Synchrotron Structural Characterization of Electrochemically

Aug 2, 2008 - Relationship between the Synthesis of Prussian Blue Pigments, Their Color, Physical Properties, and Their Behavior in Paint Layers...
0 downloads 0 Views 524KB Size
Download PDF
13264

J. Phys. Chem. C 2008, 112, 13264–13271

Synchrotron Structural Characterization of Electrochemically Synthesized Hexacyanoferrates Containing K+: A Revisited Analysis of Electrochemical Redox Paulo Roberto Bueno,*,† Fabio Furlan Ferreira,‡ David Gime´nez-Romero,§ Grazielle Oliveira Setti,*,† Ronaldo Censi Faria,| Claude Gabrielli,⊥ Hubert Perrot,⊥ Jose´ Juan Garcia-Jaren˜o,§ and Francisco Vicente§ Instituto de Quı´mica, UniVersidade Estadual Paulista, C. Postal 355, 14801-907, Araraquara, SP, Brazil, Laborato´rio Nacional de Luz Sı´ncrotron, C. Postal 6192, 13083-970, Campinas, SP, Brazil, UniVersidade Federal de Sa˜o Carlos, 13560-905, Sa˜o Carlos, SP, Brazil, UPR 15 du CNRS, Laboratoire Interface et Syste`me Electrochimique, 4 Place Jussieu, 75252 Paris, France, and Departament de Quı´mica Fı´sica. UniVersitat de Vale`ncia. C/ Dr. Moliner, 50, 46100, Burjassot, Vale`ncia, Spain ReceiVed: March 9, 2008; ReVised Manuscript ReceiVed: May 24, 2008

The presumably soluble KFe+3[Fe2+(CN)6] structure of electrochemically synthesized hexacyanoferrate materials (Prussian Blue) containing K+ ions was determined for the first time in this study. Prior to drawing conclusions from a structural analysis, the main goal was to make a precise analysis of the inferred soluble structure, that is, KFe+3[Fe2+(CN)6], which is frequently referred to in the literature as the final stable electrochemically synthesized structure. Indeed, a successful X-ray powder diffraction experiment using X-ray synchrotron radiation was made of a powder placed in a 0.5 mm diameter borosilicate glass capillary, which was obtained by removing sixty 90 nm thin films from the substrates on which they were prepared. However, the conclusions were highly unexpected, because the structure showed that the [Fe(CN)6] group was absent from ∼25% of the structure, invalidating the previously presumed soluble KFe+3[Fe2+(CN)6] structure. This information led to the conclusion that the real structure of Prussian Blue electrochemically synthesized after the stabilization process is Fe4[Fe(CN)6]3 · mH2O containing a certain fraction of inserted K+ ions. In fact, based on an electrogravimetric analysis (Gimenez-Romero et al., J. Phys. Chem. B 2006, 110, 2715 and 19352) complemented by the Fourier maps, it is possible to affirm that the K+ was part of the water crystalline substructure. Therefore, the interplay mechanism was reexamined considering more precisely the role played by the water crystalline substructure and the K+ alkali metal ion. As a final conclusion, it is proposed that the most precise way to represent the structure of electrochemically synthesized and stabilized hexacyanoferrate materials is Fe3+4[Fe2+(CN)6]3 · [K+h · OH-h · mH2O]. The importance of this result is that the widespread use of the terms soluble and insoluble in the electrochemical literature could be reconsidered. Indeed, only one type of structure is insoluble, and that is Fe4[Fe(CN)6]3 · mH2O; hence, the use of the terms soluble and insoluble is inappropriate from a structural point of view. The result of the presence of the [Fe(CN)6] vacancy group is that the water substructure cannot be ignored in the ionic interplay mechanism which controls the intercalation and redox process, as was previously confirmed by electrogravimetric analyses (Gimenez-Romero et al., J. Phys. Chem. B 2006, 110, 2715; Garcia-Jareno et al., Electrochim. Acta 1998, 44, 395; Kulesza, Inorg. Chem. 1990, 29, 2395). Introduction There is considerable interest in the study of molecularbased compounds, for example, hexacyanometallate materials, alternatively referred to as mixed transition metal valence compounds.5-10 Because of their mixed valence state, they present different spin and/or oxidation states of similar or identical transition metal atoms of the structure.5-9 For this reason, the electronic or magnetic state of the metal atoms at particular sites can be tailored by external stimuli, for example, magnetic,7,8 light,11,12 or pressure stimuli and/or electrochemical compositional modulation.13-15 A typical example of hexacyanometallate materials is Prussian Blue * [email protected]. † Universidade Estadual Paulista. ‡ Laborato ´ rio Nacional de Luz Sı´ncrotron. § Universitat de Vale ` ncia. | Universidade Federal de Sa ˜ o Carlos. ⊥ Laboratoire Interface et Syste ` me Electrochimique.

(PB) and its analogues,6,7,12-14 which are applied in a large variety of fields ranging from optical memories,12,16,17 magnetic,7,18,19 electrochromic13-15,20,21 and battery devices to biosensors22-24 and electrocatalytic systems.13 Several deposition methods are employed to obtain PB material. Among them, electrochemical synthesis is used in a series of applications,13 for instance, electrochromic and electroanalytical applications.13,15 However, the redox processes involving the structures of PB compounds are still under debate. Classically, two structures are considered in the redox processes and synthesis mechanism, that is, insoluble and soluble ones.13,25-27 Inferences about these two types of structures are clearly made in the analysis of the mechanisms of ionic interplay between K+ ions and the soluble PB structure after stabilization. The soluble structure is referred to as KFe+3[Fe2+(CN)6], that is, with a k/l ratio around unity in the general formula KhFek[Fe(CN)6]l · mH2O proposed elsewhere.1,28 Electrogravimetric techniques have been employed to gain a better under-

10.1021/jp802070f CCC: $40.75  2008 American Chemical Society Published on Web 08/02/2008

Synthesized Hexacyanoferrates Containing K+

J. Phys. Chem. C, Vol. 112, No. 34, 2008 13265

standing of the intercalation mechanisms from a detailed mass change standpoint.1,14,29 Therefore, different electrogravimetric studies have demonstrated the importance of H+ and H3O+ ions in the redox processes occurring in PB type compounds,1,3,14,29,31 showing that the water substructure cannot be ignored in the general picture of the process1 and that it is also very important in other types of mechanism, such as the changeover process.14 However, to the best of our knowledge, the crystal structure of electrochemically prepared PB (after stabilization) has never yet been determined or studied in detail by X-ray diffraction, although the chemically prepared structure is well-known and expressed as the insoluble Fe4[Fe(CN)6]3 · mH2O structure (in which a water substructure is more frequently considered10,32). For that reason, such studies are very important to determine the k/l ratio in the general formula,1 that is, KhFek[Fe(CN)6]l · mH2O, to confirm the presence of the water substructure and to infer details about the presumably soluble structure. In the electrochemical literature, the movement of H+ and H3O+ through the water substructure is rarely taken into account.1,2,14 Furthermore, with respect to the main structural framework, that is, Fe2+-CN-Fe3+ chains, the K+ ions are believed to occupy 4a sites, that is, the Fe3+ iron position in the final electrochemically stabilized structure. Therefore, it is logical to determine the k/l ratio from the intercalation process mechanism that controls the electrochemical redox process, as will be demonstrated below. Let us begin by describing the general historical picture outlined in the literature based on refs 25-27, 33 and 34, which are very extensive. The general mechanism is referred to as an established picture even today’s modern spectroelectrochemical studies. Therefore, the classical picture of the intercalation process in insoluble hexacyanometallate compounds, considered as a reversible intercalation reaction, is achieved by an electrochemical cycle, that is,25-27,33,34

Fe+34[Fe+2(CN)6]3 + 4K+ + 4e- S K4Fe+2[Fe+2(CN)6]3 (1) for K+ cationic intercalation. It is important to note that the above historical and classical statement does not take into account the water substructure of the PB compound. As a result, in the cycling voltammetric process, K+ is considered to intercalate in the interstices or in the lattice of cubic PB. Furthermore, although always written in a reversible form, eq 1 is considered to possess a parallel irreversible reaction that leads to the stabilization of a final compound called soluble PB. At the end, the stabilization process is presumed to lead to a stable conversion of Fe3+4[Fe2+(CN)6]3 (insoluble) to KFe+3[Fe2+(CN)6] (soluble) compound. In other words, it is believed that after the newly deposited PBsFe2+4[Fe2+(CN)6]3sis cycled in a solution containing potassium ions, and assuming that structural changes occur in the PB material (stabilization process), the insoluble structure is converted into a soluble one. It is especially important to note that during the structural changes PB presumably loses approximately one-quarter of the high-spin Fe3+ atoms in the structure and potassium ions occupy the interstitial 4a sites (intercalating in the place of lost Fe3+) in the structure of the insoluble PB, leading to another structure, called a soluble structure.25-27,33,34 The reversible electrochemical process related to the soluble structure is proposed as follows.25-27,33,34

KFe+3[Fe+2(CN)6] + K++e- S K2Fe+2[Fe+2(CN)6] (2) Equation 2 is known as a reversible intercalation process between the soluble PB structure and Everitt’s salt (ES).25-27,34 Thus, it is possible to observe that this classical and historical analysis ignores the water substructure and furthermore assumes that the Fe3+/Fe2+ ratio and the occupation factor of iron are equal to unity (k/l ) 1 in the general formula1 KhFek[Fe(CN)6]l), leading to a specific formula of soluble PB known as KFe+3[Fe2+(CN)6]. The most important and central idea here is that the intercalation mechanism assumes that K+ ions occupy interstitial sites in the structure due to the absence of Fe3+ ions in the positions of their 4a crystallographic sites.25-27,34 This central idea and interpretation of the process will be revised herein based on and supported by structural and electrogravimetric analyses.1,2,14,29 The only way to verify the validity of the model is to determine the structure of PB compound prepared electrochemically after stabilization, which has never been done before due to difficulties in obtaining enough material for a powder diffraction analysis. Therefore, the main purpose of this work was to determine the structure of soluble PB and then, based on the resulting structural information, reconsider the electrochemical intercalation mechanisms from a structural standpoint supported by modern ac and dc electrogravimetric techniques and experimental data.1,2,14,29 Experimental Section Details of the electrochemical preparation methodologies are given in refs 25, 35, and 36. Briefly, what is important to mention here is that the presumably soluble PB compound in question was obtained after stabilization by means of cyclic voltammetry around the insoluble Everitt’s salt T insoluble PB transition,1,33,35-37 that is, the classical cycling process. This mechanism (viewing of the redox cycling process), which is outlined in the Introduction of this paper, leads to the presumable so-called stable soluble PB structure (the irreversible process associated to eq 1). Then, assuming it was possible to obtain Fe4[Fe(CN)6]3 · mH2O in an initial step (based on the assumption of the classical mechanism25,33,35,36), subsequent voltammetric cycles (stabilization process) led to the presumable target KFe+3[Fe2+(CN)6] compound, that is, the soluble PB compound. This target compound was obtained with the main purpose of studying its crystal structure. Because the role of the water substructure was previously discussed,1 the results will now be supported by a structural analysis with some reviewed points based on the new information provided by the structural analysis. The voltammetric pattern reported by refs 1, 14, 35, and 36 was adopted as a quality control parameter, that is, a pattern presenting a narrow voltammetric peak around 0.1 and 0.2 V with reduction and oxidation peaks separated by no more than 30 mV (a classical pattern obtained after stabilization of the PB compound25). Materials with a distance greater than 30 mV between these peaks were rejected for the analysis. Under these quality control conditions, obtained by PB depositions on platinum substrates, about sixty 90 nm thin films were necessary to obtain enough material for a good structural analysis (which explains why it has always been so difficult to study electrochemically synthesized material). Note that using this methodology enabled us to accurately determine the stabilized structure that likely generates the narrow voltammetric peak pattern observed during in situ compositional variation studies.1,14 The presumably soluble PB structure obtained under this particular

13266 J. Phys. Chem. C, Vol. 112, No. 34, 2008 condition of synthesis and the details and consequences of the analysis of such structure will be discussed throughout this paper. The crystal structure of the PB compound was studied by means of synchrotron X-ray powder diffraction and Rietveld refinement. High-intensity diffraction data were collected at the X-ray Powder Diffraction (D10B - XPD) beamline38 of the Brazilian Synchrotron Light Laboratory (LNLS, Campinas, Brazil) positioned after a dipolar source. X-rays of 1.37862(2), 1.74156(2), and 1.76140(8) Å wavelengths were selected by a double-bounce Si(111) monochromator. A vertically collimated beam was used in our experiments on a ∼0.8 (vertical) × ∼2.0 mm (horizontal) spot on the sample. The experiments were performed in the vertical scattering plane, that is, perpendicular to the linear polarization of the incident photons. The wavelength and the zero point were determined from several well-defined reflections of the SRM640c silicon standard. The diffracted beam was analyzed with a pyrolitic graphite HOPG(002) and detected with a Na(Tl)I scintillation counter with a pulse-height discriminator in the counting chain. The incoming beam was also monitored by a scintillation counter for normalization of the decay of the primary beam. The sample was loaded into a 0.5 mm-diameter borosilicate glass capillary after being scraped off the thin films, and data were recorded at room temperature for 10 s at each 2θ in steps of 0.06° from 10 to 90°. The Rietveld refinement was performed using the General Structure Analysis System (GSAS)39 program. The initial structural model considered was the one proposed by Herren et al.32 and obtained from a neutron diffraction profile analysis, which provided a precise interpretation of the main structure and water substructure of the PB-type compound. The background was fitted using a 24-terms shifted Chebyschev function. The peak profile was modeled by a pseudo-Voigt profile function as parametrized by Thompson et al.30 with asymmetry corrections by Finger et al.40 and microstrain anisotropic broadening terms by Stephens.41 Data collected at ∼73 eV below the Fe K-edge (7.112 keV ) 1.7433 Å) were used to minimize the imaginary anomalous scattering term of iron. It should be mentioned that the X-ray diffractograms obtained for the three distinct wavelengths were refined simultaneously. The temperature factors (see the Uiso values in Table 2) for a given atom type were constrained to be equal to each other, with the exception of the O occupying the 24e site, which was constrained to have the same value of the C and N atoms. During the initial refinement cycles, only the background, scale factor, peak profile, and unit-cell parameters were allowed to vary. After convergence, a difference Fourier map was calculated in order to locate the K+ atom positions. The K+ atoms were added to the structure’s model and all the atomic positions, temperature factors, and occupancy factors were refined except for the Fe3+ ion. Soft constraints were used to limit the C-N bond distances within the range reported for ferricyanides from single crystal structural studies.42 Excellent fitting was attained without this constraint, albeit with doubtful values for some interatomic bond distances and angles. The weighting factor for the constraint was gradually reduced and set to zero during the final stages of the refinement. Results and Discussion Structural Analysis of Hexacyanoferrates Containing K+ Ions. The final refinement parameters are listed in Table 1, the atom positions in Table 2, and the corresponding bond distances

Bueno et al. TABLE 1: Statistical Quality Factors for the PB Structural Model without the K+ Ion and with the K+ Ion Present in 24e Sitesa

b

space group

Fm3jm

Fm3jm

quality factors

Fe4[Fe(CN)6]3 · mH2O

a (unit cell), Å volume, Å3 χ2 Rp Rwp RF2 cryst. size, Å

10.1783(3) 1054.4(1) 1.39 2.52 3.05 2.35 183(4)

Fe4[Fe(CN)6]3 · mH2Ob containing K+ 10.1780(8) 1054.3(2) 1.23 2.47 2.82 2.05 183(4)

a a is the unit cell parameter, and cryst. size is the crystallite size. Electrochemically synthesized and stabilized “soluble” PB

TABLE 2: Results of Rietveld Refinement of Electrochemically Synthesized and Stabilized Soluble PB Structure Containing K+ Alkali Metal Ionsa atom

site

x

y

z

occup.b

Uiso (Å2)

Fe(III) Fe(II) C N O O O K

4a 4b 24e 24e 24e 8c 32f 24e

0 0.5 0.304(3) 0.188(3) 0.218(6) 0.25 0.315(3) 0.325(7)

0 0.5 0 0 0 0.25 0.315(3) 0

0 0.5 0 0 0 0.25 0.315(3) 0

1 0.755(9) 0.755(9) 0.755(9) 0.245(9) 0.764(2) 0.062(1) 0.088(8)

0.014(2) 0.029(3) 0.054(3) 0.054(3) 0.054(3) 0.102(5) 0.102(5) 0.102(5)

a Estimated standard deviations are given in parenthesis in this and other tables of the text. b Occupancy of the sites.

TABLE 3: Selected Interatomic Distances and Bond Angles Calculated from Experimental X-ray Diffraction Pattern, Considering the Insoluble PB Structure As the Correct Modela interatomic distances (Å) Fe(III)-N Fe(III)-O(24e) Fe(II)-C C-N C-O(24e) O(8c)-O(32f)

2.06(7) 2.00(7) 1.86(6) 1.17(4) 1.22(4) 1.50(4)

bond angles (°) Fe(III)-O(24e)-C

180.000(0)

Fe(III)-O(24e)-N

180.000(0)

O(32f)-O(8c)-O(32f)

109.471(4)

a

All other bond angles, i.e., N-Fe(III)-O(24e), O(24e)Fe(III)-O(24e), and C-Fe(II)-C, Fe(II)-C-N, are equal to 90 or 180 degrees.

and angles in Table 3. Surprisingly, the final k/l ratio gives a number higher than unity. This is clearly incompatible with the presumably soluble PB structure proposed in the literature, in which the occupation factor is assumed to be equal to unity for ratios between 4a and 4b sites, i.e., it is assumed that k/l positions are equivalently occupied and that the K+ ions occupy the interstices of the structure (which is a spatial picture of the presumably soluble PB structure).25-27,34,43 This result is also in agreement with conclusions drawn by reference44 for analogous PB compounds. Thus, the value of the k/l ratio obtained here is critical to demonstrate the inexistence of a soluble structure, as will be discussed in detail later herein. Having concluded that the soluble structure really does not exist, the former presumably soluble structure obtained by electrochemical synthesis will be referred to hereinafter as “soluble” PB structure. Let us now continue to evaluate the structure of PB electrochemically synthesized compounds. The final observed, calculated, and differentiated X-ray powder diffraction pattern of the soluble PB structure is plotted in Figure 1 with the value

Synthesized Hexacyanoferrates Containing K+

Figure 1. Rietveld X-ray profile refinement of soluble PB structure measured with λ ) 1.37862 Å. (+) observed, (red line) calculated, and (blue line) difference. Various selected indexed reflections are indicated. The pattern indicates a k/l ratio of around 1.332(2), leading to a formula of Fe4[Fe(CN)6]3 · mH2Oa containing K+. Sample loaded into a 0.5 mm diameter borosilicate glass.

of k/l ratio calculated at around 1.325(2), leading to the general formula of KhFe4[Fe(CN)6]3 · mH2O, in which k and l are now specific for electrochemically synthesized and stabilized soluble PB structure. As is often the case, the standard deviations calculated by GSAS for the lattice parameters are lower than the true errors.45 We present the errors calculated by GSAS in the tables of this paper with the understanding that the systematic errors may be more than 1 order of magnitude higher than the calculated deviations. On the basis of this structural analysis, the water substructure must be considered (as was done by Herren et al.32) to exist in such a type of structure. Moreover, from the electrochemical point of view, this substructure was also found to be very important in the ionic interplay mechanism,46 that is, the intercalation mechanism that controls the redox processes. Therefore, with regard to the crystalline water substructure and its atomic positions, the model proposed by Herren et al.32 and calculated from a neutron diffraction analysis was considered a starting model. In their work,32 the location of water clusters was identified in the Fe4[Fe(CN)6]3 · mH2O insoluble structure,19,32 that is, a framework that differs from the one studied here only by the fact that K+ alkali metal ions are not taken into account in the structural model of Herren et al.32 Therefore, it is now important to consider the K+ position inside the electrochemically synthesized and stabilized soluble PB structure. This was done based on the knowledge of the substructure water model determined by Herren et al.32 It is important to emphasize that the structure of chemically prepared and stabilized hexacyanometallates without the alkali metal has also been exhaustively studied by Buser et al.10 using X-ray diffraction. Their structure does not significantly differ from the one obtained here in terms of the occupation factor (the differences involve chemical bonding distances). Hence, it is possible to deduce that the structure of the electrochemically prepared and stabilized PB compound is Fe4[Fe(CN)6]3 · mH2O containing K+ ions and not the one traditionally related to the mechanism involved in eq 2; in fact, there are actually Fe2+(CN)6 vacancies calculated to be around ∼25%. Furthermore, the structural model proposed here is supported by an analysis of experimental electrogravimetric data, which showed no iron loss during the electrochemical stabilization of PB compounds.35,36 This information leads to the obvious conclusion

J. Phys. Chem. C, Vol. 112, No. 34, 2008 13267 that the form of the stabilized structure is actually Fe4[Fe(CN)6]3 · mH2O (it is the insoluble structure) containing K+ alkali metal ions and not KFe+3[Fe2+(CN)6] · mH2O (the former soluble structure proposed in the literature). The final electrochemically synthesized structure, which is also the most stable one, is the insoluble structure and not the soluble one, as traditionally stated.25-27,33,34 In other words, after the cycling process, a stabilized PB compound was obtained. Its structure, determined by a structural analysis and supported by modern ac and dc electrogravimetric techniques35,36 concerning the presence of water substructure, is Fe4[Fe(CN)6]3 · mH2O containing K+ alkali metal ions. This means that there is no loss of high-spin Fe3+ during the stabilization process, indicating that there is no soluble PB or analogous structure. Indeed, the terms soluble and insoluble should be reevaluated. The fact that the two structures have a similar insoluble equilibrium constant (Ksp ) 10-40),27 that is, both soluble and insoluble structures are considered to be highly and equivalently insoluble, is consistent with the finding that the only feature that distinguishes these structures is the compound’s facility to be peptized by K+ ions.27 Indeed, there is no difference between these structures with respect to the main Fe2+-CN-Fe3+ chains. The entire difference between soluble and insoluble structures lies in the occupation fraction of the ionic entities in the water substructure, as will be clarified later herein. The value of the k/l ratio is more or less fixed at around ∼1.33 (which is equivalent to ∼25% of Fe(CN)6 vacancies), and the Fe3+ occupation in the structure does not change after the stabilization process. In other words, the electrogravimetric analysis combined with the present structural study allows us to state that there is no mass change relating to Fe3+ leakage into the electrolyte.36 In the structural model of Herren et al.,32 the (H2O)6 clusters occupy empty nitrogen sites of Fe(CN)6 vacancies, forming a kind of water substructure. When this structural model is considered for our PB-type material, the most important difference between our Fe4[Fe(CN)6]3 · mH2O containing K+ ion alkali metals and the Fe4[Fe(CN)6]3 · mH2O structure without K+ ions appears to be the difference observed in some structural factors (Fobs). Without considering the inclusion of K+ ions in the structure, a good fit can be obtained, albeit with higher values of the goodness-of-fit (χ2) and R factors (Rp, Rwp and RF2).47 According to several proposals in the literature,48 the X-ray diffraction pattern of a face-centered cubic lattice with partial occupation of octahedral sites should present the (111) reflection, as is the case of our PB compound. This reflection was better resolved, belonging entirely to the (hhh) family, when the measurement was taken with λ ) 1.76140 Å, thus minimizing the imaginary anomalous scattering term of iron. Although Table 1 indicates a slight change in the quality factors, the inclusion of K+ to the structure yielded better agreement factors. Due to the low scattering factor of water by X-rays, it is difficult to detail its crystalline substructure solely by X-ray diffraction analysis. Neutron diffraction experiments are being planned and will be conducted in a future study. However, the analysis of the Fourier map indicates that K+ is located in the Fe(CN)6 vacancies and that it is part of the water substructure. The presence of K+ in the water substructure is also supported by electrogravimetric analyses.1,2 Although it is very difficult to pinpoint the exact position of K+ in the Fe(CN)6 vacancies, that is, its position in the water substructure, some inferences are put forward here.

13268 J. Phys. Chem. C, Vol. 112, No. 34, 2008

Bueno et al. TABLE 4: Selected Interatomic Distances and Bond Angles Calculated from Our Results Considering the Soluble PB Structure Containing K+ Ions Proposed in the Present Work interatomic distances (Å) Fe(III)-N Fe(III)-O(24e) Fe(III)-K Fe(II)-C Fe(II)-K C-N C-O(24e) O(8c)-O(32f) O(32f)-O(32f)

1.92(3) 2.22(6) 3.30(7) 2.00(30) 1.79(7) 1.18(20) 0.88(9) 1.15(5) 1.88(8)

bond angles (degree) N-Fe(III)-O(24e)

90.000(0)

N-Fe(III)-O(24e)

179.966(0)

O(32f)-O(8c)-O(32f) O(8c)-O(32f)-O(32f)

109.471(4) 35.264(2)

TABLE 5: Molar Compositions of PB-type Structuresa insoluble PB Figure 2. Schematic representation of the soluble electrochemically synthesized and stabilized PB structure after the final refinement cycle. (green circle) Fe(III), (red circle) Fe(II), (magenta circle) C, (yellow circle) N, (blue circle) O, and (cyan circle) K. This figure shows one octant containing 5 atoms at sites 32f and 8c. The atoms in the other octants were removed for clarity. One Fe(CN)6 group was removed, thus revealing a vacant Fe(CN)6 site. The structure of the chemically prepared and stabilized insoluble PB is similar to this one, except for the presence of the K+ ions (and, of course, slightly different coordinates of the atoms).

At this point, it is important to clarify that the neutron diffraction profile analysis conducted by Herren et al.32 indicates two types of crystalline water clusters. The first type is a shell water structure (composed of oxygen in 24e sites), which is coordinated with the main structure framework by means of the Fe3+ atoms, which fill the empty nitrogen sites of the Fe(CN)6 vacancies. The second type of structural water is formed by oxygen occupying 8c and 32f sites, which is not coordinated to the PB-type main structural framework, that is, Fe2+-CN-Fe3+ chains. Figure 2 exemplifies the soluble PB structure containing K+ ions generated after the final refinement cycle. Several works in the literature discuss the possibility that alkali metal ions occupy interstitial positions close to the center of the cubes in PB-type structures.9,35,43,49 This assumption does not take into account the presence of coordinated or uncoordinated water molecules also occupying these interstices, which is clearly demonstrated herein. In contrast, our analysis based on a difference Fourier map drawn after the refinement, provides evidence that this ion should occupy a 24e position (0.325, 0, 0) close to the vacant Fe(CN)6 group (see Table 2 for more details) and around the coordinated water. Therefore, as expected, because the K+ ion is a cation and the 24e oxygen site (probably OH-) is in an anionic position in the coordinated water substructure, the 24e site occupied by K+ is close to that occupied by the 24e oxygen site of the water-coordinated substructure (see Table 2). This indicates a local disorder of the water substructure caused by the presence of the K+ ion, which can be better observed by comparing the oxygen interatomic distances in Tables 3 and 4. A comparison of the data in Tables 3 and 4 also reveals the markedly large interatomic distances between Fe(III)-O(24e) with (about 2.22 Å) and without (about 2.00 Å) the presence of K+ alkali metal. The differences are important because they enable one to infer that, in terms of local charge distribution, the K+ ion is equilibrated by the oxygen (probably OH-) of the coordinated water part. The local picture can be visualized as a formation of a K+(OH-) bonding structure. In general, the

K Fe(III) Fe(II) C N O

soluble PB 4 3 18 18 14

K Fe(III) Fe(II) C N O

a Comparison of soluble and insoluble calculated by X-ray diffraction profiles.

2.108 4 3.022 18.131 18.131 13.970 structural

models

structural analysis described here, which indicates that the K+ ions would occupy sites adjacent to Fe(CN)6 vacancies, was correctly inferred in a previous work.14 It is also very important to emphasize that, although the K+ ion occupies sites of coordinated water crystalline structure, its presence affects the whole structural framework. This point is crucial, because it can lead to distortions due to interatomic bonds related to C, N, and iron metals, as can be observed mainly with respect to the Fe-N, Fe-C, and Fe-O bonds in Table 4 in comparison with the values in Table 3. This particular aspect of the hexacyanoferrate compounds containing K+ ions is shown in Table 5, which compares the model proposed by Herren et al.32 for Fe4[Fe(CN)6]3 · mH2O (chemically prepared) without K+ ions and our structural model for the soluble PB compound, which considers K+ ions occupying 24e sites close to 24e sites of coordinated water crystalline structure. At this point, it is possible to conclude that the structure of the electrochemically synthesized and stabilized hexacyanoferrate compound containing K+ is very similar to that expected for the well-known insoluble structure, without K+ ions, in terms of the main structural framework, that is, occupation and positions of iron metal in the Fe2+-CN-Fe3+ chains. However, according to Table 1, the two structures differ (not insignificantly) in terms of density (d ) 1.842 g.cm-3 and d ) 1.691 g.cm-3 with and without K+ ions, respectively) and interatomic bond values, a difference that is attributed to the modification of the water substructure by the presence of K+ ions. This is very interesting because crystalline water can change tremendously, depending on the conditions of the electrochemical synthesis,25 such as pH and electrolyte environment conditions.1,2 The voltammetric pattern is likely affected by the synthesis25,35,36 only in terms of the water crystalline substructure, which can lead to a different occupation and charge compensation. Furthermore, the water’s crystalline substructure changes locally due to the presence of K+ ions, although the complete details cannot be calculated exactly by X-ray diffraction (complementary analyses will be discussed in future works). However, it can be stated that the role of the K+ ions is to create local structural disorder (by modifying the water crystalline

Synthesized Hexacyanoferrates Containing K+

Figure 3. TEM and EDS analysis of electrochemically synthesized soluble PB structure. The EDS analysis corroborates the chemical quantitative analysis, showing less than 8% of chloride atoms in this compound. The inset on the right-hand side is the X-ray spectrum analysis, which is congruent with the X-ray diffraction analysis from synchrotron radiation. Measurements of the X-ray spectrum in different parts of the sample indicate that the compound is homogeneous.

substructure). The interatomic bond distances given in Tables 3 and 4 are mean values of the whole structure obtained for three distinct wavelengths refined simultaneously. However, they probably reflect this local structural disorder and its effect on the total bond distance and density of the PB-type structure containing K+ alkali metal ions. Complementary to the structural analysis, a morphological study is depicted in Figure 3, showing the sample’s nanoscale homogeneity in a transmission electron microscopy image (TEM, Philips CM200). The sample’s morphology is congruent with the mean crystallite value of 183 Å obtained from the X-ray profile analysis, indicating that the compound really does have a nanostructured polycrystalline morphology. The X-ray scattering analysis confirms the crystalline homogeneity in the micro and nanoscale range. An energy dispersive spectroscopy (EDS) analysis was made using a Digital Spectrometer (Princeton Gamma Tech, Model PRISM), that showed the presence of a small number of chloride elements (Cl- ions). It is impossible, at this point, to pinpoint the location of the Cl- ions in the structure. However, according to the structural X-ray analysis made here (also supported by an electrogravimetric analysis), it is possible to affirm that these ions are not located inside the main structure of the PB-type compound, that is, Fe2+-CN-Fe3+ chains. Present in a minor amount (less than 8%), the Cl- ions can be ignored in the structural analysis, particularly because they do not play an important role in the ionic interplay mechanism.1,29 These ions originate from the synthesis methodology employed to obtain the insoluble structure, prior to the formation of the presumably soluble PB structure. More details are given in refs 1, 25, and 37. It is plausible to propose that Cl- ions may occupy positions inside the water’s uncoordinated or coordinated crystalline substructures. It should be noted that these chlorides do not participate in the dynamic charge compensation (in situ compositional variation) because they are undetectable by electrogravimetric techniques in the ac or dc modes.1,29 Assuming the latter picture, locally in these sites, that is, Fe(CN)6 vacancies, there will probably be three H+ vacancies (three OH- uncompensated additional crystalline water parts) and one Cl- ion for every three OH-. This reasoning is based on the fact that Fe(CN)6 vacancies have a 4- valence local charge, which must be compensated. This assumption is also congruent with the fact that the crystalline water part of the

J. Phys. Chem. C, Vol. 112, No. 34, 2008 13269 soluble structure is considered here to be nonstoichiometric, as is commonly observed in solid-state compounds. Ignoring the Cl-, which is reasonable since it does not participate in the interplay mechanism,1,2 the model of Herren et al.32 and the structural model proposed here indicate that the O/H ratio is about 6/8, that is, a fraction calculated from 1/1.33. This means that there are 6 O for every 8 H, evidencing uncompensated OH- and consistent with the fact that there is enough OH- in the water substructure to compensate the K+ positive ions, forming a K+(OH-) environment inside the water substructure. Table 5 compares the compositions extracted from the neutron diffraction analysis reported by Herren et al.32 and our structural model, which considers the presence of K+ ions in sites of coordinated water substructure present in the main PB-type structure. On the basis of the results presented in Table 5, it is possible at this point to propose an actual average composition for the electrochemically synthesized and stabilized PB-type compound, that is, Fe4[Fe(CN)6]3 · [K+hCl-xOH-h-x · (mH2O)]. There are different possibilities for compensating the K+ charge, one of which is Fe2+ in 4a sites (interchanged mechanism11,14). Another possibility is the existence of OH- in an equivalent concentration in the water substructure. This can also be inferred from the electroneutrality rule, as previously discussed, whereby the OHspecies would be necessary to stabilize equivalent K+ species in similar concentrations. However, the possibility that K+ charge is compensated by Fe2+ can be disregarded, because the Fe3+/Fe2+ ratio was calculated recently by resonant X-ray diffraction experiments50 and found a value of about 1.44. This value is very similar to the 1.33 value found here for Fe(CN)6 vacancies, which means that there is no Fe2+ compensating the electronic charged added by the K+ insertion. The fact that K+ is compensated by OH- is also evidence of the presence of K+ in the water crystalline substructure of PB compounds. On the basis of the latter assumption, the chemical formula of our compound would be considered Fe4[Fe(CN)6]3 · (K+h · OH-h · mH2O), with the Cl- in the structure ignored because it does not participate in the ionic interplay mechanism, according to the electrogravimetric analysis.1,2 Consequences of the Structural Analysis in the Ionic Interplay Mechanism. The main purpose of the present work was to determine the structure of electrochemically prepared and stabilized PB compound. This determination confirmed it to be Fe4[Fe(CN)6]3 · [K+h · OH-h · mH2O] containing about ∼25% of Fe(CN)6 vacancies comparatively to the formerly expected KFe+3[Fe2+(CN)6] soluble structure. A combined structural and electrogravimetric analysis allowed us to conclude that a soluble PB-type compound structure does not exist, since Fe3+ is not lost and K+ does not occupy a position in such sites. Hence, the intercalation and interplay mechanism must be reexamined to consider the water substructure existing in such compounds, even when it is electrochemically prepared. It is not possible to determine the value of the K+ occupation fraction (estimated here as ∼2), that is, the occupation of K+ in the structure or even the exact position of this alkali metal ion in the water crystalline substructure [here considered to be the 24e position (0.325, 0, 0)]. However, it is certainly possible to unequivocally state that K+ ions are part of the water crystalline substructure, according to the Fourier map analysis and electrogravimetric measurements taken during the stabilization of electrochemically synthesized PB compounds.35,36,51 However, despite the hindrance in pinpointing the exact position of the K+ ions, what can be unequivocally concluded

13270 J. Phys. Chem. C, Vol. 112, No. 34, 2008

Bueno et al.

is that there is only one type of structure: Fe4[Fe(CN)6]3 · mH2O. Also certain is the fact that, when electrochemically synthesized, the K+ ions are inserted in the water substructure during the stabilization process, leading to Fe4[Fe(CN)6]3 · [K+hCl-xOH-h-x · (mH2O)]. The h density number of this structure still needs to be determined in a future work using neutron diffraction, a specific type of analysis that has yet to be made of electrochemically synthesized PB compounds. However, the main structural framework, that is, the one formed by Fe2+-CN-Fe3+ chains, is very similar to the chemically synthesized structure, Fe4[Fe(CN)6]3 · mH2O, which the electrochemical literature refers to as the insoluble structure. In other words, the soluble structure depicted as K+Fe2+[Fe3+(CN)6]3 should be reexamined and revised. In general, the insoluble structure is assumed to be Fe3+4[Fe2+(CN)6]3 prior to the stabilization, in which the movement of H+ and H3O+ ionic entities inside the water substructure is rarely take into account.1,2,14 Actually, a realistic ionic interplay mechanism cannot be considered by ignoring the water substructure in the PB compounds, which is very important, for instance, in the control of electrochemical redox processes, that is, intercalation processes. Therefore, the ionic interplay mechanism proposed elsewhere1,52 can now be reevaluated more precisely by the following equations.

Fe+34[Fe+2(CN)6]3 · [K+h · OH-h · mH2O] + nK+ + ne- T Fe+34-nFe+2n[Fe+2(CN)6]3 · [K+h+n · OH-h · mH2O)] (3) Fe+34[Fe+2(CN)6]3 · [K+h · OH-h · mH2O] + nH+ + ne- T Fe+34[Fe+2(CN)6]3 · [K+h · OH-h-n · ((m + n)H2O)] (4) Fe+34[Fe+2(CN)6]3 · [K+h · OH-h · mH2O] + nH3O+ + ne- T Fe+34[Fe+2(CN)6]3 · [K+h · OH- · ((m + 2n)H2O)] (5) The consequences of a more precise analysis of the ionic interplay mechanism are very interesting from both the chemical and the electrochemical points of view and are totally consistent with the structural analysis. However, the most evident fact is that the chemical reactions and processes in electrochemically or chemically synthesized PB compounds cannot be rationalized without considering the water substructure. The physical or chemical properties of PB are closely connected to the water crystalline substructure attached to the main framework structure composed of Fe2+-CN-Fe3+ chains.11,12,14,53

soluble and insoluble in the electrochemical literature must be reconsidered insofar as they are employed to describe two different possible PB structures. In actual fact, the only type of structure that is insoluble is Fe4[Fe(CN)6]3 · mH2O. When this structure is electrochemically synthesized, the K+ ions are inserted into the water substructure during the stabilization process, leading to Fe3+4[Fe2+(CN)6]3 · [K+h · OH-h · (mH2O)]. Therefore, soluble and insoluble are inappropriate terms from a structural point of view. An analysis based on a difference Fourier map and supported by electrogravimetric measurements1,14,29 confirmed that the K+ ions occupy positions in the water crystalline substructure. However, a more precise analysis of the water substructure of electrochemically synthesized PB compounds (all strongly dependent on the processing conditions) is necessary, based on neutron diffraction analyses. Finally, it was proposed that the K+ in the water crystalline substructure is charge-compensated by OH- in this substructure, leading to Fe3+4[Fe2+(CN)6]3 · [K+h · OH-h · (mH2O)]. Therefore, the ionic exchange mechanism was reexamined based on the fact that there is no loss of Fe3+ ions into the solution during the redox process of stabilization of the material, an analysis supported by the actual structure determined in this study.35 The previous mechanism,1,2 which considers the water substructure as determinant in electro-insertion reactions, was confirmed, and details can now be better described from a more precise view based on and supported by the structure thus determined. It is important, therefore, to stress that a realistic electrochemical mechanism cannot be constructed ignoring the water crystalline substructure in which most of the K+ resides. Furthermore, the former picture of the electrochemical processes does not consider that the ratio of iron atoms in 4a and 4b sites remains fixed throughout the stabilizing electrochemical cycling process, a notion that was not confirmed by electrogravimetric analysis.35 Thus, the entire redox process and intercalation mechanism depends on the water crystalline substructure. The ionic interplay mechanism was reviewed here to draw a more precise picture, which is described in detail by eqs 3-5. Acknowledgment. This work was supported by the Sa˜o Paulo state research funding agency FAPESP. We thank LNLS for the use of its beam line. D. G.-R. acknowledges his position at the Generalitat Valenciana. D. G.-R., J. J. G.-J., and F. V. are grateful for the financial support of FEDER-CICyT project CTQ 2007-64005/BQU. Supporting Information Available: The crystallographic international file (CIF) is provided as Supporting Information of this work. This material is available free of charge via the Internet at http://pubs.acs.org.

Conclusions and Final Remarks The structural analysis in the PB compounds prepared electrochemically and stabilized after different voltammetric cycles indicates a Fe3+4[Fe2+(CN)6]3 · [K+h · OH-h · mH2O] structure very similar to that obtained by traditional chemical methods, proposed as Fe4[Fe(CN)6]3 · mH2O. This is the first time the electrochemically synthesized and stabilized PB structure is determined and studied in detail by X-ray powder diffraction and, in the present case, using an X-ray synchrotron light source. The inference of the existence of a k/l ratio higher than unity was reached, determining that the [Fe(CN)6] group is absent from ∼25% of the structural main framework composed of Fe2+-CN-Fe3+ chains. Hence, the widespread use of the terms

References and Notes (1) Gimenez-Romero, D.; Bueno, P. R.; Garcia-Jareno, J. J.; Gabrielli, C.; Perrot, H.; Vicente, F. J. Phys. Chem. B 2006, 110, 2715. (2) Gimenez-Romero, D.; Bueno, P. R.; Garcia-Jareno, J. J.; Gabrielli, C.; Perrot, H.; Vicente, F. J. Phys. Chem. B 2006, 110, 19352. (3) Garcia-Jareno, J.; Sanmatias, A.; Navarro-Laboulais, J.; Vicente, F. Electrochim. Acta 1998, 44, 395. (4) Kulesza, P. J. Inorg. Chem. 1990, 29, 2395. (5) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K. Science 1996, 272, 704. (6) Champion, G.; Escax, V.; Moulin, C. C. D.; Bleuzen, A.; Villain, F. O.; Baudelet, F.; Dartyge, E.; Verdaguer, N. J. Am. Chem. Soc. 2001, 123, 12544. (7) Hayami, S.; Gu, Z.-Z.; Shiro, M.; Einaga, Y.; Fujishima, A.; Sato, O. J. Am. Chem. Soc. 2000, 122, 7126.

Synthesized Hexacyanoferrates Containing K+ (8) Yamamoto, T.; Umemura, Y.; Sato, O.; Einaga, Y. J. Am. Chem. Soc. 2005, 127, 16065. (9) Bleuzen, A.; Escax, V.; Ferrier, A.; Villain, F.; Verdaguer, M.; Mu¨nsch, P.; Itie´, J.-P. Angw. Chem. Int. Ed. 2004, 43, 3728. (10) Buser, H. J.; Schwarzenbach, D.; Petter, W.; Ludi, A. Inorg. Chem. 1997, 16, 2704. (11) Sato, O.; Iyoda, T.; Fujishima, A.; Hashimoto, K. Science 1996, 271, 49. (12) Sato, O.; Einaga, Y.; Iyoda, T.; Fujishima, A.; Hashimoto, K. J. Electrochem. Soc. 1997, 144, L11. (13) de Tacconi, N. R.; Rajeshwar, K.; Lezna, R. O. Chem. Mater. 2003, 15, 3046. (14) Bueno, P. R.; Gimenez-Romero, D.; Gabrielli, G.; Garcia-Jareno, J. J.; Perrot, H.; Vicente, F. J. Am. Chem. Soc. 2006, 128, 17146. (15) Mortimer, R. J.; Reynolds, J. R. J. Mater. Chem. 2005, 15, 2226. (16) Sato, O. J. Photochem. Photobiol. C 2004, 5, 203. (17) Sato, O.; Einaga, Y.; Fujishima, A.; Hashimoto, K. Inorg. Chem. 1999, 38, 4405. (18) Gu, Z. Z.; Sato, O.; Iyoda, T.; Hashimoto, K.; Fujishima, A. Chem. Mater. 1997, 9, 1092. (19) Kumar, A.; Yusuf, S. M.; Keller, L. Phys. ReV. B 2005, 71, 2720. (20) Ellis, D.; Eckhoff, M.; Neff, V. D. J. Phys. Chem. B 1981, 85, 1225. (21) Rajan, K. P.; Neff, V. D. J. Phys. Chem. 1982, 86, 4361. (22) Garcia-Jareno, J. J.; Gabrielli, C.; Perrot, H. Electrochem. Commun. 2000, 2, 195. (23) O’Halloran, M. P.; Pravda, M.; Guilbault, G. G. Talanta 2001, 55, 605. (24) Zou, Y. J.; Sun, L. X.; Xu, F. Biosens. Bioelectron. 2007, 22, 2669. (25) Itaya, K.; Akahoshi, H.; Toshima, S. J. Electrochem. Soc. 1982, 129, 1498. (26) Mortimer, R. J.; Rosseinsky, D. R. J. Chem. Dalton Trans. 1984, 2059. (27) Neff, V. D. J. J. Electrochem. Soc. 1978, 125, 886. (28) Kim, K.; Jureviciute, I.; Bruckenstein, S. Electrochim. Acta 2001, 46, 4133. (29) Gabrielli, C.; Garcia-Jareno, J. J.; Keddam, M.; Perrot, H.; Vicente, F. J. Phys. Chem. B 2002, 106, 3182. (30) Thompson, P.; Cox, D. E.; Hasting, J. B. J. Appl. Crystallogr. 1987, 20, 79. (31) Garcia-Jareno, J. J.; Navarro-Laboulais, J.; Vicente, F. Electrochim. Acta 1996, 41, 835.

J. Phys. Chem. C, Vol. 112, No. 34, 2008 13271 (32) Herren, F.; Fischer, P.; Ludi, A.; Halg, W. Inorg. Chem. 1980, 19, 956. (33) Mortiner, R. J.; Rosseinsky, D. R. J. Electroanal. Chem. 1983, 151, 133. (34) Lundgren, C. A.; Murray, R. W. Inorg. Chem. 1988, 27, 933. (35) Gimenez-Romero, D.; Agrisuelas, J.; Garcia-Jareno, J. J.; Gregori, J.; Gabrielli, C.; Perrot, H.; Vicente, F. J. Am. Chem. Soc. 2007, 129, 7121. (36) Agrisuelas, J.; Gabrielli, C.; Garcia-Jareno, J. J.; Gimenez-Romero, D.; Gregori, J.; Perrot, H.; Vicente, F. J. Electrochem. Soc. 2007, 154, F134. (37) Rosseinsky, D. R.; Glasser, L.; Jenkins, D. B. J. Am. Chem. Soc. 2004, 128, 10472. (38) Ferreira, F. F.; Granado, E., Jr.; Kycia, S. W.; Bruno, D., Jr. J. Synchrotron Radiat. 2006, 13, 46. (39) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS), Los Alamos National Laboratory Report LAUR 86-748; Los Alamos National Laboratory: 2001. (40) Finger, L. W.; Cox, D. E.; Jephcoat, A. P. J. Appl. Crystallogr. 1994, 27, 892. (41) Stephens, P. W. J. Appl. Crystallogr. 1999, 32, 281. (42) Gravereau, P.; Garnier, E. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1984, 40, 1306. (43) Keggin, J. F.; Milles, F. D. Nature 1936, 137, 577. (44) Jeerage, K. M.; Steen, W. A.; Schwartz, D. T. Chem. Mater. 2002, 14, 530. (45) Post, J. E.; Bish, D. L., Rietveld refinement of crystal structures using powder X-ray diffraction data. In Modern Powder Diffraction, ReViews in Mineralogy, Bish, D. L.; Post, J. E., Eds. Mineralogical Society of America: Washington, D.C., 1989; Vol. 20, pp 277-308.. (46) de Tacconi, N. R.; Rajeshwar, K.; Lezna, R. O. J. Electroanal. Chem. 2006, 587, 42. (47) Toby, B. H. Powder Diffr. 2006, 21, 67. (48) Escax, V.; Bleuzen, A.; Cartier dit Moulin, C.; Villain, F.; Goujon, A.; Varret, F.; Verdaguer, M. J. Am. Chem. Soc. 2001, 123, 12536. (49) Escax, V.; Bleuzen, A.; Itie´, J. P.; Mu¨nsch, P.; Varret, F.; Verdaguer, M. J. Phys. Chem. B 2003, 107, 4763. (50) Ferreira, F. F.; Bueno, P. R.; Setti, G. O.; Gimenez-Romero, D.; Garcia-Jareno, J. J.; Vicente, F. Appl. Phys. Lett. 2008, 92, 264103. (51) Garcia-Jareno, J. J.; Navarro-Laboulais, J.; Sanmatias, A.; Vicente, F. Electrochim. Acta 1998, 43, 1045. (52) GarciaJareno, J. J.; NavarroLaboulais, J.; Vicente, F. Electrochim. Acta 1997, 42, 1473. (53) Sato, O. Acc. Chem. Res. 2003, 36, 692.

JP802070F