a Co3+ → Co2+ reduction upon heating to agree with an increased anisotropy. In order to discriminate between these scenarios, X-ray absorption experiments were carried out at the Co K-edge to obtain direct information on the valence state and local structure around the Co ions in the two series. XAFS Study: Co Localization in Series 1 (Fe2O3:Coads). To clarify the localization of the Co dopants in the Fe2O3:Coads sample directly obtained after surface treatment (series 1, 100 °C), we first compared its X-ray absorption fine structures with those of the same sample after thermal annealing and with Co(NO3)2·6H2O and CoFe2O4 bulk standards. In the former reference sample, Co2+ ions are 6-fold coordinated to oxygen atoms of water molecules, whereas in the CoFe2O4 reference, Co2+ ions are mostly located in the octahedral sites of a spinel structure. The corresponding XAFS spectra are displayed in Figure 3a. For the as-synthesized sample (Fe2O3:Coads, 100 °C), the features observed above the main absorption edge are clearly consistent with those found in the Co2+ aquo complex. When focusing on the X-ray absorption near edge structures (XANES), similarities are even more pronounced (Figure 3b). K-edge XANES spectra are typically subdivided into an intense

Figure 2. Blocking temperature versus annealing temperature for the Fe2O3:Coads (1) and Fe2O3:Co (2) series and for undoped particles (series 3). Lines are guides for the eyes.

Fe2O3, series 3) and of coprecipitated (Fe2O3:Co, series 2) samples treated at the same temperature. For series 1, the values of the coercive field (μ0Hc) and of the blocking temperature (TB) were found to be comparable to the ones of undoped particles, whereas they strongly depart from those of the coprecipitated sample (Figure 2 and Table 2). This result suggests that Co is either adsorbed at the particle surface or located within the first few atomic layers of the outer shell, which is commonly considered to be a dead magnetic layer mainly composed of weakly coupled and misaligned spins.43,44 As a consequence, the presence of Co slightly affects the overall magnetic properties before heat treatment. In contrast, large changes were reported for both μ0Hc and TB after annealing in the 200−300 °C range, which should originate from thermally activated diffusion of Co in the bulk. The cusp at about 300 °C should thus be representative of a full incorporation of the dopants. Surprisingly, the blocking temperature was still found to increase above 300 °C, but at a much reduced rate (Figure 2). It is interesting to note that the evolution of TB between 300 and 800 °C for the Fe2O3:Coads samples is similar to the trend reported not only for undoped γ-Fe2O3 particles, but also for the coprecipitated samples. We showed in ref 25 that the small increase in magnetic anisotropy observed on heating for γ-Fe2O3 particles should be ascribed to an enhanced surface contribution, presumably associated with the removal of surface hydroxyl groups and formation of new oxo bridges. Because of 19676



Article

Figure 4. Experimental EXAFS spectrum (circles) and calculated spectrum (solid line) assuming a single shell of oxygen neighbors around Co, for Co(NO 3)2·6H2O and for the as-synthesized Fe2O3:Coads sample. Dashed line: calculated spectrum corresponding to a first shell of O atoms and a second shell of Fe atoms as in bulk CoFe2O4.

the analysis of the Fe2O3:Coads sample treated at 100 °C, the amplitude reduction factor was subsequently fixed to 0.63 and the Debye−Waller factor to 0.07 Å, leading to fitted parameters of 2.10 ± 0.01 Å for the Co−O distance and 5.6 ± 1.0 for the number of oxygen neighbors (R-factor 8%). The mean Co−O distance does not significantly differ from that of the Co2+ hexaaquo complex, and the difference in coordination number is typically within the error bar. A second fit was performed after introduction of a second shell representative of iron atoms in the octahedral and tetrahedral sites of a spinel structure, leading to 5.9 ± 1.1 oxygen first neighbors at a distance of 2.09 ± 0.01 Å, a maximum of 0.4 Fe atoms at 2.99 ± 0.01 Å, and 2.1 ± 1.8 additional iron atoms at 3.50 ± 0.01 Å (R-factor 9%). Starting Co−Fe distances were derived from previous X-ray diffraction (cif file, ICSD-39131) and EXAFS47 investigations of CoFe2O4. Deviations of the three first next-neighbor distances with respect to these starting values were constrained to be equal, in order to decrease the number of refined parameters. As shown in Figure 4, the corresponding fit (dashed line) does not lead to a better agreement considering the signal-to-noise ratio at high k values. In addition, the number of second neighbors derived from the fit is rather low, with values comparable to the standard deviation. Hence, it is not possible to draw a definite conclusion on the nature of a hypothetical second coordination shell, but EXAFS data suggest a reduced number of Fe neighbors, about one to two per Co dopant. The mechanism for Co uptake from iron oxides is assumed to involve exchange reaction between Co2+ ions in solution and protons of surface hydroxyl groups.21 Our results agree with a scenario of cobalt dopants bounded to the surface Fe atoms through hydroxo or oxo bridges, but elucidation of the mono- or multidentate complexation scheme as discussed in refs 36 and 48, is out of the scope of the present data quality. The main point is that both XANES and EXAFS findings are consistent with the fact that the Co2+ species are not located within a spinel lattice, not even in the outermost surface layers. This result disagrees with the conclusions drawn by Uheida et

Figure 3. (a) Normalized XAFS spectra of the Fe2O3:Coads samples (series 1) and of the Co(NO3)2·6H2O and CoFe2O4 standards. (b) Enlarged view of the XANES region for selected samples and standards. Inset: the 7700−7715 eV pre-edge range.

line attributed to the 1s → 4p dipole-allowed transition superimposed to the main edge, and small pre-edge structures usually assigned to quadrupolar 1s → 3d and dipolar 1s → 4p transitions, the latter occurring through the hybridization of 3d and 4p states of neighboring transition metal ions.45,46 For the Fe2O3:Coads sample obtained after synthesis, the XANES spectrum consists of a broad main line, with no substructures, similar to the one of Co(NO3)2·6H2O. It thus departs from the spectrum of CoFe2O4, which is structured into three narrow lines, labeled P1, P2, and P3. Note that the energy of the absorption edge (defined as the inflection point of the main line) is similar for Fe2O3:Coads (100 °C), hydrated cobalt nitrate, and CoFe2O4, giving conclusive evidence that cobalt is present at the +II valence state. One would expect a shift of the main line of about 5 eV for an increase of the oxidation state, from +II (CoO, CoFe2O4) to +III (Co3O4), as shown in the Supporting Information (Figure S4). Similarity between the Fe2O3:Coads sample and the Co(NO3)2·6H2O reference was further confirmed by comparing the k2-weighted EXAFS signals (Figure 4). The two sets of data show comparable oscillations, indicative of similar local environments. Data corresponding to the Co2+ aquo complex were first refined considering Co2+ ions surrounded by a single shell of six oxygen atoms, and using single and multiple scattering paths. A Co−O distance of 2.08 ± 0.02 Å was inferred from the fit, with an amplitude reduction factor of 0.63 ± 0.05 and a Debye−Waller factor of 0.07 ± 0.04 Å. The quality of the refinement is illustrated by the good agreement between experimental and calculated spectra (R-factor 3%). For 19677



Article

al. which pointed out that Co should diffuse at room temperature into the crystal structure, by exchange with Fe2+ ions for magnetite particles and through vacancy filling for maghemite particles.21 These XAFS results are however consistent with the previous magnetic study, which indicates hardly any change of magnetocrystalline anisotropy for the γFe2O3 particles after surface treatment with Co2+ ions. In contrast, the XAFS spectrum of the Fe2O3:Coads sample annealed at 600 °C strongly departs from that of Co(NO3)2· 6H2O, but is comparable to the one of CoFe2O4 (Figure 3). In particular, the XANES region exhibits three distinct substructures in the main line. The multiple oscillations in the EXAFS region are also indicative of several coordination shells around the absorbing Co atoms, similar to those found in CoFe2O4. Note that the sample spectrum markedly differs from that of Co3O4 (Figure S4, Supporting Information), indicating that the Co distribution at the surface of the γ-Fe2O3 particles is quite homogeneous and annealing did not give rise to a segregated Co3O4 phase. As reported for the as-synthesized particles, the edge energy suggests a +II oxidation state for most of the Co ions (Figure 3b). All these findings corroborate the localization of the Co2+ dopants in the spinel lattice after heat treatment at 600 °C. To discuss the modification in Co bonding upon heating, XAFS data were monitored at intermediate annealing temperatures (Figure 3a and Figure S5 in Supporting Information). As shown in the Supporting Information, Figure S5, the edge position does not shift in energy after the various heat treatments. The hypothesis of a Co3+ → Co2+ reduction process, which could explain part of the increase in magnetic anisotropy observed in the 200−300 °C range (see Table 2), must be ruled out. Additional information on the cobalt localization, in octahedral or tetrahedral positions, can be obtained from the pre-edge structures. The 1s → 3d electronic transitions are forbidden by dipole selection rules and hence have a weak intensity. As a result, noncentrosymmetric (e.g., tetrahedral) positions result in more intense pre-edge features compared to centrosymmetric (e.g., octahedral) positions, due to dipole−quadrupole mixing. The inset of Figure 3b presents an enlarged view of the pre-edge region, where the spectra of the Fe2O3:Coads samples treated at 100 and 600 °C are compared to those of Co(NO3)2·6H2O and of Co doped Al2O3. This latter blue pigment is characteristic of Co2+ ions in the tetrahedral sites of a γ-Al2O3 spinel structure.49 The preedge peak is clearly more intense for Al2O3:Co (ca. 0.08 in unit of step edge) than for Co(NO3)2·6H2O (normalized peak height of ca. 0.01), confirming a 4-fold symmetry for the former and a 6-fold coordination for the latter. For the Fe2O3:Coads samples, no distinct peak is detected in the pre-edge region, supporting the idea that cobalt is mainly located in octahedral environment. Co diffusion would thus only proceed via octahedral cationic sites, presumably because maghemite presents 17% vacancies on these sites. Note, however, that similar results were reported for the vacancy-free magnetite structure upon Co2+ doping, with a preferential localization of the dopants in the octahedral sites of the spinel structure.50,51 Cobalt localization in octahedral environment was also corroborated by comparing the Fourier transform moduli of the k2χ(k) data (Figure 5a). The first peak observed in the corresponding χ(R) plots accounts for oxygen nearest neighbors and the next peaks account for a second coordination shell composed of iron in tetrahedral and octahedral sites. If Co sits in an octahedral position, the second coordination shell is

Figure 5. Fourier transforms of k2-weighted EXAFS oscillations for (a) Fe2O3:Coads (series 1) and for (b) coprecipitated Fe2O3:Co (series 2) samples annealed between 100 and 600 °C.

split into two subpeaks; the first one is characteristic of the distance between Co and Fe atoms located in octahedral sites, at ∼3.0 Å with a 6-fold multiplicity, and the second subpeak features the distance between Co and Fe atoms located in tetrahedral sites, at ∼3.5 Å, also with a 6-fold multiplicity. In contrast, if the Co absorber is located in a tetrahedral site, the second coordination shell only accounts for Fe atoms in octahedral sites, at ∼3.5 Å, with a 12-fold multiplicity.47 As a result, even a small fraction of Co in tetrahedral sites would lead to a large intensity of the 3.5 Å subpeak. The relative intensity between these two subpeaks gives thus useful insights in cobalt coordination.15 As shown in Figure 5, the intensity ratio between these peaks does not significantly change upon annealing. The intensity of the subpeak at about 3.2 Å is systematically smaller than the one of the peak at 2.6 Å. This result confirms that no substantial site redistribution of Co over the two cationic sites occurs upon heating. Consequently, the large increase in anisotropy discussed in the magnetic study should not account for a symmetry change around the Co2+ dopants. To get further information on the diffusion mechanism, we have plotted in Figure 6 the area of the peaks corresponding to the first and second coordination shells as a function of annealing temperature. The amplitude of the first peak remains almost constant, i.e., representative of a 6-fold coordination, over the whole temperature range. Variations are only observed for the peaks associated with second neighbors. For the assynthesized Fe2O3:Coads sample, Co2+ ions are merely adsorbed onto the particle surface with no clear second peaks. For heat treatments carried out between 200 and 300 °C, the overall number of second neighbors increases in a steep way and then reaches a plateau. This evolution has to be related to the progressive incorporation of the Co dopants in the γ-Fe2O3 lattice, while the plateau above 300 °C reflects a full insertion. These observations are again in good accordance with the results of the magnetic study summarized in Figure 2. Between 19678



Article

to be rather precise as the P2 and P3 substructures are well resolved and their relative intensity shows a marked dependence on annealing. As shown in Figure 7, all experimental

Figure 6. Peak area related to the first and second coordination shells around Co versus annealing temperature, for samples of series 1 (Fe2O3:Coads) and series 2 (coprecipitated Fe2O3:Co). The dashed lines are guides for the eyes.

Figure 7. (left) XANES spectra of Fe2O3:Coads samples (series 1) heat treated at selected temperatures and fitted to a linear combination of the spectra of the Fe2O3:Coads samples annealed at 100 and 600 °C. (right) Similar analysis for the coprecipitated samples (series 2) using as references the XANES spectrum of the Fe2O3:Coads sample dried at 100 °C and that of the coprecipitated sample heat treated at 600 °C. Experimental spectra are plotted in bold gray lines and fits are plotted in black dashed lines.

200 and 300 °C, the change in magnetic anisotropy is mostly indicative of an increased magnetocrystalline contribution due to the incorporation of highly anisotropic Co2+ ions in the octahedral sites of γ-Fe2O3. From 300 to 800 °C, the small change in magnetic anisotropy is no longer representative of a magnetocrystalline contribution, but rather of a surface effect. It is difficult to access structural information beyond the second coordination shell, even for the CoFe2O4 bulk standard. Hence, it is not possible to discriminate between a scenario of homogeneous doping within the particles or the confinement of the dopants near the surface. Note that filling of vacant Fe sites in the volume may not be the only mechanism for Co incorporation due to electrostatic considerations. Exchange with Fe3+ ions, which allows for a partial electro-compensation, could also take place. Indeed, for 7 nm undoped γ-Fe2O3 nanoparticles a partial ordering of the Fe vacancies, with space group lowering from Fd3m ̅ to P4332, has been observed below 300 °C.25 This ordering typically involves a migration of Fe3+ ions within the octahedral sites. This temperature range thus allows for a diffusion of Fe within the maghemite lattice. The previous analysis suggested that, for annealing temperatures between 100 and 300 °C, part of the Co2+ dopants remain adsorbed on the surface. This assumption can be verified by modeling the XANES spectra by linear combinations of reference spectra. Indeed, the XANES spectra, characteristic of either adsorbed or bulk Co2+ species, present very different shapes, i.e., a broad line with no substructures for the former and narrow lines with well-defined substructures (P1, P2, and P3) for the latter (see Figure 3b). In the following, we used the spectrum of the as-synthesized Fe2O3:Coads sample (dried at 100 °C) to model the contribution of Co2+ ions immobilized at the surface. The spectrum of the Fe2O3:Coads sample heat treated at 600 °C served as a reference for a full insertion of the dopants in the lattice. At intermediate annealing temperatures, experimental data were fitted to a linear combination of these two spectra. This procedure was found

spectra were perfectly reproduced by this model. Fitted parameters are summarized in Table 3. For instance, annealing at 250 °C makes half of the initially adsorbed ions diffuse into maghemite. After heat treatment at 300 °C, 92 ± 2% of the Co2+ ions are incorporated, confirming the previous results based on the analysis of the χ(R) plots. Note that a similar threshold temperature of 250 °C was recently mentioned by Mayanovic et al. for Co diffusion in magnetite nanoparticles dispersed in supercritical aqueous fluids.52 The thermal evolution of the relative fraction between adsorbed and bulk Co2+ species suggests that the limiting step for Co diffusion in γ-Fe2O3 corresponds to its incorporation in the crystal lattice, presumably due to charge compensation. XAFS Study: Co Localization in Series 2 (Fe2O3:Co). Similar XAFS analyses were carried out for the coprecipitated Fe2O3:Co samples. Figure 8a presents the corresponding spectra, compared to those of Co(NO 3 ) 2 ·6H 2 O and CoFe2O4. Contrary to what was reported for series 1, all spectra present strong characteristic features of the CoFe2O4 reference. The energy of the main absorption edge is similar for Co(NO3)2·6H2O, CoFe2O4, and Fe2O3:Co samples heat treated between 100 and 600 °C (Figure 8b), indicating a +II valence state for Co. As in series 1, cobalt should be mainly located in an octahedral environment, as no distinct peak could be detected in the 7700−7715 eV pre-edge range. Absence of Co2+ redistribution over the tetrahedral and octahedral sites was also confirmed by analyzing the area of the peaks in the pseudoradial distribution function (RDF), χ(R) (see Figures 5b and 6). The peak area corresponding to oxygen nearest neighbors remains globally constant and, thus, mostly representative of a 6-fold coordination. In addition, the 19679



Article

Table 3. Percentage of Co2+ Adsorbed at the Surface of γ-Fe2O3 Derived from the Fit of the XANES Spectra Fe2O3:Coads 100 °C %Fe2O3:Coads (100 °C) %Fe2O3:Coads (600 °C)

%Fe2O3:Coads (100 °C) %Fe2O3:Co (600 °C)

100 0

250 °C

275 °C

300 °C

600 °C

48 ± 1 52 ± 1 Fe2O3:Co

20 ± 2 80 ± 2

8±2 92 ± 2

0 100

100 °C

250 °C

275 °C

600 °C

23 ± 3 77 ± 3

14 ± 3 86 ± 3

13 ± 3 87 ± 3

0 100

spectra of the coprecipitated samples by a weighted sum of two reference spectra. The first one corresponds to the assynthesized Fe2O3:Coads sample, which is representative of Co2+ ions adsorbed onto γ-Fe2O3 particles. We used the spectrum of the coprecipitated sample heat treated at 600 °C as the reference for a full Co2+ incorporation in the lattice. As in series 1, calculated spectra compare well to the experimental data (see Figure 7) and the corresponding fitted parameters are summarized in Table 3. For these Co doped γ-Fe2O3 particles synthesized by direct coprecipitation of Co2+, Fe2+, and Fe3+, 23% of the dopants were not inserted in the bulk phase after synthesis but were merely immobilized at the surface. According to these fits, half of the initially adsorbed Co ions were incorporated into the lattice after a 250 °C heat treatment as in series 1. After annealing at 600 °C, the area of the second peaks in the RDF is similar in the two series (see Figure 6) and should reflect a full incorporation of the dopants. Note that previous studies have shown that the reflux treatment in the presence of Fe(NO3)3 that was used after coprecipitation passivates the surface of CoFe2O4 particles against the dissolution and release of Co2+ ions in acidic solution.53 The presence of adsorbed Co2+ ions in the present study should thus be related to the coprecipitation step, rather than to a postsynthesis leaching process. To further justify the previous analyses, we have plotted in Figure 9 the percentage of Co ions inserted in the spinel lattice Figure 8. (a) Normalized XAFS spectra of Fe2O3:Co samples (series 2) annealed at different temperatures and of Co(NO3)2·6H2O and CoFe2O4 standards. (b) Enlarged view of the XANES region.

intensity ratio between the two subpeaks at 2.6 and 3.2 Å, which correspond to iron second neighbors in octahedral and tetrahedral sites, hardly changed after annealing. Synthesis routes based on coprecipitation of Co and Fe precursors are usually assumed to lead to a homogeneous distribution of the Co dopants in the γ-Fe2O3 lattice. If so, the present XAFS measurements fail to explain the increase in magnetic anisotropy observed between 200 and 300 °C as there is no evidence for some rearrangement of the dopants between the cationic sites. In contrast, the peaks representative of the second coordination shell in the RDF (Figure 5b) show an enhanced intensity upon heating. This evolution is consistent with that of the blocking temperature (Figure 2) and indicates an increasing number of neighbors upon heat treatment. One has again to invoke an incomplete dopant incorporation at the end of synthesis. Part of the dopants should have been immobilized at the particle surface, so their evolution upon heating is strictly similar to the one reported for series 1. To corroborate this interpretation, and get quantitative information on the effective doping concentration, we modeled the XANES

Figure 9. Percentage of Co ions inserted in γ-Fe2O3 according to Table 3 versus the area of the peaks associated with the second coordination shell of Co in series 1 and 2.

(values listed in Table 3) as a function of the area of the peaks corresponding to Co−Fe nearest neighbors. There is a clear linear relationship between these two parameters for the two series of samples. As the area of the peaks in the RDF is proportional to the coordination number (N2) times the percentage of bulk Co ions, the proportionality indicates that N2 is a constant. In the late 1970s, cobalt doping was 19680



■

intensively studied to increase the coercivity of iron oxide particles for magnetic recording applications.20,22,47,54 Doping involved either surface treatment (for instance alkaline precipitation of cobalt in the presence of preformed γ-Fe2O3 particles) or coprecipitation of cobalt and iron precursors. The present study shows that in the case of nanoparticles, with a large surface-to-volume ratio, a significant fraction of dopants could remain adsorbed on the surface, reducing de facto the true doping level.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors are indebted to the European Synchrotron Radiation Facility for provision of synchrotron X-ray beamtime. Construction of the spectrometer was financially supported by the French National Agency (ANR Projects NANOSURF and MESONNET), CEREGE laboratory (UMR 7330, Aix en Provence, France), and the INSU CNRS institute. The authors also thank M. Poggi for the TEM observations. P. Bonville is acknowledged for his help with the simulations of the ZFC curves. The research described here has been supported by Triangle de la Physique (Contract No. 2008-051T). The visit of C.V. to the University of California Santa Barbara was hosted by the IMI Program of the National Science Foundation under Award No. DMR 08-43934. The authors also acknowledge the National Science Foundation under Grant NSF DMR-1121053, the Material Science Laboratory, and the MRL staff for the use of the Quantum Design MPMS-5 SQUID magnetometer.

■

CONCLUSION Fine tuning of the magnetic anisotropy of γ-Fe2O3 nanoparticles can be achieved at fixed particle size through cobalt doping. In this comparative study, we tentatively correlate the anisotropy constant value to the effective doping level in samples prepared by two techniques, i.e., surface treatment or coprecipitation. The first doping strategy involved immobilization of Co2+ ions on preformed magnetite particles, followed by further oxidation in maghemite and diffusion of the dopants upon heating. In the second approach, the Co2+ dopants were directly coprecipitated with the iron precursors. The two series of samples were subjected to heat treatments up to 800 °C, after dispersion of the particles in a silica matrix to avoid grain growth. Co localization and its valence state were subsequently determined by X-ray absorption spectroscopy at the Co K-edge. For the samples doped by surface treatment, both XANES and EXAFS findings were consistent with a scenario of Co2+ species adsorbed at the surface of the iron oxide, which do not affect magnetic anisotropy. Upon annealing at temperatures higher than 250 °C, the Co2+ ions were gradually inserted in the lattice, preferentiallyif not onlyin the octahedral sites of the spinel structure. As expected, a large increase in magnetocrystalline anisotropy was associated with this diffusion process. Fitting the XANES spectra by linear combinations of reference spectra, representative of adsorbed Co2+ species or of Co2+ ions in the octahedral sites of γ-Fe2O3, allowed us to quantitatively assess the effective doping level and to monitor Co2+ diffusion upon annealing. Above 300 °C, incorporation of the dopants was virtually complete. Nevertheless, the question of their diffusion length, restricted or not to the uppermost layers, remains open. For heat treatments above 300 °C, a slight increase in magnetic anisotropy was still observed but had to be related to the dehydroxylation of terminal Fe atoms, and thus to a contribution of surface anisotropy. Surprisingly, a large increase in magnetic anisotropy was also observed in the 200− 300 °C range for the coprecipitated samples. XAFS analyses showed that, for this direct coprecipitation protocol, a substantial part of the dopants were merely adsorbed at the particle surface, up to 20% for 7 nm particles. Previous investigations devoted to the change in magnetic properties upon Co doping in systems based on iron oxide nanoparticles should thus be revised accordingly.

■

Article

■

REFERENCES

(1) Tran, N.; Webster, T. J. Magnetic Nanoparticles: Biomedical Applications and Challenges. J. Mater. Chem. 2010, 20, 8760−8767. (2) Pankhurst, Q. A.; Thanh, N. K. T.; Jones, S. K.; Dobson, J. Progress in Applications of Magnetic Nanoparticles in Biomedicine. J. Phys. D: Appl. Phys. 2009, 42, 224001. (3) Pankhurst, Q. A.; Connolly, J.; Jones, S. K.; Dobson, J. Applications of Magnetic Nanoparticles in Biomedicine. J. Phys. D: Appl. Phys. 2003, 36, R167. (4) Luigjes, B.; Woudenberg, S. M. C.; de Groot, R.; Meeldijk, J. D.; Galvis, H. M. T.; de Jong, K. P.; Philipse, A. P.; Erné, B. H. Diverging Geometric and Magnetic Size Distributions of Iron Oxide Nanocrystals. J. Phys. Chem. C 2011, 115, 14598−14605. (5) Levy, M.; Quarta, A.; Espinosa, A.; Figuerola, A.; Wilhelm, C.; Garcia-Hernandez, M.; Genovese, A.; Falqui, A.; Alloyeau, D.; Buonsanti, R.; et al. Correlating Magneto-Structural Properties to Hyperthermia Performance of Highly Monodisperse Iron Oxide Nanoparticles Prepared by a Seeded-Growth Route. Chem. Mater. 2011, 23, 4170−4180. (6) Fortin, J.-P.; Wilhelm, C.; Servais, J.; Ménager, C.; Bacri, J.-C.; Gazeau, F. Size-Sorted Anionic Iron Oxide Nanomagnets as Colloidal Mediators for Magnetic Hyperthermia. J. Am. Chem. Soc. 2007, 129, 2628−2635. (7) Hergt, R.; Dutz, S.; Röder, M. Effects of Size Distribution on Hysteresis Losses of Magnetic Nanoparticles for Hyperthermia. J. Phys.: Condens. Matter 2008, 20, 385214. (8) Sun, S.; Zeng, H. Size-Controlled Synthesis of Magnetite Nanoparticles. J. Am. Chem. Soc. 2002, 124, 8204−8205. (9) Kwon, S. G.; Piao, Y.; Park, J.; Angappane, S.; Jo, Y.; Hwang, N.M.; Park, J.-G.; Hyeon, T. Kinetics of Monodisperse Iron Oxide Nanocrystal Formation by “Heating-Up” Process. J. Am. Chem. Soc. 2007, 129, 12571−12584. (10) Carrey, J.; Mehdaoui, B.; Respaud, M. Simple Models for Dynamic Hysteresis Loop Calculations of Magnetic Single-Domain Nanoparticles: Application to Magnetic Hyperthermia Optimization. J. Appl. Phys. 2011, 109, 083921. (11) Lévy, M.; Wilhelm, C.; Siaugue, J.-M.; Horner, O.; Bacri, J.-C.; Gazeau, F. Magnetically Induced Hyperthermia: Size-Dependent Heating Power of γ-Fe2O3 Nanoparticles. J. Phys.: Condens. Matter 2008, 20, 204133. (12) Hergt, R.; Hiergeist, R.; Zeisberger, M.; Glockl, G.; Weitschies, W.; Ramirez, L. P.; Hilger, I.; Kaiser, W. A. Enhancement of AC-

ASSOCIATED CONTENT

* Supporting Information S

TEM micrographs and size histograms for series 1, 2, and 3; Xray diffraction patterns measured for different annealing temperatures; magnetization loops recorded at 10 K after annealing in series 1 and 2; XANES spectra of series 1 and of selected reference samples. This material is available free of charge via the Internet at http://pubs.acs.org. 19681



Article

Losses of Magnetic Nanoparticles for Heating Applications. J. Magn. Magn. Mater. 2004, 280, 358−368. (13) Demortière, A.; Panissod, P.; Pichon, B. P.; Pourroy, G.; Donnio, B.; Bégin-Colin, S. Size-Dependent Properties of Magnetic Iron Oxide Nanocrystals. Nanoscale 2011, 3, 225−232. (14) Slonczewski, J. C. Origin of Magnetic Anisotropy in CobaltSubstituted Magnetite. Phys. Rev. 1958, 110, 1341−1348. (15) Fantechi, E.; Campo, G.; Carta, D.; Corrias, A.; de Julian Fernandez, C.; Gatteschi, D.; Innocenti, C.; Pineider, F.; Rugi, F.; Sangregorio, C. Exploring the Effect of Co Doping in Fine Maghemite Nanoparticles. J. Phys. Chem. C 2012, 116, 8261−8270. (16) Chakrabarti, S.; Mandal, S. K.; Chaudhuri, S. Cobalt Doped γFe2O3 Nanoparticles: Synthesis and Magnetic Properties. Nanotechnology 2005, 16, 506. (17) Hu, L.; de Montferrand, C.; Lalatonne, Y.; Motte, L.; Brioude, A. Effect of Cobalt Doping Concentration on the Crystalline Structure and Magnetic Properties of Monodisperse CoxFe3−xO4 Nanoparticles within Nonpolar and Aqueous Solvents. J. Phys. Chem. C 2012, 116, 4349−4355. (18) Vichery, C.; Maurin, I.; Boilot, J.-P.; Gacoin, T. Post-synthesis Heat Treatments of γ-Fe2O3 Nanoparticles Embedded in a Refractory Matrix: From Annealing of Structural Defects to Doping. J. Appl. Phys. 2012, 111, 07B541. (19) Salazar-Alvarez, G.; Sort, J.; Uheida, A.; Muhammed, M.; Suriñach, S.; Baro, M. D.; Nogués, J. Reversible Post-synthesis Tuning of the Superparamagnetic Blocking Temperature of γ-Fe2O3 Nanoparticles by Adsorption and Desorption of Co(II) Ions. J. Mater. Chem. 2007, 17, 322−328. (20) Tokuoka, Y.; Umeki, S.; Imaoka, Y. Anisotropy of CobaltAdsorbed γ-Fe2O3 Particles. J. Phys., Colloq. 1977, 38, C1-337−C1340. (21) Uheida, A.; Salazar-Alvarez, G.; Bjö rkman, E.; Yu, Z.; Muhammed, M. Fe3O4 and γ-Fe2O3 Nanoparticles for the Adsorption of Co2+ from Aqueous Solution. J. Colloid Interface Sci. 2006, 298, 501−507. (22) Tewari, P. H.; Campbell, A. B.; Lee, W. Adsorption of Co2+ by Oxides from Aqueous Solution. Can. J. Chem. 1972, 50, 1642−1648. (23) Maurin, I.; Dantelle, G.; Boilot, J.-P.; Gacoin, T. A Protected Annealing Process for the Production of High Quality Colloidal Oxide Nanoparticles with Optimized Physical Properties. J. Mater. Chem. C 2013, 1, 13−22. (24) Zhang, H. Y.; Gu, B. X.; Zhai, H. R.; Lu, M.; Miao, Y. Z.; Zhang, S. Y. Anisotropy and Faraday Effect in Co Spinel Ferrite Films. J. Appl. Phys. 1994, 75, 7099. (25) Vichery, C.; Maurin, I.; Bonville, P.; Boilot, J.-P.; Gacoin, T. Influence of Protected Annealing on the Magnetic Properties of γFe2O3 Nanoparticles. J. Phys. Chem. C 2012, 116, 16311−16318. (26) Langford, J. I. The Use of the Voigt Function in Determining Microstructural Properties from Diffraction Data by Means of Pattern Decomposition. In Accuracy in Powder Diffraction II; Prince, E., Stalick, J. K., Eds.; NIST Special Publication 846; National Institute of Standards and Technology: Boulder, CO, USA, 1992; pp 110−126. (27) Rodriguez-Carvajal, J. Recent Advances in Magnetic Structure Determination by Neutron Powder Diffraction. Physica B 1993, 192, 55−69. (28) Proux, O.; Nassif, V.; Prat, A.; Ulrich, O.; Lahera, E.; Biquard, X.; Menthonnex, J.-J.; Hazemann, J.-L. Feedback System of a LiquidNitrogen-Cooled Double-Crystal Monochromator: Design and Performances. J. Synchrotron Radiat. 2006, 13, 59−68. (29) Hazemann, J.-L.; Proux, O.; Nassif, V.; Palancher, H.; Lahera, E.; Da Silva, C.; Braillard, A.; Testemale, D.; Diot, M. A.; Alliot, I.; et al. High-Resolution Spectroscopy on an X-ray Absorption Beamline. J. Synchrotron Radiat. 2009, 16, 283−292. (30) Collart, E.; Shukla, A.; Gélébart, F.; Morand, M.; Malgrange, C.; Bardou, N.; Madourib, A.; Pelouard, J.-L. Spherically Bent Analyzers for Resonant Inelastic X-Ray Scattering with Intrinsic Resolution below 200 meV. J. Synchrotron Radiat. 2005, 12, 473−478. (31) Proux, O.; Micha, J. S.; Régnard, J. R.; Traverse, A.; Dieny, B.; Ernult, F.; Bayle-Guillemaud, P.; Hazemann, J.-L.; Giacomoni, L.

Structural Investigations of Co/ZrO2 Discontinuous Multilayers by Xray Absorption Fine Structure Spectroscopy. J. Phys.: Condens. Matter 2003, 15, 7237. (32) Llorens, I.; Lahera, E.; Delnet, W.; Proux, O.; Braillard, A.; Hazemann, J.-L.; Prat, A.; Testemale, D.; Dermigny, Q.; Gelebart, F.; et al. High Energy Resolution Five-Crystal Spectrometer for High Quality Fluorescence and Absorption Measurements on an X-ray Absorption Spectroscopy Beamline. Rev. Sci. Instrum. 2012, 83, 063104. (33) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: Data Analysis for X-ray Absorption Spectroscopy using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537−541. (34) Ankudinov, A.; Ravel, B.; Rehr, J.; Conradson, S. Real-Space Multiple-Scattering Calculation and Interpretation of X-ray-Absorption Near-Edge Structure. Phys. Rev. B 1998, 58, 7565−7576. (35) Auffan, M.; Rose, J.; Proux, O.; Borschneck, D.; Masion, A.; Chaurand, P.; Hazemann, J.-L.; Chaneac, C.; Jolivet, J.-P.; Wiesner, M. R.; et al. Enhanced Adsorption of Arsenic onto Maghemites Nanoparticles: As(III) as a Probe of the Surface Structure and Heterogeneity. Langmuir 2008, 24, 3215−3222. (36) Tamura, H.; Katayama, N.; Furuichi, R. The Co2+ Adsorption Properties of Al2O3, Fe2O3, Fe3O4, TiO2, and MnO2 Evaluated by Modeling with the Frumkin Isotherm. J. Colloid Interface Sci. 1997, 195, 192−202. (37) Sorescu, M.; Grabias, A.; Tarabasanu-Mihaila, D.; Diamandescu, L. From Magnetite to Cobalt Ferrite. J. Mater. Synth. Process. 2001, 9, 119−123. (38) Dormann, J. L.; Fiorani, D.; Tronc, E. On the Models for Interparticle Interactions in Nanoparticle Assemblies: Comparison with Experimental Results. J. Magn. Magn. Mater. 1999, 202, 251−267. (39) Morup, S.; Hansen, M. F.; Frandsen, C. Magnetic Interactions Between Nanoparticles. Beilstein J. Nanotechnol. 2010, 1, 182−190. (40) Klokkenburg, M.; Erné, B. H.; Mendelev, V.; Ivanov, A. O. Magnetization Behavior of Ferrofluids with Cryogenically Imaged Dipolar Chains. J. Phys.: Condens. Matter 2008, 20, 204113. (41) Peddis, D.; Cannas, C.; Musinu, A.; Piccaluga, G. Coexistence of Superparamagnetism and Spin-Glass Like Magnetic Ordering Phenomena in a CoFe2O4-SiO2 Nanocomposite. J. Phys. Chem. C 2008, 112, 5141−5147. (42) Cannas, C.; Musinu, A.; Ardu, A.; Orru, F.; Peddis, D.; Casu, M.; Sanna, R.; Angius, F.; Diaz, G.; Piccaluga, G. CoFe2O4 and CoFe2O4/SiO2 Core/Shell Nanoparticles: Magnetic and Spectroscopic Study. Chem. Mater. 2010, 22, 3353−3361. (43) Kachkachi, H.; Ezzir, A.; Noguès, M.; Tronc, E. Surface Effects in Nanoparticles: Application to Maghemite γ-Fe2O3. Eur. Phys. J. B 2000, 14, 681−690. (44) Millan, A.; Urtizberea, A.; Silva, N. J. O.; Palacio, F.; Amaral, V. S.; Snoeck, E.; Serin, V. Surface Effects in Maghemite Nanoparticles. J. Magn. Magn. Mater. 2007, 312, L5−L9. (45) de Groot, F.; Vankó, G.; Glatzel, P. The 1s X-ray Absorption Pre-Edge Structures in Transition Metal Oxides. J. Phys.: Condens. Matter 2009, 21, 104207. (46) Yamamoto, T. Assignment of Pre-Edge Peaks in K-edge X-ray Absorption Spectra of 3d Transition Metal Compounds: Electric Dipole or Quadrupole? X-ray Spectrom. 2008, 37, 572−584. (47) Nordhei, C.; Ramstad, A. L.; Nicholson, D. G. Nanophase Cobalt, Nickel and Zinc Ferrites: Synchrotron XAS Study on the Crystallite Size Dependence of Metal Distribution. Phys. Chem. Chem. Phys. 2008, 10, 1053−1066. (48) Tamura, H.; Matijevic, E.; Meites, L. Adsorption of Co2+ Ions on Spherical Magnetite Particles. J. Colloid Interface Sci. 1983, 92, 303−314. (49) Cava, S.; Tebcherani, S.; Pianaro, S.; Paskocimas, C.; Longo, E.; Varela, J. Structural and Spectroscopic Analysis of γ-Al2O3 to α-Al2O3CoAl2O4 Phase Transition. Mater. Chem. Phys. 2006, 97, 102−108. (50) Lelis, M.; Porto, A.; Gonçalves, C.; Fabris, J. Cation Occupancy Sites in Synthetic Co-Doped Magnetites as Determined with X-ray Absorption (XAS) and Mössbauer Spectroscopies. J. Magn. Magn. Mater. 2004, 278, 263−269. 19682



Article

(51) Sorescu, M.; Grabias, A.; Tarabasanu-Mihaila, D.; Diamandescu, L. Influence of Cobalt and Nickel Substitutions on Populations, Hyperfine fields, and Hysteresis Phenomenon in Magnetite. J. Appl. Phys. 2002, 91, 8135. (52) Mayanovic, R. A.; Yan, H.; Anderson, A. J.; Meredith, P. R.; Bassett, W. A. In Situ X-ray Absorption Spectroscopic Study of the Adsorption of Ni2+ on Fe3O4 Nanoparticles in Supercritical Aqueous Fluids. J. Phys. Chem. C 2012, 116, 2218−2225. (53) Soler, M. A. G.; Lima, E. C. D.; da Silva, S. W.; Melo, T. F. O.; Pimenta, A. C. M.; Sinnecker, J. P.; Azevedo, R. B.; Garg, V. K.; Oliveira, A. C.; Novak, M. A.; et al. Aging Investigation of Cobalt Ferrite Nanoparticles in Low pH Magnetic Fluid. Langmuir 2007, 23, 9611−9617. (54) Berkowitz, A. E.; Parker, F. E.; Hall, E. L.; Podolsky, G. Toward a Model for Co-Surface-Treated Fe-Oxides. IEEE Trans. Magn. 1988, 24, 2871−2873.

19683


Introduction of Cobalt Ions in γ-Fe2O3 Nanoparticles by Direct

Recommend Documents