Structure Characterization and Properties of K-Containing Copper

Article pubs.acs.org/IC

Structure Characterization and Properties of K‑Containing Copper Hexacyanoferrate Dickson O. Ojwang,† Jekabs Grins,† Dariusz Wardecki,† Mario Valvo,‡ Viktor Renman,‡ Lennart Hag̈ gström,‡ Tore Ericsson,‡ Torbjörn Gustafsson,‡ Abdelfattah Mahmoud,§,⊥ Raphael̈ P. Hermann,§,∥ and Gunnar Svensson*,† †

Department of Materials and Environmental Chemistry, Arrhenius Laboratory, Stockholm University, SE-10691 Stockholm, Sweden Department of Chemistry, Ångström Laboratory, Uppsala University, Box 538, SE-75121 Uppsala, Sweden § Jülich Centre for Neutron Science JCNS and Peter Grünberg Institut PGI, JARA-FIT, Forschungszentrum Jülich GmbH, D-52425 Jülich, Germany ∥ Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, 37831 Tennessee, United States ‡

S Supporting Information *

ABSTRACT: Copper hexacyanoferrate, CuII[FeIII(CN)6]2/3·nH2O, was synthesized, and varied amounts of K+ ions were inserted via reduction by K2S2O3 (aq). Ideally, the reaction can be written as CuII[FeIII(CN)6]2/3·nH2O + 2x/3K+ + 2x/3e− ↔ K2x/3CuII[FeIIxFeIII1−x(CN)6]2/3·nH2O. Infrared, Raman, and Mössbauer spectroscopy studies show that FeIII is continuously reduced to FeII with increasing x, accompanied by a decrease of the a-axis of the cubic Fm3m ̅ unit cell. Elemental analysis of K by inductively coupled plasma shows that the insertion only begins when a significant fraction, ∼20% of the FeIII, has already been reduced. Thermogravimetric analysis shows a fast exchange of water with ambient atmosphere and a total weight loss of ∼26 wt % upon heating to 180 °C, above which the structure starts to decompose. The crystal structures of CuII[FeIII(CN)6]2/3·nH2O and K2/3Cu[Fe(CN)6]2/3·nH2O were refined using synchrotron X-ray powder diffraction data. In both, one-third of the Fe(CN)6 groups are vacant, and the octahedron around CuII is completed by water molecules. In the two structures, difference Fourier maps reveal three additional zeolitic water sites (8c, 32f, and 48g) in the center of the cavities formed by the −Cu−N−C−Fe− framework. The K-containing compound shows an increased electron density at two of these sites (32f and 48g), indicating them to be the preferred positions for the K+ ions. dyestuff.5 PBAs constitute a vast family of compounds with a general formula AxM[M′(CN)6]z·nH2O (M = e.g., Fe, Cu, Ni, Co, Ti, Zn, In, Ga, Cd; M′ = e.g., Fe, Ru, Os, Rh, Mn; A = a monovalent cation or NH4+).6−10 Their physical and chemical properties have been extensively studied and employed in various fields, including solid-state electrochemistry,11 electrochromism,12 hydrogen storage,13 electrocatalysis,14 cationic and electronic conductivity,15 charge storage devices,11,16 corrosion protection,17 waste recovery,18 molecular magnets,19 and optics.20 PB and PBAs are typically synthesized through water-based precipitation reactions that produce extremely fine powders.21 The structure of PB was first characterized by Keggin and Miles in 1936, on the basis of X-ray powder diffraction (XRPD) data, using the space group Fm3̅m and a ≈ 10.2 Å.9 For the alkali-free PB, a slightly modified structural model, currently the accepted one, was given in the 1970s by Buser et al. from single-crystal data.22,23 For single crystals, the symmetry is

1. INTRODUCTION The current energy scenario requires a progressive shift toward the use of renewable and more sustainable energy sources, posing also an increasing demand for new and improved electrical energy storage systems. The latter should offer better performances, combined with low costs and a safe operation. Potassium-ion secondary batteries with Prussian blue analogues (PBAs) as electrodes and aqueous solutions containing K+ as electrolyte have shown promising electrochemical results, including rapid kinetics, long cycle lives, and high cycling rates.1−4 In addition, they are comparatively environmental friendly and inexpensive to make. The electrical energy is stored by insertion and extraction of K+ in the electrodes during their electrochemical reactions. The chemical formula for alkali-free Prussian blue (PB), the so-called “insoluble form”, may be written as FeIII[FeII(CN)6]3/4·nH2O, where n = 3.5−4, and that of an alkali-containing PB, the so-called “soluble form”, may be written as AIFeIII[FeII(CN)6]·nH2O, where A = Li, Na, K, etc., and n = 1−5. Since the first report in 1724,5 PB has been produced industrially for centuries for use as pigment and © 2016 American Chemical Society

Received: February 8, 2016 Published: June 3, 2016 5924

DOI: 10.1021/acs.inorgchem.6b00227 Inorg. Chem. 2016, 55, 5924−5934

Article

Inorganic Chemistry

Figure 1. (a) Illustration of the cubic PB/PBA structure with space group symmetry Fm3̅m and a ≈ 10.2 Å. For a PBA, M = e.g., Cu, Ni, Co, Mn, Cr, Cd, Os, and Zn. Zeolitic water molecules are shown in the cavities formed by the Fe(CN)6 octahedra.10,24,40 (b) Fe(CN)6 groups were removed to reveal vacant [Fe(CN)6]4− sites. The vacant sites are occupied by coordinated water, completing the coordination sphere of surrounding metals. In (c) the inserted alkali ions are shown, while the zeolitic waters were excluded for clarity.

found to be Pm3̅m, but for the fine powders usually obtained it is best approximated by Fm3̅m. The locations of the water molecules in alkali-free PB were subsequently determined by Herren et al. from neutron powder diffraction data.24 The accepted general structure model for a PBA is illustrated in Figure 1a−c. The face-centered unit cell contains four formula units of AxM[M′(CN)6]z·nH2O. The structure includes M(NC)6 and M′(CN)6 octahedra that are linked by sharing of CN groups. The linked octahedra form a three-dimensional framework of linearly repeating −NC−M′−CN−M−NC− units (M and M′ = FeII/FeIII).22,25 The framework is similar to the octahedral frameworks of the cubic ReO3 and AxMO3 perovskite structures but with the O atoms in the structures replaced by larger CN groups.26 Voids or cavities in the framework constitute locations for water molecules and/or alkali ions. While only the Fm3m ̅ structure model is relevant here, note that a number of PBAs exhibit structures with different connections of octahedra and lower space group symmetries.26−31 It should also be remembered that most PB/ PBA structures are inherently disordered and that local ordering of structural motifs are to be expected.22,24,32 For alkali-free PB, FeIII[FeII(CN)6]3/4·nH2O, the 4a sites are fully occupied by Fe(III) ions, whereas the Fe(II) ions of Fe(II)(CN)6 octahedra occupy only 75% of the 4b sites. As the chemical formula suggests, the strong [FeII(CN)6]4− complex acts as a structural entity, and for every vacant Fe(II) site, six adjacent CN groups are also missing, implying that the 24e sites occupied by C and N atoms also have a site occupancy of 75%. The water molecules were found by Herren et al.24 to partially occupy three different sites; (i) the 8c site (1/4,1/4,1/4) at the center of the cavity of the framework coordinated by a cubeoctahedron of CN groups, (ii) a 32f site, with four of the sites forming a tetrahedron around the center of the cavity (8c site), and (iii) a 24e site located between the positions of C and N. The water in the cavities, (i,ii), has been designated as “zeolitic water”, while those at the 24e site, as “coordinated water”. The presence of both coordinated water and N of the CN groups implies a mixed octahedral coordination for the Fe(III) ions. The maximum number of water molecules in the unit cell is 14, 8 in the cavities and 6 on the 24e sites at the vacant CN groups. Structures of alkali-containing PB are less characterized, in particular, regarding locations of alkali ions and water. Some authors suggest that the alkali ions are distributed over sites forming a tetrahedron in the cavities,33 while others claim them to be located at the center of the cavities.34 Bueno et al.35 suggested in a recent study by synchrotron X-ray diffraction of a K-rich PB, synthesized electrochemically from a K-free one,

that K is not located in the cavities, but at a 24e position near that of the coordinated water. Full insertion of K+ into insoluble PB gives Prussian white (PW), according to the equation11,36−38 PB(Fe III 4[Fe II(CN)6 ]3 ) + 4K+ + 4e− ↔ PW(K4Fe II 4[Fe II(CN)6 ]3 )

(1)

The structure of PW is very similar to that of PB, except that all Fe atoms are 2+, and each cavity contains one K atom.9 The insertion of alkali ions is accompanied by a change in stoichiometry, electronic structure of the Fe, and crystal structure.33,39 Although numerous reports on the electrochemical processes associated with PB materials can be found in literature, there is very limited information on the structural changes upon transformation to PW. A PBA that shows good electrochemical properties, and is therefore considered for use as an electrode material in batteries, is copper hexacyanoferrate, CuII[FeIII(CN)6]2/3·nH2O (CuHCF).1 The electrochemical reaction with K+ ions can be expressed as Cu II[Fe III(CN)6 ]2/3 ·nH 2O + 2x /3K+ + 2x /3e− ↔ K 2x /3Cu II[Fe II xFe III1 − x(CN)6 ]2/3 ·nH 2O

(2)

It is of significant interest to understand the similarities and structural differences of the end members in this system. However, it seems not possible to synthesize K2x/3CuII[FeIIxFeIII1−x(CN)6]2/3·nH2O, x = 1.0 by the direct methods described for PB synthesis.41,42 These methods all yield single-phase CuII[FeIII(CN)6]2/3·nH2O. Nevertheless, reduction of FeIII and insertion of K+ can be achieved with K2S2O3 (aq) according to the reaction 2, where 0.0 ≤ x ≤ 1.0. In the present study we report on the synthesis of these compounds and their characterization via various analytical techniques.

2. EXPERIMENTAL METHODS 2.1. Characterization. All precursors were of analytical-grade purity, purchased from Sigma-Aldrich. The preparation of the compound CuII[FeIII(CN)6]2/3·nH2O was based on the direct method of synthesis.41 Equal volumes, 50 mL, of 0.08 M Cu(NO3)2 and 0.04 M K3Fe(CN)6 solutions were simultaneously mixed in 25 mL of distilled water under constant stirring at room temperature. The stoichiometric proportions of the resulting phase can be expressed as Cu 2 +(aq) + 2/3[Fe III(CN)6 ]3 − (aq) = Cu II[Fe III(CN)6 ]2/3 ·nH 2O(s)

(3) 5925


Article

Inorganic Chemistry Potassium ions were inserted into the CuII[FeIII(CN)6]2/3·nH2O (s) structure by reducing FeIII with 0.1 M K2S2O3 (aq) according to

of the two end compositions, x = 0.0 and x = 1.0, were collected at room temperature in the range of 3.5−40°. Structural analysis was performed by the Rietveld method,43 implemented in the TOPAS program.44

Cu II[Fe III(CN)6 ]2/3 · nH 2O + 2x /3·K+ + 2x /3·S2O32 − ↔ K 2x /3Cu II[Fe II xFe III1 − x(CN)6 ]2/3 · nH 2O + 1x /3· S4 O6 2 −

3. RESULTS AND DISCUSSION 3.1. Color and Morphology of Samples. The assynthesized sample, of nominal composition CuII[FeIII(CN)6]2/3·nH2O and an oxidation state of 3+ for all Fe, is tawny brown. During the insertion process of K+ ions, the color changes from tawny brown to deep claret for x = 1.0, of nominal composition K2/3CuII[FeII(CN)6]2/3·nH2O, indicating a successful reduction of FeIII to FeII upon addition of K2S2O3 (aq). A potentiometric titration of fresh CuII[FeIII(CN)6]2/3· nH2O with K2S2O3 (aq) shows a sharp decrease in redox potential for x = 1.0, which accords to a complete reduction of FeIII to FeII. Equivalent results were obtained when using Na2S2O3 (aq). Obtained SEM micrographs for the two end compositions are shown in Figure 2a,b. The samples consist of polydispersed particles 20−50 nm in size. The particles aggregate to form a porous network. It should be mentioned that despite numerous

(4) Three series of samples A, B, and C, with 0.0 ≤ x ≤ 1.0 and Δx = 0.2, were prepared. In the series, x = 1.0 corresponds to a 1:1 molar ratio of added K2S2O3 to Fe. To check the completeness of the reaction, samples with x = 1.2 were synthesized as well. Each synthesis was treated in a similar way, 10 min of stirring and 18 h of aging time, before filtration. The resulting precipitates were separated from the mother liquor through centrifugation and washed several times with distilled water and subsequently freeze-dried. The final products were crushed and milled carefully to give fine powders for further analysis. A JEOL JSM-7401F scanning electron microscope (SEM), operated at an accelerating voltage of 2 kV and with a working distance of 3 mm, was used to study sample morphologies. Prior to the analysis, the samples were dispersed on a carbon tape and mounted on an aluminum stub. Energy-dispersive X-ray spectroscopy (EDS) analyses were performed on a HITACHI TM3000 microscope to determine the cation compositions and for semiquantitative estimation of the C, N, and O contents. The powders were pressed into pellets to obtain flat dense surfaces and thereby improve the accuracy of the EDS analysis. In addition to EDS, inductively coupled plasma-optical emission spectroscopy (ICP) was used to confirm the cation compositions (Cu, Fe, and K). Quantitative analyses of C, H, and N were obtained through elemental combustion. The ICP and combustion analyses were performed by Medac Ltd., United Kingdom. Infrared spectra were recorded on a Varian 610-IR spectrometer equipped with a DTGS detector in the mid-IR range (400−4000 cm−1) using attenuated total reflectance. The spectra were acquired with a resolution of 4 cm−1 for 64 cumulative scans. Raman microspectroscopy was performed at room temperature with a Renishaw InVia Raman spectrometer equipped with a charge-coupled device detector. A solid-state laser, Renishaw, with an excitation wavelength of 532 nm and a maximum power of 500 mW, was used for all measurements. A rather low power of 0.1%, nominally 0.5 mW, was applied during the analysis of all the samples, and their exposure to the laser beam was minimized between subsequent acquisitions to prevent possible degradation of the materials. Prior to the analyses, the instrument was calibrated using the characteristic Raman line at 520.6 cm−1 from a Si (001) wafer. The acquisitions of the spectra were run in an extended scan mode to have a wide range of wavenumbers (100−3200 cm−1) for the samples. To obtain a satisfactory signal-tonoise ratio, the acquisition time for each scan was set to 20 s with the total number of scans fixed to 20. Images were also taken by the builtin optical microscope before and after each analysis to assess surfaces of the various samples after probing. Mössbauer measurements were performed at room temperature on a spectrometer with a constant acceleration type of vibrator and a 57 CoRh source. The samples were ground and mixed with BN to obtain a concentration of 25 mg/cm2 for the absorber. Calibration spectra were recorded from an Fe metal foil. The resulting spectra were folded and analyzed using the least-squares Mössbauer fitting program Recoil. The thermal stability, dehydration temperature, and water content of the synthesized compounds were studied with a PerkinElmer TGA7 thermogravimetric analyzer (TGA) under a flowing N2 gas atmosphere at 30 to 600 °C and with samples placed in Pt crucibles. Room-temperature XRPD patterns were recorded in Bragg− Brentano geometry by means of a PANalytical X’pert PRO X-ray diffractometer operated at 45 kV and 40 mA and using CuKα1 (λ = 1.5406 Å) radiation. Silicon was used as an internal standard. The diffraction patterns were recorded in the 2θ range 10−70°, with a step size of 0.026° and total measuring time of 2 h. The samples were spread evenly on zero-background Si plates. High-resolution XRPD patterns were recorded at the ID22 beamline (λ = 0.4009 Å) of the ESRF synchrotron radiation facility in Grenoble, France. The patterns

Figure 2. SEM images of ground samples of (a) CuII[FeIII(CN)6]2/3· nH2O (x = 0.0) and (b) K2/3CuII[FeII(CN)6]2/3·nH2O (x = 1.0) obtained using an accelerating voltage of 2 kV. 5926


Article

Inorganic Chemistry

3.3. Mössbauer Spectra. Room-temperature 57Fe Mössbauer spectra were used to determine the oxidation states of Fe in the compounds. The spectra are shown in Figure 5, and

attempts where parameters such as concentration, pH, mixing speed, temperature and so on were varied, we have not succeeded to increase the size of the crystallites. Thus, it seems difficult to increase the particle size via this method.16 As seen in the SEM images, the size and morphologies of the crystallites do not change upon insertion of K+ ions. 3.2. Lattice Parameters. XRPD patterns of samples with 0.0 ≤ x ≤ 1.0, are shown in Figure 3. All peaks can be indexed with an F-centered cubic unit cell.

Figure 3. XRPD patterns of K2x/3Cu[Fe(CN)6]2/3·nH2O, 0.0 ≤ x ≤ 1.0 (sample series B).

Although the patterns are very similar, an increase of the 200 peak intensity relative to that of the 220 peak, together with small peak shifts to higher 2θ angles, can be observed with increasing x. The unit-cell parameter decreases from a = 10.1108(4) Å for x = 0.0 to a = 10.0267(4) Å for x = 0.8 as depicted in Figure 4. The decrease of a by 0.8% is a combined

Figure 5. Mössbauer spectra recorded at room temperature. The red and blue marked peaks represent FeII and FeIII contributions, respectively.

fitting parameters of their interpretation are presented in Table 1. The ascription of valence states of low-spin (LS) FeII and FeIII is based on knowledge of similar compounds from the literature.47,48 However, the assignments of nonmagnetic Mössbauer spectra are often nonconclusive, and there are several sets of parameters that can give a good fit with observed data. On the one hand, the spectrum for x = 0.0 reveals an asymmetric doublet structure, which we interpret as emanating purely from LS FeIII. On the other hand, the spectrum for x = 1.0 shows a single line, somewhat broader than the natural line width of 0.20 mm/s, and is ascribed to LS FeII. We consequently fitted all spectra using a model with one FeII doublet and one or two FeIII doublets. For x = 0.0 we used two FeIII doublets, with CS of −0.160(10) mm/s and −0.131(10) mm/s to fit the spectra, while for 0.2 ≤ x ≤ 0.8 the FeIII doublets for x = 0.0 could be merged into one doublet, in addition to the FeII doublet. In that way, the fitting results became very consistent as can be seen in Table 1. The CS for FeII is on the average of −0.093(10) mm/s throughout the whole series. The magnitude of the QS for FeII shows a small increase with increasing x, 0.195(10) mm/s for x = 1.0. For 0.2 ≤ x ≤ 0.8 the FeIII signal shows rather consistent CS values and constant QS, on the average of −0.159(10) mm/s and 0.523(10) mm/s, respectively. The two doublets can be accredited to FeIII in [FeIII(CN)6]3− and may be attributed to

Figure 4. Lattice parameter a vs x. The errors in a are smaller than the symbols.

effect of (i) a shortening of the Cu−N bond distance and (ii) the extra electron introduced to the Fe−C−N−Cu π bonding system upon reduction of [FeIII(CN)6]3− to [FeII(CN)6]4−.1,45 The unit-cell parameter variation with x also suggests that the reduction of FeIII to FeII might be essentially complete for x = 0.8, which is inconsistent with the performed potentiometric titration. Correlations between changes in composition and unit-cell parameter have also been reported in other studies of electrochemical insertion/extraction of Li+, K+, and Mg2+ ions in PBA electrodes and during redox reactions of PB.1,45−47 5927


Article

Inorganic Chemistry Table 1. Mössbauer Parametersa for K2x/3Cu[Fe(CN)6]2/3·nH2O, 0.0 ≤ x ≤ 1, Samples LS Fe x

CS

QS

II

LS Fe W

I

0.0 0.2 0.4 0.6 0.8 1.0

−0.092 −0.094 −0.092 −0.091 −0.097

0.143 0.149 0.156 0.182 0.194

0.254 0.245 0.251 0.239 0.246

43 64 81 95 100

III

CS

QS

W

I

−0.160 −0.131 −0.159 −0.157 −0.160 −0.160

0.521 0.250 0.533 0.572 0.520 0.470

0.245 0.241 0.266 0.279 0.317 0.280

65 35 57 36 19 5

a

CS = center shift (mm/s) vs natural Fe at 295 K, QS = absolute value of the electric quadrupole splitting (mm/s), W = full width at half-maximum (mm/s) of the individual Lorentzians. Errors in CS, QS, and W are 0.010 mm/s. I = relative spectral intensities with an error of 5%.

ions with different coordination environments, for example, structure distortions and Fe(CN)6 vacancies. The CS values observed here are in good agreement with those reported by Mizuno et al.47 in a study of the electrochemical insertion of Mg2+ in CuHCF. The compositional variation of the determined relative amount of FeIII shown in Figure 6 is very similar to nominal and what we observe for the unit-cell parameters for x < 1.0.

Figure 7. IR spectra (sample series B).

extended X-ray absorption near edge structure (XANES) studies indicate the presence of only CuII (Figure S1). For the as-synthesized CuII[FeIII(CN)6]2/3·nH2O, a sharp v(CN) band is found at 2178 cm−1, together with a very weak shoulder at 2093 cm−1, which can be assigned to FeIII−CN−CuII and FeII−CN−CuII links, respectively.52−54 With increasing x, there is a significant change in the relative peak intensities already for x = 0.2, as manifested by a strong increase of the 2093 cm−1 band intensity. As x increases for x = 0.4−1.0, the intensity of the 2178 cm−1 band diminishes, while that of the 2093 cm−1 band continues to increase. Similar correlations can be obtained based on the relative intensities of Fe−C stretching and Fe− CN linear bending modes. In the v(OH) region, there are four bands at 3645m, 3589w, 3380vs, b, and 3223m, sh cm−1. The first two bands indicate the presence of weakly hydrogen-bonded water molecules, while the last two indicate the presence of strongly hydrogen-bonded water molecules in the structure. In addition, the δ(HOH) band for the IR spectrum of x = 1.0 exhibits four component bands at 1678w, 1653vw, sh, 1635w, sh, and 1608s cm−1. Apart from the major vibrational bands mentioned above, other minor features were also observed at lower frequencies. The IR frequency values for the studied series are within the range of those reported by Avila et al.31 The information obtained from Raman is complementary to that of IR (see Figure S2). More than one Raman feature is, however, commonly associated with a single IR bandwidth and frequently with small shifts in energy.55 For PB and related compounds, the Fe cations experience ideally a local Oh point group symmetry with an inversion center. The three stretching

Figure 6. Relative amount of FeIII in % vs x from Mössbauer spectroscopy data (sample series B). Each data point has an error of 5%.

3.4. Infrared and Raman Spectra. The IR spectra of hexacynoferrates show four types of absorption bands from vibrations related to the octahedral [Fe(CN)6] unit: v(CN), δ(FeCN), v(FeC), and δ(CFeC), and two types of internal modes of zeolitic and coordinated water, v(OH) and δ(HOH). Free CN‑ has v(CN) of ∼2080 cm−1 in aqueous solutions.19 The CN− bind to the metal by a σ-bond by donating electrons from its weakly antibonding 5σ orbital to the metal and in a πbond by accepting electrons from the metal in a back-donation to its antibonding π-orbital. The σ-donation tends to raise the v(CN), while the back bonding tends to decrease it. Here, the latter effect is dominant. The final resulting shift in v(CN) depends on the electronegativity of the metals bound to the cyano-group and on their oxidation states and coordination numbers.19 The IR spectra of the investigated series are shown in Figure 7. The v(CN) band position in hexacyanoferrates allows unequivocal differentiation between ferrocyanides and ferricyanides.30,49−52 The v(CN) band positions in FeIII−CN−CuII and FeII−CN−CuII links are very close to 2178 and 2093 cm−1, respectively. Their intensities can thus be used to estimate FeIII and FeII contents in the samples. In other studies, the presence of FeII−CN−CuIII links have been suggested.51 However, our 5928


Article

Inorganic Chemistry modes are assigned to the A1g, Eg, and T1u vibrations.55,56 The latter band is Raman forbidden, and its presence is caused by structural distortions that lead to deviations from ideal Oh symmetry. For the x = 0.0, at least three bands can be observed in the v(CN) stretching region: 2194, 2157, and 2115 cm−1, which correspond to the A1g, Eg, and T1u modes, respectively. With increasing x, the relative intensities of the bands change. The observed band evolutions and peak shifts are similar to those reported for Na iron hexacyanoferrates.57,58 The peaks in the lower-frequency region (3.8 Å, indicating that the interaction between the Fe−CN−Cu framework and the zeolitic water and K in the cavities most probably takes place via the π-orbital system and not directly via the Fe and Cu cations. Finally, the distances between the O and K at the partially occupied sites 8c, 32f, and 48g range between 0.5 and 0.8 Å.

Figure 13. Iobs, Ical, and Iobs − Ical for the end compositions (a) x = 0.0 and (b) x = 1.0. The two patterns differ significantly in the intensities of 200 and 220 reflections.

C−N bond distance to ∼1.16 Å, which is within the range reported in the literature.22,31,35,67 For x = 0.0, Biso was fixed to 2 Å2 for the three zeolitic water positions, while the sofs of the corresponding O atoms as well as the x-coordinates for the 32f and 48g positions were refined. For x = 1.0, Biso was fixed to 1 Å2 for the three zeolitic water positions. The smallest residual intensities were obtained by distributing the amount of K from ICP analysis statistically over the 32f and 48g sites, with no K at the 8c site. The sofs for K were fixed to fit the ICP value, whereas the sofs of O atoms on the three zeolitic water positions and the x-coordinates for the 32f and 48g K/O atom positions were refined. A refinement where both the K and the O content were allowed to vary resulted in a K content being only 20% of that from the ICP analysis. The Rietveld analyses show that the main structural difference between the compounds is the higher electron density at the sites 32f and 48g for x = 1.0, the difference for the 8c site being small, see Tables 3a and 3b. It should be noted that the K content of 0.06 for x = 0.0 is too low at the 32f and 48g sites to be included in the refinement. Compositions corresponding to the refinements and from combined ICP and TG data are Cu[Fe(CN)6]0.70(1)·1.9(1)·H2O and K0.06(3)Cu[Fe(CN)6]0.67(3)·2.9(5)·H2O, and K2/3Cu[Fe(CN)6]0.68(1)· 2.2(1)·H2O and K0.61(3)Cu[Fe(CN)6]0.64(3)·3.8(5)·H2O, respectively. The water contents for the refined models are in comparison significantly smaller. Several explanations can be advanced, for example, that the water molecules are very disordered, as evidenced by fast kinetics for water exchange, and therefore mainly contribute to the amorphous background in the XRPD patterns. There are several reports in the literature suggesting that inserted alkali metal ions occupy the 8c site together with the zeolitic water.9,39,32 However, in this study the refinements and the difference Fourier map strongly suggest the O atoms to be distributed over the three sites 8c, 32f, and 48g and the K atoms over the two latter sites. The location of K atoms found here thus differs from the structure proposed by Bueno et al.,35 where K atoms occupy a 24e site, while it agrees with the structure described by Keggin and Miles.9

4. CONCLUSIONS Following the direct method of synthesis described in the literature, the compound CuII[FeIII(CN)6]2/3·nH2O was successfully prepared. In a following step, reductive titration by K2S2O3(aq) was employed to reduce FeIII to FeII and insert K+ into the CuII[FeIII(CN)6]2/3·nH2O to give K2x/3Cu[Fe(CN)6]2/3·nH2O. IR, Raman, and Mössbauer spectroscopy studies all reveal a reduction of FeIII to FeII with increasing x, as also shown by a continuous decrease of the unit-cell parameter. A reduction of FeIII to FeII is found to start immediately upon addition of K2S2O3 (aq), whereas ICP and EDS analyses of the K content show that the insertion of K+ starts first when a significant fraction, ∼20%, of the FeIII is already reduced. This indicates the presence of additional charge-compensating ions, thought to be H+ and OH−. The mechanisms involved in the insertion of K+ and associated reduction of FeIII are thus found to be chemically complicated and presently not fully understood. The SEM images show that sample morphology is independent of the composition. The samples consist of 20−50 nm sized agglomerated polydisperse particles that form a porous assembly of aggregates. The small particle size and porous nature of the aggregates is advantageous for use in battery electrodes, as it provides a high surface area to volume ratio that aids a rapid charge transport throughout the material. 5932


Article

Inorganic Chemistry

(8) Weiser, H. B.; Milligan, W. O.; Bates, J. B. Phys. Chem. 1942, 46, 99−111. (9) Keggin, J. F.; Miles, F. D. Nature 1936, 137, 577−578. (10) Gómez, A.; Reguera, E. Int. J. Inorg. Mater. 2001, 3, 1045−1051. (11) Neff, V. D. J. Electrochem. Soc. 1985, 132, 1382. (12) Bueno, P. R.; Gimenez-Romero, D.; Ferreira, F. F.; Setti, G. O.; Garcia-Jareno, J. J. Phys. Chem. C 2009, 113, 9916−9920. (13) Kaye, S. S.; Long, J. R. Catal. Today 2007, 120, 311−316. (14) de Lara González, G. L.; Kahlert, H.; Scholz, F. Electrochim. Acta 2007, 52, 1968−1974. (15) Kulesza, P. J. Inorg. Chem. 1990, 29, 2395−2397. (16) Yamada, S.; Kuwabara, K.; Koumoto, K. Mater. Sci. Eng., B 1997, 49, 89−94. (17) Deng, Z. J. Electrochem. Soc. 1991, 138, 1911−1918. (18) Ganzerli Valentini, M. T.; Stella, R.; Maggi, L.; Ciceri, G. J. Radioanal. Nucl. Chem. 1987, 114, 105−112. (19) Dunbar, K.R.; Heintz, R. A. Prog. Inorg. Chem. 1996, 45, 283− 391. (20) Moore, J. G.; Lochner, E. J.; Ramsey, C.; Dalal, N. S.; Stiegman, A. E. Angew. Chem., Int. Ed. 2003, 42, 2741−2743. (21) Davidson, D.; Welo, L. A. J. Phys. Chem. 1927, 32, 1191−1196. (22) Buser, H.; Schwarzenbach, D.; Petter, W.; Ludi, A. Inorg. Chem. 1977, 16, 2704−2710. (23) Buser, H. J.; Ludi, A. J. Chem. Soc., Chem. Commun. 1972, 1299. (24) Herren, F.; Fischer, P.; Ludi, A.; Halg, W. Inorg. Chem. 1980, 19, 956−959. (25) Giorgetti, M.; Berrettoni, M. Inorg. Chem. 2008, 47, 6001−6008. (26) Verdaguer, M.; Girolami, G. S. Magnetic Prussian Blue Analogs. In Magnetism Molecules to Materials V; Miller, J. S., Drillon, M., Eds.; Wiley-VCH: New York, 2005; Vol. 5, pp 283−346. (27) Kareis, C. M.; Lapidus, S. H.; Her, J.-H.; Stephens, P. W.; Miller, J. S. J. Am. Chem. Soc. 2012, 134, 2246−2254. (28) Ng, C. W.; Ding, J.; Gan, L. M. J. Phys. D: Appl. Phys. 2001, 34, 1188−1192. (29) Rodríguez-Hernández, J.; Reguera, E.; Lima, E.; Balmaseda, J.; Martínez-García, R.; Yee-Madeira, H. J. Phys. Chem. Solids 2007, 68, 1630−1642. (30) Reguera, E.; Balmaseda, J.; Quintana, G.; Gomez, A.; FernandezBertran, J. Polyhedron 1998, 17, 2353−2361. (31) Avila, M.; Reguera, L.; Rodríguez-Hernández, J.; Balmaseda, J.; Reguera, E. J. Solid State Chem. 2008, 181, 2899−2907. (32) Samain, L.; Grandjean, F.; Long, G. J.; Martinetto, P.; Bordet, P.; Strivay, D. J. Phys. Chem. C 2013, 117, 9693−9712. (33) Escax, V.; Bleuzen, A.; Cartier dit Moulin, C.; Villain, F.; Goujon, A.; Varret, F.; Verdaguer, M. J. Am. Chem. Soc. 2001, 123, 12536−12543. (34) Ludi, A.; Güdel, H. U. Inorg. Chem. 1973, 14, 1−21. (35) Bueno, P. R.; Ferreira, F.; Gimenez-Romero, D.; Oliveira Setti, G.; Faria, R. C.; Gabrielli, C.; Perrot, H.; Garcia-Jareno, J.; Vicente, F. J. Phys. Chem. C 2008, 112, 13264−13271. (36) Itaya, K.; Ataka, T.; Toshima, S. J. Am. Chem. Soc. 1982, 104, 4767−4772. (37) Lundgren, C. A.; Murray, R. W. Inorg. Chem. 1988, 27, 933− 939. (38) Agrisuelas, J.; García-Jareño, J. J.; Vicente, F. J. Phys. Chem. C 2012, 116, 1935−1947. (39) Bleuzen, A.; Escax, V.; Ferrier, A.; Villain, F.; Verdaguer, M.; Münsch, P.; Itié, J.-P. Angew. Chem., Int. Ed. 2004, 43, 3728−3731. (40) Ludi, A.; Güdel, H. U. Helv. Chim. Acta 1968, 51, 2006−2016. (41) Wilde, R. E.; Ghosh, N. S.; Marshall, B. J. Inorg. Chem. 1970, 9, 2512−2516. (42) Morrison, J. O.; Chatham, N. J.; Perkins, B. H. Manufacture of iron blue pigments. U.S. Patent 2,592,169, 1952. (43) Rietveld, H. M. Acta Crystallogr. 1967, 22, 151−152. (44) Coelho, A. A. Topas-Academic V5; Coelho Software: Brisbane, Australia, 2012. (45) Dostal, A.; Kauschka, G.; Reddy, S. J.; Scholz, F. J. Electroanal. Chem. 1996, 406, 155−163.

Thermogravimetrical studies show that the water content is strongly dependent on the ambient humidity and temperature but independent of the K content. The water exchange for the samples is very fast at room temperature. The structure of the CuII[FeIII(CN)6]2/3·nH2O appears to be very similar to that of K2/3Cu[Fe(CN)6]2/3·nH2O in terms of their main −Cu− NC−Fe−CN−Cu− framework. Difference Fourier maps provide evidence that the additional K+ ions in the latter structure share 32f and 48g sites with the zeolitic water, while only water is located on the 8c site. This finding differs from the structure model proposed by Bueno et al.,35 according to which the K+ ions are located on a 24e site close to a vacant [Fe(CN)6] site.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b00227. Preliminary results of the X-ray absorption near edge structure (PDF) X-ray crystallographic data (CIF) X-ray crystallographic data (CIF) Numeric data (PDF) Numeric data (PDF) Iobs, Icalc, and Diff data (PDF) Iobs, Icalc, and Diff data (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +08-164505. Fax: +08-152187. E-mail: gunnar. [email protected]. Present Address

⊥ LCIS/GREENMAT, Institute of Chemistry B63APTIS, Institute of Physics, Universitè de Liège, B-4000 Liege, Belgium.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the “Consortium for Crystal Chemistry, C3” within the Röntgen Ångström cluster, Swedish Research Council VR, for providing financial support. R.P.H. acknowledges support from the Materials Sciences and Engineering Division, Office of Basic Energy Sciences, U.S. Department of Energy. We appreciate valuable comments from Prof. J. Mink, Hungarian Academy of Sciences, Budapest, Hungary.

■

REFERENCES

(1) Wessells, C. D.; Huggins, R.; Cui, Y. Nat. Commun. 2011, 2, 550. (2) Soloveichik, G. L. Annu. Rev. Chem. Biomol. Eng. 2011, 2, 503− 527. (3) Wessells, C. D.; Peddada, S. V.; Huggins, R. A.; Cui, Y. Nano Lett. 2011, 11, 5421−5425. (4) Wessells, C. D.; Peddada, S. V.; McDowell, M. T.; Huggins, R. A.; Cui, Y. J. Electrochem. Soc. 2012, 159, A98−A103. (5) Brown, J. Philos. Trans. R. Soc. London. 1724, 33, 17−24. (6) Ludi, A.; Güdel, H. U.; Ruegg, M. Inorg. Chem. 1970, 9, 2224− 2227. (7) Kishore, D.; Gupta, G. P.; Lal, K. C.; Srivastava, T. N. Z. Phys. Chem. 1981, 126, 127−128. 5933


Article

Inorganic Chemistry (46) Mizuno, Y.; Okubo, M.; Kagesawa, K.; Asakura, D.; Kudo, T.; Zhou, H.; Oh-ishi, K.; Okazawa, A.; Kojima, N. Inorg. Chem. 2012, 51, 10311−10316. (47) Mizuno, Y.; Okubo, M.; Hosono, E.; Kudo, T.; Oh-ishi, K.; Okazawa, A.; Kojima, N.; Kurono, R.; Nishimura, S.-I.; Yamada, A. J. Mater. Chem. A 2013, 1, 13055−13059. (48) Okubo, M.; Asakura, D.; Mizuno, Y.; Kudo, T.; Zhou, H.; Okazawa, A.; Kojima, N.; Ikedo, K.; Mizokawa, T.; Honma, I. Angew. Chem., Int. Ed. 2011, 50, 6269−6273. (49) Pharr, C. M.; Griffiths, P. R. Anal. Chem. 1997, 69, 4673−4679. (50) Jones, L. H. Inorg. Chem. 1963, 2, 777−780. (51) Ng, C. W.; Ding, J.; Shi, Y.; Gan, L. M. J. Phys. Chem. Solids 2001, 62, 767−775. (52) Bertran, J. F.; Pascual, J. B.; Hernandez, M.; Rodriguez, R. React. Solids 1988, 5, 95−100. (53) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds; Wiley-Interscience Publication: New York, 1978; p 178. (54) Xia, L.; Mccreery, R. L. J. Electrochem. Soc. 1999, 146, 3696− 3701. (55) Kettle, S. F. A.; Diana, E.; Marchese, E. M. C.; Boccaleri, E.; Stanghellini, P. J. Raman Spectrosc. 2011, 42, 2006−2014. (56) Kettle, S. F. A.; Diana, E.; Stanghellini, P. L. Eur. J. Inorg. Chem. 2010, 2010, 3920−3929. (57) You, Y.; Yu, X.-Q.; Yin, Y.-X.; Nam, K.-W.; Guo, Y.-G. Nano Res. 2015, 8, 117−128. (58) Wang, L.; Song, J.; Qiao, R.; Wray, L. A.; Hossain, M. A.; Chuang, Y.-D.; Yang, W.; Lu, Y.; Evans, D.; Lee, J.-J.; Vail, S.; Zhao, X.; Nishijima, M.; Kakimoto, S.; Goodenough, J. B. J. Am. Chem. Soc. 2015, 137, 2548−2554. (59) Barsan, M. M.; Butler, I. S.; Fitzpatrick, J.; Gilson, D. F. R. J. Raman Spectrosc. 2011, 42, 1820−1824. (60) Samain, L.; Gilbert, B.; Grandjean, F.; Long, G. J.; Strivay, D. J. Anal. At. Spectrom. 2013, 28, 524−535. (61) Bueno, P. R.; Gimenez-Romero, D.; Gabrielli, C.; Garcıa-Jareno, J. J.; Perrot, H.; Vicente, F. J. Am. Chem. Soc. 2006, 128, 17146−17152. (62) Giorgetti, M.; Guadagnini, L.; Tonelli, D.; Minicucci, M.; Aquilanti, G. Phys. Chem. Chem. Phys. 2012, 14, 5527−5537. (63) Martínez-Garcia, R.; Knobel, M.; Reguera, E. J. Phys. Chem. B 2006, 110, 7296−7303. (64) Inoue, H.; Narino, S.; Yoshioka, N.; Fluck, E. Z. Naturforsch., B: J. Chem. Sci. 2000, 55, 685−690. (65) Ganguli, S.; Bhattacharya, M. J. Chem. Soc., Faraday Trans. 1 1983, 79, 1513−1522. (66) Ware, M. J. Chem. Educ. 2008, 85, 612−621. (67) Tokoro, H.; Shiro, M.; Hashimoto, K.; Ohkoshi, S. Z. Anorg. Allg. Chem. 2007, 633, 1134−1136. (68) Baerlocher, Ch.; Hepp, A. The X-ray Rietveld System (XRS-82), a set of computer programs for the Rietveld refinement of X-ray powder data; Inst. für Krist., ETH: Zürich, 1982. (69) Beall, G. W.; Milligan, W. O.; Korp, J.; Bernal, I. Inorg. Chem. 1977, 16, 2715−2718.

5934


Structure Characterization and Properties of K-Containing Copper

Recommend Documents