Puzzling Mechanism behind a Simple Synthesis of Cobalt and Cobalt

Mar 11, 2014 - Puzzling Mechanism behind a Simple Synthesis of Cobalt and Cobalt Oxide ...... London starts world's largest air-pollution monitoring n...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/cm

Puzzling Mechanism behind a Simple Synthesis of Cobalt and Cobalt Oxide Nanoparticles: In Situ Synchrotron X‑ray Absorption and Diffraction Studies Malwina Staniuk,†,§ Ofer Hirsch,†,§ Niklaus Kran̈ zlin,†,§ Rahel Böhlen,† Wouter van Beek,‡ Paula M. Abdala,‡ and Dorota Koziej*,† †

Laboratory for Multifunctional Materials, Department of Materials, ETH Zurich, Vladimir-Prelog Weg 5, 8093 Zurich, Switzerland Swiss-Norwegian Beamlines at European Synchrotron Research Facility, 6 Rue Jules Horowitz, 38043 Grenoble, France



S Supporting Information *

ABSTRACT: Here, we show a simple approach to synthesize cobalt and cobalt oxide nanoparticles in an organic solvent. We find that the cubic Co3O4 nanoparticles can be easily obtained, even at temperatures as low as 80 °C. Moreover, exactly the same reaction at 180 °C leads to metallic Co nanoparticles. Thus, in addition to the synthetic efforts, we study the mechanism of occurrence of oxidation and reduction of a Co2+ precursor in benzyl alcohol. Remarkably, the in situ X-ray absorption and diffraction measurements of the synthesis at 140 °C reveal that oxidation of Co2+ to Co3+/2+ and reduction of Co2+ to Co0 reactions take place simultaneously. It is followed by a rapid formation of Co3O4 nanoparticles and its consecutive solid-state reduction to CoO. In parallel, metallic Co nanoparticles begin to grow. In addition, Multicomponent Curve Resolution−Alternating Least Squares (MCR-ALS) analysis of X-ray absorption spectroscopy (XAS) data efficiently reveals the nontrivial interdependence between four different reactions. Our strategy to control reduction and oxidation of Co-based nanoparticles as they grow opens up an elegant pathway for the one-pot-synthesis of the hybrid materials for energy-related applications.



INTRODUCTION Co-based nanoparticles are known for their application in Liion batteries,1−4 Fischer−Tropsch synthesis,5,6 and water splitting.2,7 However, their utility is critically dependent on the redox behavior that is closely linked with their size, shape, and crystal structure. The bulk Co hexagonal (hcp) and cubic (fcc) phases exist, with the hcp crystal structure being favored at low temperatures. Entering the nanoscale, where surface freeenergy considerations guide the periodic arrangement of the Co atoms into a crystal lattice, the situation is reversed and the fcc phase is more stable than the hcp crystal configuration.8,9 Straightforward approaches to obtain metallic Co nanoparticles involve the reduction of cobalt metal salts to zerovalent metallic cobalt. The evolving nanoparticulate phase, synthesized well below the hcp-to-fcc phase-transition temperature is strongly dependent on the solvent properties, the reducing agent, and the binding ligands, if present.10,11 In alkaline aqueous media, hydrazine typically has been used as a reducing agent, whereas in organic solvents strongly binding surfactants such as oleic acid or trioctylphosphine oxide (TOPO) trigger the resulting morphology of the Co nanocrystals.12−16 Within organic solvents, polyols are somewhat special, because they can simultaneously take the role of being a reducing agent and a binding ligand.9 In some cases of liquid-phase-synthesized Co particles, coexistence of the cubic and hexagonal phases is observed. © 2014 American Chemical Society

The phase formation and also the reduction and oxidation behaviors are influenced by the prevailing crystallite size. In the hundreds of nanometers size range, the oxidation of cobalt to Co3O4, as well as the reduction of Co3O4 to cobalt, proceed in two steps, with cubic CoO as an intermediate phase.17 Below a threshold size of ∼20 nm, however, the redox behavior of Cobased nanoparticles becomes unpredictable; it is not only governed by the fast kinetics, because of the limited size, but also the ordering of nanoparticles,18 oxide supports,19 the presence of subsurface carbon species,20 or even the electron beam.21 As a consequence, phenomena that do not occur in the bulk materials are now observed; for example, formation of hexagonal CoO20 and hollowing of nanoparticles due to Kirkendall effect.16,18 Such phenomena dramatically impact the materials performance in diverse applications. To illustrate that point, we will examine hexagonal CoO. It turns out that hexagonal CoO is much more stable than its cubic counterpart and, as such, cannot be readily reduced to cobalt. Therefore, in the realistic condition for Fischer−Tropsch synthesis, the rapid formation of hexagonal CoO at the surface of Co nanoparticles hampers their catalytic activity.20,22 Also, with respect to other applications, the nontrivial redox behavior of Co-based Received: December 16, 2013 Revised: February 21, 2014 Published: March 11, 2014 2086

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

fitted into a brass housing and sealed with a cap pressed against the top of the sample container by four screws. Tightly adapted alumina bricks surrounding the brass housing thermally isolate the measurement cell for better temperature control. A temperature sensor right at the bottom of the PEEK container controls the resistive heating of the reaction solution via a feedback loop. The resistive elements are embedded on the both sides of the brass housing. Ex Situ Characterization. PXRD patterns from the powders synthesized ex situ were collected on an X’Pert Pro powder diffractometer (PANalytical B.V., The Netherlands), equipped with a diffracted-beam curved graphite crystal monochromator, operating in reflection mode under constant irradiated area conditions with Cu Kα radiation (45 kV, 40 mA). PXRD patterns from powders in glass capillaries were also recorded at the Swiss−Norwegian Beamline (SNBL) BM01B at the European Synchrotron Research Facility (ESRF), with the same optics and detector as described below for the in situ PXRD measurements, using a wavelength of 0.50479 Å. XAS data of the corresponding powder pellets were measured from 7.6 keV to 8.6 keV in continuous scanning mode with two ion chambers: one before the sample and one after the sample. The powders were diluted in cellulose to compensate for self-absorption and pressed into pellets. Scanning electron microscopy (SEM) images were taken on a LEO microscope (Carl Zeiss AG, Germany) to monitor the shape and morphology of the nanoparticles. Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HR-TEM) investigations were performed on a Philips Model Tecnai F30 system at 300 kV. In Situ XAS/PXRD Studies. A schematic representation of the setup used for the in situ studies is given in Figure 1. The PXRD data were collected at the SNBL BM01B at the ESRF using a wavelength of 0.50541 Å. The beam spot size at the front window of the sample cell containing the reaction solution was ∼0.7 mm × ∼0.7 mm, and the beam passed through the sample over a length of 1 mm. The sample volume

nanoparticles is highly relevant. Thus, after years of synthetic efforts to tune the properties of Co-based nanoparticles,3,23−28 understanding the origins of the redox phenomena is gaining in importance.29 Experimental measurements using in situ X-ray absorption and diffraction techniques provide access to the dependence between the Co-based nanoparticles and their performance as a catalyst.5 However, the proper elaboration of a general strategy to design the function-specific Co-based nanoparticles and their hybrids is only possible when the dependence between the oxidation state of Co-based nanoparticles and their growth is well understood. Therefore, here, we present a simple method to synthesize Co nanoparticles with the oxidation state of cobalt ranging from 0 to 3+, which gives us an opportunity to simultaneously study the growth and redox behavior of nanoparticles. We show that exactly the same educts react at 80 °C to cobalt oxide nanoparticles, whereas, at 180 °C, they react to metallic cobalt nanoparticles. Surprisingly, the in situ X-ray absorption and diffraction measurements of the synthesis at 140 °C reveal considerable higher complexity of the redox reaction than usually anticipated for the nanoparticle synthesis in benzyl alcohol. We observe that oxidation reaction of Co2+ to Co3+/2+ and the reduction reaction of Co2+ to Co0 occur simultaneously. It is followed by a rapid formation of Co3O4 nanoparticles and its consecutive solid-state reduction to CoO. In parallel, metallic Co nanoparticles begin to grow. Finally, the Multivariate Curve Resolution (MCR) analysis of XAS data reveals the interdependence between the four aforementioned reactions and helps to elucidate how balancing between the reduction and oxidation of the precursor alters the final composition and size of Co-based nanoparticles.



EXPERIMENTAL SECTION Synthetic Procedures. Chemicals. Anhydrous benzyl alcohol (≥99%), cobalt(II) oxide (99.99%), and cobalt(II,III) oxide (99.5%) were supplied by Sigma−Aldrich, cobalt(II) isopropoxide by Alfa Aesar, absolute ethanol (≥99%) and acetone by J.T. Baker of technical grade. The chemicals were used without further purification. Synthesis. Cobalt(II) isopropoxide (0.5 mmol) was added in an oxygen- and water-free atmosphere to 5 mL of anhydrous benzyl alcohol. The precursor changes the color from pink to deep violet immediately after contact with the benzyl alcohol. The mixture was placed in a 10-mL vessel and heated in the CEM microwave at a heating rate of 1 °C/s for 30 min at 80, 100, 120, 140, 160, and 180 °C. The brownish precipitate was washed three times with ethanol and acetone and centrifuged. The powders were dried in a hood at room temperature. In Situ Synthesis. The in situ synthesis was performed analogous to microwave synthesis, described above but the reaction was heated in a specially designed in situ cell that allows heating of the reaction solution with the heating rate comparable to that of the CEM microwave. A quantity of 0.83 mL of the reaction solution was transferred to a polyether ether ketone (PEEK) container (with a total volume of 1.66 mL). The temperature of the cell was increased to 140 °C, at a heating rate of 1 °C/s, and maintained at that temperature for 220 min. In Situ Cell. The measurement cell is assembled from three distinct parts (see Figure 1). The container for the reaction solution consists of PEEK, confining the sampled volume at the bottom part and facing the fluorescence detector by a thin window (wall thickness of 0.5 mm). The PEEK container is

Figure 1. Illustration of the setup used at the Swiss−Norwegian Beamline (SNBL) at the European Synchrotron Radiation Facility (ESRF) for the mechanistic in situ studies. Time-resolved powder Xray diffraction (PXRD) and X-ray absorption spectroscopy (XAS) measurements are recorded. The X-ray beam is aligned to penetrate the cell at the bottom part, where the thickness of the cell-window allows the emission and efficient detection of fluorescence radiation by the fluorescence detector. The diffracted beam is detected behind the cell, using a two-dimensional (2D) X-ray detector, revealing the evolution of crystalline species with time. 2087

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

Figure 2. (a) X-ray diffraction patterns of nanopowders obtained at different reaction temperatures from 80 °C to 180 °C. References of cubic CoO, hexagonal cobalt, cubic cobalt, and Co3O4 were simulated from International Centre for Diffraction Data (ICDD) File Card Nos. 01-071-1178, 01071-4238, 01-071-4652, and 01-073-1701, respectively. (b) HR-TEM of Co3O4 synthesized at 80 °C. (c) SEM images of Co synthesized at 180 °C. Insets show different morphologies of cubic and hexagonal Co nanoparticles. Scale bar = 1 μm.



RESULTS Synthesis of Cobalt and Cobalt Oxide Nanoparticles. X-ray diffraction patterns shown in Figure 2a reveal that, at 80 °C, the reaction of cobalt isopropoxide with benzyl alcohol already results in cubic Co3O4 nanoparticles. Moreover, the nanoparticles are spherical in shape, with a size between 3 nm and 6 nm, and tend to assemble into large structures, as is shown in the HR-TEM image given as Figure 2b. Such a direct formation of Co3O4 in benzyl alcohol is very uncommon for nonaqueous chemical route43−45 and contradicts the idea that benzyl alcohol usually acts as a reducing solvent.46,47 Generally, the reaction of Co2+ precursors in nonaqueous solution result in CoO or Co(OH)2 nanoparticles that can transform to Co3O4 only after subsequent heating in air.3,23,28,48 An exception here is pyrolysis of cobalt fatty acid salts in nonaqueous solution to Co3O4.49 Interestingly, for other spinel oxides, the direct formation of Mn3O4 and Fe3O4 in benzyl alcohol from Mn2+ and Fe2+ precursors was previously reported, however, without any comment on the oxidation mechanism.48 To explore the mechanism of Co3O4 formation, we change the reaction temperatures up to 180 °C (Figures 2a and 2c). Surprisingly, even at a temperature of 140 °C, the Co3O4 phase is still stable and the average crystal size of nanoparticles remains unchanged (5 nm). We find that 140 °C seems to be a threshold temperature for oxidation of Co2+ to Co3+ in benzyl alcohol. Already at slightly higher temperature (160 °C), we do not observe any traces of Co3O4; however, the presence of mixtures of CoO and Co nanoparticles are evident. Moreover, at an even higher temperature (180 °C), we solely find metallic cobalt nanoparticles. However, according to the diffraction pattern shown in purple in Figure 2a, both cubic and hexagonal metallic Co phases are present. In the SEM image shown in Figure 2c, we see that the majority of nanoparticles exhibit spherical morphologies and only a small amount of nanoparticles have bipyramidal morphologies. These observations suggest that spherical cobalt nanoparticles with cubic crystal structure are the main product of the reaction at 180 °C. Moreover, detailed analysis of all diffraction patterns reveals traces of hexagonal cobalt phase at 46° 2θ (for a 1.5054 Å radiation source) in all powders, independent from the reaction temperature, and indicates either contamination of the precursor with metallic cobalt or, more likely, the formation of metallic cobalt in parallel to the main reaction.

was constantly irradiated during the time of reaction. A Dexela Model 2923 two-dimensional (2D) CMOS X-ray detector was installed at a distance of 408 mm from the sample, with a pixel size of 75 μm × 75 μm. The PXRD data were extracted from the 2D ring patterns by radial integration with Fit2D.10,30 XAS measurements were recorded with a Vortex EM fluorescence detector equipped with Xia digital electronics. The energy resolution of the Co K-edge X-ray absorption near edge structure (XANES) scans was 0.5 eV/step and also was maintained during the recording of the extended X-ray absorption fine structure (EXAFS) patterns. PXRD and XAS measurement were taken every 9.5 min. Data Handling. PEEK as an Internal Standard for PXRD. PEEK has found application as a material that can be used in Xray absorption for cell inlets and windows, because of its high chemical resistance, temperature, and pressure endurance and its low density (1.32 g/cm3).31−35 Here, even though it is a semicrystalline polymer, we also extend its application to PXRD, using its diffraction peaks as internal standard to compensate for the beam intensity changes. Further details are given in Figures 1 and 2 in the Supporting Information (SI). EXAFS. We perform the EXAFS data reduction using the ATHENA software.36 For the normalization, we use the preedge range from 7600.0 eV to 7637.0 eV post-edge from 7865.0 eV to 8592.0 eV. The value of E0 was set to 7715.8 eV, which is the position of the maximum of the first derivative for the first in situ scan and kept constant for the consecutive scans and powder references. The photon energy was converted to photoelectron wavenumbers k. The resulting χ(k)-function was k2 weighted and the range between 2.8 Å−1 and 8.9 Å−1 was Fourier-transformed, using a Hanning window function. MCR-ALS Method. XANES spectra in the energy range of 7690−7770 eV were analyzed with the MCR-ALS method.37−42 The number of components was chosen on the basis of Singular Value Decomposition (SVD) results. The initial spectra of components were estimated based on SIMPLISMA method with the noise level of 5%. ALS algorithm with following constraints was applied: (1) non-negativity of spectra and concentrations, (2) unimodality of concentrations (tolerance: 10%), (3) convergence criterion of 0.1. Further details are given in the SI. 2088

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

Figure 3. Reactions mechanism behind the synthesis of Co-based nanoparticles, as revealed by in situ studies.

We hypothesize that, when the temperature is sufficiently high, benzyl alcohol fully suppresses the oxidation of Co2+ to Co3+ and instead of it, rapidly reduces the precursor to Co0 in the solution. Within this picture, cubic cobalt nanoparticles nucleate directly from the solution rather than by the reduction of cobalt oxides. In the following, we validate this prediction by measuring the changes of oxidation state and crystal structure of Co-based nanoparticles during the synthesis. Tackling the Complex Reaction Mechanism via In Situ X-ray Diffraction and Absorption Studies. To elucidate the mechanism behind the oxidation and reduction of Co-based nanoparticles as they grow, we measure in situ X-ray absorption and diffraction during synthesis in benzyl alcohol at an intermediate temperature of 140 °C. Figure 3 gives an overview on the complexity of the reaction. In order to give an account for the underlying mechanism, we first analyze the oxidation state of Co and short-range ordering around Co atoms (Figures 4 and 5) and then the crystal structure and size of Co-based nanoparticles (Figure 6). Finally, we identify the dependence between different reactions by utilizing the MCR-ALS method to analyze XAS data (Figure 7). (a). Precursor. We analyze the XANES spectra measured at Co K-edge of the precursor and the reference compounds shown in Figure 4a to determine the exact oxidation state of the precursor; however, this turns out not to be straightforward. In general, the absorption edge position reflects the oxidation state of Co.50−55 In the case of Co isopropoxide powder, the edge position coincides with that of Co3+/Co2+ in Co3O4, while the edge position of Co isopropoxide dissolved in benzyl alcohol is shifted toward lower energies and is closer to that of Co2+ in CoO. To further explore the oxidation state of Co isopropoxide, we analyze the position and shape of the whiteline (the first resonance after the absorption edge). Remarkably, the position of the whiteline of Co isopropoxide dissolved in benzyl alcohol and CoO overlaps (marked as “B” at 7726.0 eV in Figure 4a) and can be precisely assigned to the electric-dipole-allowed transition of the 1s core electron to the unoccupied 4p bound state in Co2+(1s1c3d74p1),56,57 while for Co3O4, the whiteline is observed at 7730.0 eV and corresponds to the transition of the 1s core electron to the unoccupied 4p bound state in Co3+ (1s1c3d64p1, where c denotes a 1s core hole).57,58 Even though the crystal structure of Co isopropoxide is unknown, the X-ray diffraction pattern of the powder precursor

Figure 4. (a) Co K-edge XANES spectra of the solution at a reaction time of 0 min (cyan), 28.5 min (green), 38 min (blue), and 218.5 min (pink) (black and gray traces represent reference compounds (cobalt isopropoxides, Co foil, CoO, and Co3O4)). Dashed vertical lines emphasize the characteristic features of reference compounds: “A” (7709.5 eV) is the pre-edge feature, “B” (7726.0 eV) and “C” (7730.0 eV) are the whitelines of CoO and Co3O4, respectively; and “a” (7736.0 eV) and “b” (7741.5 eV) indicate the post-edge features of CoO. (b) Co K edge EXAFS spectra, where the k2-weighted spectrum is shown as a χ(k) function. The black lines in the different plots represent reference spectra of (1) cobalt isopropoxide, (2) Co3O4, (3) CoO, and (4) cobalt foil. (The red, green, and blue scans represent the in situ data after 9.5, 28.5, and after 38 min, respectively. The arrows indicate changes and first occurrences of different features during the reaction.) Reference plots (2)−(4) were scaled into the intensity range of the in situ data.

shows intense diffraction peaks at low 2θ angles, which are characteristic for complex structures of metal alkoxides, as shown in Figure 3 in the SI.59 No traces of metallic hexagonal cobalt are found, which rules out the idea that hexagonal cobalt phase is contaminating the initial precursor. We also notice that, as soon as Co isopropoxide is added to benzyl alcohol, it dissolves, which manifests in the absence of the diffraction peaks in the solution. (b). Multiple Oxidation States of Co during Reaction. The aforementioned Figure 3 in the SI evokes the complexity of the reaction mechanism. The in situ XANES spectra at Co K-edge shown in Figures 4 and 5 give the experimental insights into 2089

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

rapid changes of the oxidation state and the electronic configuration of cobalt species in benzyl alcohol at 140 °C. The precursor dissolved in solution, as it was discussed above, exhibits oxidation state Co2+. However, upon heating, we observe instantaneous changes of the spectra features. For example, we note that, after 38 min (blue scan in Figure 4a and red scan in Figure 5), the absorption energy edge is shifting toward higher energies of ∼1.5 eV, which indicates the oxidation of Co2+ to Co3+. Simultaneously, the pre-edge feature at 7709.5 eV caused by 1s → 3d transition, strongly increases, indicating reduction of Co2+ to Co0.50,56,57 To further elucidate these two contradictive observations, we probe the local environment of the Co atom, using EXAFS, as shown in Figure 4b. The data quality of Co K-edge EXAFS spectra of the cobalt species in solution decreases, in comparison with the quality of spectra of reference pellets. Nevertheless, the qualitative inspection of the EXAFS spectrum after 28.5 min of reaction suggests two components, Co and Co3O4, respectively (see Figure 4b, as well as Figure 4 in the SI). Different than the precursor in the solution, a maximum at values slightly higher than k = 2 Å−1, appears, followed by a plateau. We interpret this maximum as the formation of Co3O4. We also observe a small bump at k = 3.5 Å−1, which can be a feature of metallic cobalt.52,56 The position of the peak of the Fourier transform of the EXAFS function (see Figure 4 in the SI (green line)) assigned to the first coordination sphere (Co− O) is close to the corresponding peak in the Co3O4 standard but with smaller amplitude, because of the small particle size. This is consistent with what we expected from our ex situ experiment, shown in Figure 2a. Furthermore, from Figures 4b and 5, as well as Figure 4 in the SI, it is obvious that, after Co3O4 and Co are formed (after 28.5 min), the reaction does not terminate but, instead, proceeds even further. In the Fourier transform of the EXAFS, we observe a shift toward higher R-values in the first coordination sphere (Co−O), as shown in Figure 4 in the SI (blue line). In addition, the position of the features of the EXAFS function is similar to that one of the CoO standard with smaller amplitude, which is indicative of small particles (see Figure 4b) Particularly, we observe a shift of the maximum from 2 Å−1 to 1.75 Å−1, which suggests the formation of CoO. Other characteristics of CoO, such as a maximum at 4 Å−1 and the two minima at 5.5 Å−1 and 7 Å−1, are also visible; however, we note that the spectrum does not match perfectly with the reference spectra and the magnitude of the double maximum at 6 Å−1 suggests that metallic cobalt is also present.52,56 (c). Structure and Size of Nanoparticles. The changes in EXAFS spectra are rather small; therefore, to validate the mechanism of nanoparticles growth and phase transition introduced in Figure 3, we additionally investigate the X-ray diffraction patterns. To reveal these small changes, we subtract the scan after 28.5 min from subsequent scans, as shown in Figure 6a. This figure is only used to qualitatively illustrate the changes. It is apparent that, as the reaction progresses, several concurrent changes in the diffraction pattern are taking place. Based on the diffraction pattern of the “quenched” reaction products shown in Figure 6b, we assign the increasing diffraction peaks to cubic CoO and cubic Co (reddish colors) and, remarkably, the decreasing diffraction peaks to Co3O4 (bluish colors). Furthermore, the in situ studies imply that CoO is growing at the expense of Co3O4. Surprisingly, metallic Co grows independently of CoO and Co3O4.

Figure 5. In situ Co K-edge XANES spectra taken during synthesis of Co-based nanoparticles from cobalt isopropoxide in benzyl alcohol at 140 °C. Spectra taken at room temperature (RT), and after 38 min of the reaction at 140 °C are marked gray and red, respectively. Dashed lines emphasize the characteristic post-edge features of CoO.

Figure 6. (a) Relative changes of the diffraction patterns during the reaction at 140 °C. The pattern recorded after 28.5 min has been subtracted; thus, the positive peaks represent the increases of CoO and Co concentration and negative peaks represent decreases of Co3O4 concentration. This figure is only used to qualitatively emphasize the relative changes. (b) Ex situ patterns of the reaction products ((a) 1 h and (b) 3 h) and simulated reference patterns (Co (red), CoO (black), Co3O4 (blue)). (c) Size of nanoparticles based on the FWHM of the (111) and (200) peak for Co and CoO, respectively. (d) Changes of the peak height of Co(111) and CoO(200) and reference peak of PEEK. The peak of PEEK is used as an internal standard. The peak fitting of the Co(111) and CoO(200) reflections were based on the raw data.

2090

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

Figure 7. Results of MCR-ALS analysis of in situ Co K-edge XAS data: (a) determination of the number of independent components, based on singular value decomposition (SVD); (b) spectra of individual components determined by MCR-ALS analysis and measured spectra of cobalt isopropoxide, CoO, Co3O4, and Co foil. Dashed lines emphasize the characteristic pre-edge, edge, and post-edge features of reference compounds. (c) Interdependence between five Co components recovered by MCR-ALS.

To verify these correlations, we fit CoO(111) and Co(200) reflections, as shown in Figures 6c and 6d. Here, we use the unsubtracted data and, instead, use the intensity of PEEKs reflections as an internal standard. Unfortunately, the Co3O4 reflections overlap with the reflections of the internal standard, which disables the quantitative analysis of Co3O4. Please refer to the SI for further details on the fitting procedure. We use the Scherrer equation to calculate the crystallite size of CoO and Co, based on the full width at half maximum (FWHM) of the corresponding peaks, without correcting for instrumental broadening, as shown in Figure 6c. Interestingly, the initial size of CoO is 12 nm and just slightly increases to 14 nm during the reaction, whereas the intensity increases dramatically from 38 min to 76 min and then saturates, as shown in Figure 6d. The increase of the integral intensity of CoO peak directly reflects the increase of CoO concentration. In the case of the Co nanoparticles, we observed that the crystallite size is growing from 7 nm to 13 nm and their concentration steadily increases. We summarize that the concentration of CoO nanoparticles is rapidly increasing and then remains almost unchanged, whereas the size of the Co nanoparticles is steadily growing from 7 nm to 13 nm. This behavior implies that Co nanoparticles cannot be formed by the reduction of CoO but rather by nucleation from the solution. (d). Overview on the Reaction at 140 °C. The analysis of in situ PXRD helps to identify the crystalline species and reveal complex interdependence between them. XAS data provides the information on all speciesprecursor, crystalline nanoparticles, and intermediate species that are forming in the solutionand this information is often used in conjunction with linear combination analysis (LCA) to quantify the individual components.60 However, the electronic structure of nanoparticles strongly differs with size, crystallinity, and ligands at their surface.61−64 Thus, reference XANES spectra of Co3O4 powders, CoO powders and Co foil do not represent well the electronic structure of growing nanoparticles. In addition, for the intermediate species in the solution (for example, Co0), no reference spectra are available. Therefore, instead of LCA, we use the MCR-ALS method to analyze all of the information included in in situ XANES spectra recorded during the synthesis of Co-based nanoparticles, as shown in Figure 5. Here, we apply the MCR-ALS to recover response profiles from

the data matrix without any knowledge of correlations in the system. We estimate the number of components that should be retained for analysis on the basis of SVD results (for more details, see the SI). Five components explain the variance in the data well, as shown in Figure 7a. At more than five components, the eigenvalues start to align in a linear fashion and correspond to the noise in the data. We use SIMPLISMA algorithm, which is a variance-based method, to find the initial spectra of the components, whose concentration significantly differ from scan to scan, compared to other components. Thus, the spectrum of the single component, recovered by the ALS algorithm and shown in Figure 7b, do not necessarily directly correspond to the spectrum of the single chemical compounds. It is rather a combination of the compounds, which evolve together/at the same time as shown in Figure 7c and schematically in Figure 3. Here, component 1 is Co2+ isopropoxide dissolved in benzyl alcohol. Components 2−5 do not represent a single chemical compound but, rather, four reactions. Component 2 is the oxidation reaction of Co2+ to Co3+/2+. Component 2 appears while we observe the formation of Co3O4. However, in the moment of nucleation, Co3O4 does not consume the entire cobalt isopropoxide in the solution. Therefore, the spectrum of component 2 is a superimposition of Co3O4 and cobalt isopropoxide. Component 3 represents the reduction reaction of Co2+ in the solution to Co0. This is underpinned by two facts: the increase of absorption at 7709.5 eV reflects reduction to Co0 and the absence of diffractions peaks, which are characteristic for Co, excludes the nucleation of Co nanoparticles. Spectators in this reaction are educts and products of components 2 and 4. Thus, the spectrum of component 3 displays, in addition to the features of Co0, also features that are characteristic for Co3O4 and CoO. Component 4 represents the solid-state reduction of Co3O4 to CoO. This reaction terminates after 76 min, which is the time at which all the Co3O4 is consumed. Consistently, PXRD data after 76 min of reaction show that no further increase of size or concentration of CoO occurs. Spectators in this reaction are educts and products of components 3 and 5. Thus, the spectrum of component 4 contains some small features of metallic cobalt. 2091

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

oxide are significantly different. During the crystallization of Co nanoparticles, the reduction of precursor is very rapid and is followed by the steady growth of nanoparticles (see Figures 6c and 6d, as well as 3). During the crystallization of cobalt oxide (Co3O4) nanoparticles, both the oxidation of precursor and the nucleation of nanoparticles are very rapid. In addition, we observe that the formation of cubic CoO is not an independent reaction path, but rather a follow-up process of Co3O4 formation and its consecutive solid-state reduction. The underlying crystallization mechanism of nanoparticles cannot be explained with Classical Nucleation Theory,73,74 which predicts a unique activation energy for any given set of experimental conditions. Instead, the dispersive kinetic model, which takes into account the distribution of activation energies, gives a better description of our reaction.75,76 It is important to mention that both educts and products from the synthesis of Co nanoparticles are highly catalytically active. The representative reactions include the conversion of benzyl alcohol to benzaldehyde, toluene, and benzoic acid, as schematically drawn in Figure 8; however, this list of reported reactions is, by

Component 5 represents nucleation from solution, as well as the growth and precipitation of metallic Co nanoparticles. The spectator in this reaction is CoO, which is a product of component 4. Finally, the product of the reaction is a mixture of CoO and Co. Using MCR-ALS to analyze our in situ XAS data, we overcome a problem to differentiate educts and spectators from products of four parallel reactions and now can elucidate the complex reaction scheme previously shown in Figure 3. In the first stage, two parallel reactions are taking place: part of Co2+ (component 1) oxidizes to Co3+/2+ (component 2), and at the same time, another part of Co2+ reduces to Co0 (component 3). It is followed by a rapid formation of Co3O4 nanoparticles and its consecutive solid-state reduction to CoO (component 4). In parallel, metallic Co nanoparticles grow from solution (component 5). In addition, we do not observe the dependence between formation of metallic Co nanoparticles (components 3 and 5) and cobalt oxides (components 2 and 4), confirming the PXRD results that metallic nanoparticles form exclusively from the solution.



GENERAL DISCUSSION AND CONCLUSIONS Using in situ X-ray absorption and diffraction techniques, we demonstrate how benzyl alcohol alters the oxidation states of Co(II) isopropoxide, leading to the entire range of Co3O4, CoO, and Co nanoparticles. Our results highlight the essential role played by the initial interaction between precursor and solvent in determining the composition of Co-based nanoparticles. In addition, it was interesting to observe that the syntheses does not follow the expected classical nucleation reaction, but, instead, show time-dependent kinetics. In the following, we attempt to rationalize this effect. We generally observe that the reaction temperature determines the reactivity of Co isopropoxide with benzyl alcohol. In the first approximation, we can distinguish two main and independent reactions, leading to either metallic cobalt or cobalt oxide nanoparticles. At a high temperature of 180 °C, benzyl alcohol reduces cobalt(II) to metallic cobalt. So far, there have been no reports on the synthesis of metallic Co nanoparticles in benzyl alcohol. However, the reduction of Co isopropoxide is not surprising, since alcohols generally are widely utilized in the synthesis of Co nanoparticles from cobalt salts.65−68 The most studied are polyols, but also benzyl alcohol was recently used for the synthesis of Ni, Cu, Bi, and Ag nanoparticles.69−71 In this light, we think that the most likely reaction is the oxidation of benzyl alcohol to benzaldehyde, while reducing the cobalt ions.66,69 At a low temperature of 80 °C, benzyl alcohol partially oxidizes Co2+ to Co3+/2+. The oxidative behavior of benzyl alcohol is rather unusual, but it has recently been observed during the synthesis of cerium(IV) dioxide from cerium(III) isopropoxide.72 Even though the organic compounds, as a result of such an oxidation reaction, are unknown, it is reasonable to hypothesize that the oxidation of Co2+ is accompanied by a reduction of benzyl alcohol to toluene. However, the in situ experiment at an intermediate temperature of 140 °C reveals that the transition between these two reaction scenarios is not sharp and is dependent not only on the reaction temperature, as anticipated based in Figure 2, but also on the reaction time. At 140 °C, two chemical reactions in the solution that lead to the formation of metallic cobalt and cobalt oxide are concomitant. However, the nucleation and growth rates of metallic cobalt and cobalt

Figure 8. Schematic of the disproportionation reaction of benzyl alcohol (reaction 1.1), the catalytic reaction of benzyl alcohol over the cobalt catalyst (reactions 2.1 and 2.2), and the catalytic reaction of toluene over the cobalt catalyst (reaction 3.1).

far, not exhaustive.77−83 An alternative scenario is the disproportionation of benzyl alcohol to benzaldehyde and toluene (schematically shown in Figure 8, reaction 1.1),84 and their subsequent reaction with Co isopropoxides. Based on the results presented here, we conclude that, whichever are the additional catalytic reactions during the synthesis of nanoparticles, they are altering the reaction mechanism and, thus, the steady-state nucleation conditions are not fulfilled, leading to different rates of formation of Co, Co3O4, and CoO nanoparticles. In summary, the mechanism of the reaction can be quantitatively understood by the conjunction of the two parallel reactions: oxidation and reduction of Co2+, where organics present in the solution are reduced and oxidized, respectively. The determination of the interdependence between different inorganic species would not be possible without MCR-ALS analysis of in situ XAS data. Although studying the catalytic role of educts, products, and byproducts of the nanoparticles synthesis exceeds the scope of this paper, their impact on the nucleation reaction is evident.



OUTLOOK The beauty of this synthesis lies in the applicability of the same chemical educts to obtain Co-based products with different compositions and oxidation states, simply by adjusting the reaction temperature. These results are fundamental on the way 2092

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

(5) Jacobs, G.; Ma, W.; Gao, P.; Todic, B.; Bhatelia, T.; Bukur, D. B.; Davis, B. H. Catal. Today 2013, 214, 100−139. (6) Chu, W.; Wang, L.-N.; Chernavskii, P. A.; Khodakov, A. Y. Angew. Chem., Int. Ed. 2008, 47 (27), 5052−5055. (7) Guo, S.; Zhang, S.; Wu, L.; Sun, S. Angew. Chem., Int. Ed. 2012, 51 (47), 11770−11773. (8) Ram, S. Mater. Sci. Eng., A 2001, 304−306, 923−927. (9) Liu, B.; Guan, J.-g.; Wang, Q.; Zhang, Q.-j. Mater. Trans. 2005, 46 (8), 1865−1867. (10) de Silva, R. M.; Palshin, V.; de Silva, K. M. N.; Henry, L. L.; Kumar, C. S. S. R. J. Mater. Chem. 2008, 18 (7), 738−747. (11) Song, Y. J.; Modrow, H.; Henry, L. L.; Saw, C. K.; Doomes, E. E.; Palshin, V.; Hormes, J.; Kumar, C. S. S. R. Chem. Mater. 2006, 18 (12), 2817−2827. (12) Gibson, C. P.; Putzer, K. J. Science 1995, 267 (5202), 1338− 1340. (13) Erasmus, W. J.; van Steen, E. Ultrason. Sonochem. 2007, 14 (6), 732−738. (14) Sun, S. H.; Murray, C. B. J. Appl. Phys. 1999, 85 (8), 4325− 4330. (15) Puntes, V. F.; Krishnan, K. M.; Alivisatos, A. P. Science 2001, 291 (5511), 2115−2117. (16) Ha, D.-H.; Moreau, L. M.; Honrao, S.; Hennig, R. G.; Robinson, R. D. J. Phys. Chem. C 2013, 117 (27), 14303−14312. (17) Martin, M.; Koops, U.; Lakshmi, N. Solid State Ionics 2004, 172 (1−4), 357−363. (18) Yang, Z.; Lisiecki, I.; Walls, M.; Pileni, M.-P. ACS Nano 2013, 7 (2), 1342−1350. (19) Jacobs, G.; Ji, Y.; Davis, B. H.; Cronauer, D.; Kropf, A. J.; Marshall, C. L. Appl. Catal., A 2007, 333 (2), 177−191. (20) Papaefthimiou, V.; Dintzer, T.; Dupuis, V.; Tamion, A.; Tournus, F.; Hillion, A.; Teschner, D.; Haevecker, M.; KnopGericke, A.; Schloegl, R.; Zafeiratos, S. ACS Nano 2011, 5 (3), 2182−2190. (21) Yang, Z.; Walls, M.; Lisiecki, I.; Pileni, M.-P. Chem. Mater. 2013, 25 (11), 2372−2377. (22) Tsakoumis, N. E.; Dehghan, R.; Johnsen, R. E.; Voronov, A.; van Beek, W.; Walmsley, J. C.; Borg, Ø.; Rytter, E.; Chen, D.; Rønning, M.; Holmen, A. Catal. Today 2013, 205, 86−93. (23) Nam, K. M.; Shim, J. H.; Han, D.-W.; Kwon, H. S.; Kang, Y.-M.; Li, Y.; Song, H.; Seo, W. S.; Park, J. T. Chem. Mater. 2010, 22 (15), 4446−4454. (24) Sun, G.; Zhang, X.; Cao, M.; Wei, B.; Hu, C. J. Phys. Chem. C 2009, 113 (17), 6948−6954. (25) Wang, H.; Si, H.; Zhao, H.; Du, Z.; Li, L. S. Mater. Lett. 2010, 64 (3), 408−410. (26) He, T.; Chen, D. R.; Jiao, X. L.; Wang, Y. L. Adv. Mater. 2006, 18 (8), 1078−1082. (27) Sietsma, J. R. A.; Meeldijk, J. D.; den Breejen, J. P.; VersluijsHelder, M.; van Dillen, A. J.; de Jongh, P. E.; de Jong, K. P. Angew. Chem., Int. Ed. 2007, 46 (24), 4547−4549. (28) Yang, J.; Liu, H.; Martens, W. N.; Frost, R. L. J. Phys. Chem. C 2010, 114 (1), 111−119. (29) Huang, Y. Y.; Yao, T.; Sun, Z. H.; Wei, S. Q. J. Phys.: Conf. Ser. 2013, 430 (1), 012033. (30) Abdala, P. M.; Mauroy, H.; Van Beek, W. J. Appl. Cryst. 2014, 47, 449−457. (31) Koziej, D.; Rossell, M. D.; Ludi, B.; Hintennach, A.; Novak, P.; Grunwaldt, J. D.; Niederberger, M. Small 2011, 7 (3), 377−387. (32) Grunwaldt, J. D.; Ramin, M.; Rohr, M.; Michailovski, A.; Patzke, G. R.; Baiker, A. Rev. Sci. Instrum. 2005, 76, 054104. (33) Michailovski, A.; Grunwaldt, J. D.; Baiker, A.; Kiebach, R.; Bensch, W.; Patzke, G. R. Angew. Chem., Int. Ed. 2005, 44 (35), 5643− 5647. (34) Fulton, J. L.; Balasubramanian, M.; Pham, V. T.; Deverman, G. S. J. Synchrotron Rad. 2012, 19, 949−953. (35) Makosch, M.; Kartusch, C.; Sa, J.; Duarte, R. B.; van Bokhoven, J. A.; Kvashnina, K.; Glatzel, P.; Fernandes, D. L. A.; Nachtegaal, M.;

to controlling the oxidation states in the liquid-phase synthesis of nanoparticles, which is still one of the main challenges. Moreover, because nowadays the hybrids of metallic and oxide nanoparticles are replacing the single phase powders, we expect that our work will be relevant to several applications, including the design of electrodes for Li-ion batteries.85 Even though we have revealed the complexity of the inorganic reaction, important questions about the organic side of the reaction are still open. The most intriguing path is to determine the species formed in benzyl alcohol, which are the organic counterparts of the oxidation of Co2+. However, because of the aforementioned complexity of the reaction and fast reaction rate, conventionally used techniques such as Fourier transform infrared (FTIR) spectroscopy, ultravioletvisible (UV-vis) spectroscopy, and gas chromatography (GC) might not be sensitive enough. Thus, instead, we think that the total X-ray scattering and atomic pair distribution function analysis should help.86,87 This technique, although highly challenging, would allow investigation of the structure of Co clusters or complexes in the solutions.



ASSOCIATED CONTENT



AUTHOR INFORMATION

S Supporting Information *

Additional text and figures give details of MCR-ALS analysis, PXRD pattern of precursor and PEEK cell, in situ PXRD row data, PXRD peak fitting procedure (with respect to a PEEK internal standard), and FT representation of EXAFS data. This material is available free of charge via the Internet at http:// pubs.acs.org. Corresponding Author

*E-mail: [email protected]. Author Contributions §

These authors contributed equally.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the Swiss National Science Foundation (Nos. 200021_137637 and 200021_140581), ETH Zürich (No. ETH-05-12-2), and by the Electron Microscopy of ETH Zurich (EMEZ). We thank Prof. Markus Niederberger (ETH Zurich) for helpful discussions. Bruno Jörg (ETH Zurich) is greatly acknowledged for outstanding technical support in the designing and fabrication of the in situ cell, and Martin Süess (EMEZ) and Dr. Mor Baram (School of Engineering and Applied Sciences, Harvard University) are greatly acknowledged for their help with HR-TEM. We thank European Research Synchrotron Facility for beam time allocation, as well as Gopinathan Sankar UC London, whose in situ cell inspired our cell developments.



REFERENCES

(1) Zhang, M.; Jia, M.; Jin, Y.; Shi, X. Appl. Surf. Sci. 2012, 263, 573− 578. (2) Zheng, X. F.; Shen, G. F.; Li, Y.; Duan, H. N.; Yang, X. Y.; Huang, S. Z.; Wang, H. E.; Wang, C.; Deng, Z.; Su, B. L. J. Mater. Chem. A 2013, 1 (4), 1394−1400. (3) Xiao, X.; Liu, X.; Zhao, H.; Chen, D.; Liu, F.; Xiang, J.; Hu, Z.; Li, Y. Adv. Mater. 2012, 24 (42), 5762−5766. (4) Zhang, L.; Hu, P.; Zhao, X.; Tian, R.; Zou, R.; Xia, D. J. Mater. Chem. 2011, 21 (45), 18279−18283. 2093

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094

Chemistry of Materials

Article

Kleymenov, E.; Szlachetko, J.; Neuhold, B.; Hungerbühler, K. Phys. Chem. Chem. Phys. 2012, 14 (7), 2164−2170. (36) Ravel, B.; Newville, M. J. Synchrotron Radiat. 2005, 12 (4), 537− 541. (37) Jaumot, J.; Gargallo, R.; de Juan, A.; Tauler, R. Chemom. Intell. Lab. Syst. 2005, 76 (1), 101−110. (38) Márquez-Alvarez, C.; Rodríguez-Ramos, I.; Guerrero-Ruiz, A.; Haller, G. L.; Fernández-García, M. J. Am. Chem. Soc. 1997, 119 (12), 2905−2914. (39) Nunes, C. A.; Resende, E. C.; Aes, I. G. R.; Anastácio, A. S.; Guerreiro, M. C. J. Appl. Spectrosc. 2011, 65 (6), 692−697. (40) Carvalho, H. W. P.; Pulcinelli, S. H.; Santilli, C. V.; Leroux, F.; Meneau, F.; Briois, V. Chem. Mater. 2013, 25 (14), 2855−2867. (41) Ressler, T.; Wong, J.; Roos, J.; Smith, I. L. Environ. Sci. Technol. 2000, 34 (6), 950−958. (42) Conti, P.; Zamponi, S.; Giorgetti, M.; Berrettoni, M.; Smyrl, W. H. Anal. Chem. 2010, 82 (9), 3629−3635. (43) Garnweitner, G.; Niederberger, M. J. Mater. Chem. 2008, 18 (11), 1171−1182. (44) Niederberger, M.; Garnweitner, G. Chem.Eur. J. 2006, 12 (28), 7282−7302. (45) Kränzlin, N.; Ellenbroek, S.; Duran-Martin, D.; Niederberger, M. Angew. Chem., Int. Ed. 2012, 51 (19), 4743−4746. (46) Kotas, J.; Vesely, Y.; Pac, J. Chem. Zvesti 1974, 28 (5), 646−658. (47) Wen, Z. G.; Zheng, F.; Jiang, Z. R.; Li, M. X.; Luo, Y. X. J. Mater. Sci. 2013, 48 (1), 342−347. (48) Bilecka, I.; Djerdj, I.; Niederberger, M. Chem. Commun. 2008, No. 7, 886−888. (49) Jana, N. R.; Chen, Y. F.; Peng, X. G. Chem. Mater. 2004, 16 (20), 3931−3935. (50) Andonova, S.; de Á vila, C. N.; Arishtirova, K.; Bueno, J. M. C.; Damyanova, S. Appl. Catal., B 2011, 105 (3−4), 346−360. (51) Saib, A. M.; Borgna, A.; van de Loosdrecht, J.; van Berge, P. J.; Niemantsverdriet, J. W. Appl. Catal., A 2006, 312, 12−19. (52) Vralstad, T.; Oye, G.; Ronning, M.; Glomm, W. R.; Stocker, M.; Sjoblom, J. Microporous Mesoporous Mater. 2005, 80 (1−3), 291−300. (53) Choy, J. H.; Jung, H.; Yoon, J. B. J. Synchrotron Rad. 2001, 8, 599−601. (54) Morishita, M.; Ochiai, S.; Kakeya, T.; Ozaki, T.; Kawabe, Y.; Watada, M.; Sakai, T. J. Electrochem. Soc. 2009, 156 (5), A366−A370. (55) Chou, T. L.; Chan, T. S.; Chen, J. M.; Yamauchi, H.; Karppinen, M. J. Solid State Chem. 2013, 202, 27−32. (56) Choi, H. C.; Lee, S. Y.; Kim, S. B.; Kim, M. G.; Lee, M. K.; Shin, H. J.; Lee, J. S. J. Phys. Chem. B 2002, 106 (36), 9252−9260. (57) Sankar, G.; Sarode, P. R.; Rao, C. N. R. Chem. Phys. 1983, 76 (3), 435−442. (58) Tsai, Y. W.; Hwang, B. J.; Ceder, G.; Sheu, H. S.; Liu, D. G.; Lee, J. F. Chem. Mater. 2005, 17 (12), 3191−3199. (59) Bradley, D. C.; Mahrotra, R. C.; Rothwell, I. P.; Singh, A. Alkoxo and Aryloxo Derivatives of Metals; Academic Press: London, 2001. (60) Olliges-Stadler, I.; Stoetzel, J.; Koziej, D.; Rossell, M. D.; Grunwaldt, J.-D.; Nachtegaal, M.; Frahm, R.; Niederberger, M. Chem.Eur. J. 2012, 18 (8), 2305−2312. (61) Liu, X.; Bauer, M.; Bertagnolli, H.; Roduner, E.; van Slageren, J.; Phillipp, F. Phys. Rev. Lett. 2006, 97, 253401. (62) Jentys, A. Phys. Chem. Chem. Phys. 1999, 1, 4059−4063. (63) Welther, A.; Bauer, M.; Mayer, M.; Jacobi von Wangelin, A. ChemCatChem 2012, 4 (8), 1088−1093. (64) Behrens, S.; Bonnemann, H.; Matoussevitch, N.; Gorschinski, A.; Dinjus, E.; Habicht, W.; Bolle, J.; Zinoveva, S.; Palina, N.; Hormes, J.; Modrow, H.; Bahr, S.; Kempter, V. J. Phys.: Condens. Matter 2006, 18 (38), S2543−S2561. (65) Park, J.; Joo, J.; Kwon, S. G.; Jang, Y.; Hyeon, T. Angew. Chem., Int. Ed. 2007, 46 (25), 4630−4660. (66) Toshima, N.; Yonezawa, T. New J. Chem. 1998, 22 (11), 1179− 1201. (67) Fievet, F.; Lagier, J. P.; Blin, B.; Beaudoin, B.; Figlarz, M. Solid State Ionics 1989, 32−33, 198−205.

(68) Ung, D.; Soumare, Y.; Chakroune, N.; Viau, G.; Vaulay, M. J.; Richard, V.; Fievet, F. Chem. Mater. 2007, 19 (8), 2084−2094. (69) Tukhtaev, R. K.; Yukhin, Y. M.; Udalova, T. A.; Bokhonov, B. B.; Lyakhov, N. Z. In IFOST 2008: Proceedings of the Third International Forum on Strategic Technologies, 2008, Novosibirsk− Tomsk, Russia, June 23−29, 2008; pp 239−242 (ISBN 978-1-42442319-4). (70) Bokhonov, B. B.; Yukhin, Y. M. Russ. J. Inorg. Chem. 2007, 52 (6), 922−926. (71) Kränzlin, N.; Niederberger, M. Adv. Mater. 2013, 25 (39), 5599−5604. (72) Niederberger, M.; Garnweitner, G.; Ba, J. H.; Polleux, J.; Pinna, N. Int. J. Nanotechnol. 2007, 4 (3), 263−281. (73) Lamer, V. K. Ind. Eng. Chem. 1952, 44 (6), 1270−1277. (74) Lamer, V. K.; Dinegar, R. H. J. Am. Chem. Soc. 1950, 72 (11), 4847−4854. (75) Cölfen, H.; Mann, S. Angew. Chem., Int. Ed. 2003, 42 (21), 2350−2365. (76) Skrdla, P. J. J. Phys. Chem. C 2011, 116 (1), 214−225. (77) Adam, F.; Ooi, W. T. Appl Catal., A 2012, 445, 252−260. (78) Cicco, S. R.; Latronico, M.; Mastrorilli, P.; Suranna, G. P.; Nobile, C. F. J. Mol. Catal. AChem. 2001, 165 (1−2), 135−140. (79) Gross, B. H.; Mebane, R. C.; Armstrong, D. L. Appl. Catal., A 2001, 219 (1−2), 281−289. (80) Ilyas, M.; Saeed, M. Int. J. Chem. React. Eng. 2010, 8, DOI: 10.2202/1542-6580.2162 (ISSN 1542-6580). (81) Kotas, J.; Vesely, K.; Pac, J. Chem. Zvesti 1974, 28 (5), 638−645. (82) Goetz, R. W.; Orchin, M. J. Org. Chem. 1962, 27 (10), 3698−. (83) Meng, Y.; Liang, B.; Tang, S. W. Appl. Catal., A 2012, 439, 1−7. (84) Haffad, D.; Kameswari, U.; Bettahar, M. M.; Chambellan, A.; Lavalley, J. C. J. Catal. 1997, 172 (1), 85−92. (85) Zhang, H. G.; Braun, P. V. Nano Lett. 2012, 12 (6), 2778−2783. (86) Tyrsted, C.; Jensen, K. M. O.; Bojesen, E. D.; Lock, N.; Christensen, M.; Billinge, S. J. L.; Iversen, B. B. Angew. Chem., Int. Ed. 2012, 51 (36), 9030−9033. (87) Jensen, K. M. O.; Christensen, M.; Juhas, P.; Tyrsted, C.; Bojesen, E. D.; Lock, N.; Billinge, S. J. L.; Iversen, B. B. J. Am. Chem. Soc. 2012, 134 (15), 6785−6792.

2094

dx.doi.org/10.1021/cm500090r | Chem. Mater. 2014, 26, 2086−2094