Structure Porous Copper Prussian Blue Analogues: Nature of Their

Within porous Prussian blue to copper the highest H2 storage capacity is observed. Such behavior finds explanation in the crystal structure for Cu3[M(...
0 downloads 3 Views 2MB Size
J. Phys. Chem. C 2010, 114, 5043–5048

5043

Structure of Porous Copper Prussian Blue Analogues: Nature of Their High H2 Storage Capacity J. Jime´nez-Gallegos,† J. Rodrı´guez-Herna´ndez,†,‡ H. Yee-Madeira,§ and E. Reguera*,†,‡ Centro de InVestigacio´n en Ciencia Aplicada y Tecnologı´a AVanzada-Unidad Legaria and Escuela Superior de Fı´sica y Matema´ticas, Instituto Polite´cnico Nacional, Me´xico, D.F., Me´xico, and Instituto de Ciencia y Tecnologı´a de Materiales, UniVersidad de La Habana, La Habana, Cuba ReceiVed: NoVember 4, 2009; ReVised Manuscript ReceiVed: January 26, 2010

Within porous Prussian blue to copper the highest H2 storage capacity is observed. Such behavior finds explanation in the crystal structure for Cu3[M(CN)6]2 with M ) Fe, Co, Ir. The crystal structure of Prussian blue analogues is usually solved and refined with a cubic unit cell in the Fm3jm space group, which corresponds to a random vacancy distribution. However, a careful evaluation of X-ray diffraction powder patterns of copper-containing compositions reveals a deviation from that structural model. The crystal structure for the considered series of copper hexacyanometallates(III) was found to be also cubic but in the Pm3jm space group related to a nonrandom vacancy distribution. To this model 50% of vacancies for the building block, [M(CN)6], corresponds, which is quite different from the value of 33.3% in the Fm3jm structural model. Mo¨ssbauer spectra and high pressure H2 adsorption isotherms support the assignment of the Pm3jm space group for the studied series of copper Prussian blue analogues. The implications of a nonrandom vacancy distribution on the physical properties of these materials are discussed. 1. Introduction Transition metal hexacyanometallates have received renewed interest in the past two decades, mainly due to their role as prototypes of molecular magnets1 and more recently in studies related to H2 storage.2–11 The CN ligand behaves as a bridging group linking neighboring metal centers. Its ability to subtract electrons from the metal linked at the C end, through a π-backdonation mechanism, and to locate the removed charge on the N end, explains the magnetic properties of hexacyanometallates. The CN group allows the interaction between the electron clouds of neighboring metal centers distant about 5 Å and cooperative magnetic interaction results. When the CN group coordinates only two metal centers (M, T) and the sTsNtCsMsCtNsTs chains are practically linear, a three-dimensional framework of cubic symmetry is formed. In such a structure the two metal centers have octahedral coordination and the compounds are known as Prussian blue (PB) analogues. Many PB analogues have a porous framework related to the existence of systematic vacancies of the building unit, the octahedral block [M(CN)6]. At the surface of a cavity created by a vacancy, six metal centers with an incomplete coordination sphere are found. In the as-synthesized material, these coordination sites are occupied by water molecules. The cavity filling is completed by additional water molecules hydrogen bonded to the coordinated ones. Both weakly bonded and coordinated water molecules can be removed by moderate heating, usually below 100 °C, preserving the solid framework.2,9 The resulting free space is appropriate for small molecule storage, among them H2.2–6,9,11 The availability on the surface of the cavity of metal centers with an open coordination sphere * To whom correspondence should be addressed. E-mail: ereguera@ yahoo.com. † Centro de Investigacio´n en Ciencia Aplicada y Tecnologı´a AvanzadaUnidad Legaria, IPN. ‡ Universidad de La Habana. § Escuela Superior de Fı´sica y Matema´ticas del IPN.

has stimulated the evaluation of this family of porous solids for H2 storage.2–6,9,11 Porous solids with exposed transition metal sites at the surface of cavities are being intensively studied for hydrogen storage.12–18 Prussian blue analogues are commonly obtained as fine powders, and their crystal structure must be solved and refined from X-ray diffraction (XRD) powder patterns. The crystal structures of PB analogues are usually described using a structural model based on a cubic unit cell in the Fm3jm space group.2–6,19,20 This model supposes a random distribution for the vacancies within the framework.19 For copper a deviation from that structural model has been observed. Its XRD powder pattern belongs to the Pm3jm space group,9,21 but the corresponding crystal structure has not been solved and reported. From a structural study using neutron diffraction of PB samples, Fe4[FeCN)6]3 · yH2O, recrystallized in concentrated HCl solution, the pattern indexing in the Pm3jm space group has been reported.22 Within divalent transition metal hexacyanocobaltates(III), the highest H2 storage capacity has been observed for copper,2,4 a behavior also observed for the iridium series,9 which could be related to the crystal structure adopted by the copper analogues. In this contribution the refined crystal structures for three copper hexacyanometallates(III) (Fe, Co, Ir) are discussed. The derived structural model for copper PB analogues is supported by Mo¨ssbauer spectra recorded for mixed Cu3-xFex[Co(CN)6]2 · yH2O samples and also by high pressure H2 adsorption isotherms. The implications of a nonrandom vacancy distribution on the physical properties of these materials are discussed. 2. Experimental Section The samples to be studied, Cu3[M(CN)6]2 · yH2O, were obtained mixing 0.01 M aqueous solutions of copper(2+) sulfate and of the involved potassium hexacyanometallates (M ) Fe,

10.1021/jp910544j  2010 American Chemical Society Published on Web 03/01/2010

5044

J. Phys. Chem. C, Vol. 114, No. 11, 2010

Jime´nez-Gallegos et al.

TABLE 1: Crystallographic Data for the Refined Structures data collection diffractometer wavelength, Å 2θ range (deg) step size (deg) time per step unit cell a (Å) formula units per cell refinement no. of contributing reflections no. of distance constraints no. of refined parameters structural parameters profile parameters Rp Rwp RB S

Cu3[Fe(CN)6] · yH2O

Cu3[Co(CN)6] · yH2O

Cu3[Ir(CN)6] · yH2O

Bruker D8 Cu KR, 1.541 83 5-100 0.025 18 Pm3jm 10.0779 (6) 1

Bruker D8 Cu KR, 1.541 83 5-100 0.025 18 Pm3jm 10.0379 (3) 1

Bruker D8 Cu KR, 1.541 83 5-100 0.025 15 Pm3jm 10.2909 (2) 1

142 5

140 5

150 5

34 8 5.04 7.06 8.13 1.4

34 8 4.64 6.48 9.70 1.39

34 8 5.38 6.89 9.15 1.28

TABLE 2: Hydration Degree and Dehydration Temperature (Estimated from TG Curves), and Frequency (in cm-1) for ν(CN) and ν(OH) (Coordinated and Hydrogen Bonded Water Molecules)a ν(OH) H2O compound

hydration degreeb

dehydration temp (°C)

ν(CN)

coordinated

hydrogen bonded

Cu3Fe2 Cu3Co2 Cu3Ir2

10 10 10.6

80 75 70

2174, 2104 W 2189 2191

3653 As, 3614 S 3615 As, 3608 S 3657 As, 3601 S

3384 Br 3455 As, 3446 S 3422 Br

a

Br, broad band. As, asymmetric; S, symmetric; W, weak. b Number of water molecules per formula unit.

Co, Ir) using a synthetic procedure already reported.9 By the same synthetic route, Cu3-xFex[Co(CN)6]2 · yH2O mixed samples were prepared. The mixed compositions were precipitated from aqueous solutions containing Cu(2+) and Fe(2+) in 8:2 and 9:1 atomic ratios. For comparison, Fe3[Co(CN)6]2 · yH2O was also prepared and its Mo¨ssbauer spectrum recorded. All the reagents used were of analytical grade from Sigma-Aldrich. The nature of the obtained samples was established from IR spectrocopy, X-ray energy-dispersive spectroscopy (EDS) analyses, Mo¨ssbauer spectroscopy, and thermogravimetric (TG) data. The metal atomic ratios estimated from EDS spectra agree with the expected formula units for the simple metal series. For the mixed series the iron content remains below its amount in the used solution, which was attributed to a greater affinity of the CN ligand for the copper atom. The XRD powder patterns were collected in Bragg-Brentano geometry using Cu KR radiation in a D8 Advance diffractometer (from Bruker). Some XRD powder patterns were also recorded under vacuum and on cooling at the XPD 10B beamline of the LNLS synchrotron radiation facility (Brazil). The experimental conditions are summarized in Table 1. The structural refinement from the collected XRD powder patterns was performed with the FullProf code,23 and the peak shape was modeled as a convolution of a peak asymmetry and a pseudo-Voigt function. Mo¨ssbauer spectra were recorded at room temperature using a constant acceleration spectrometer (from Wissel) operated in the transmission mode and a 57Co/Rh source. The obtained spectra were fitted using pseudo-Lorentzian peaks to estimate the values for isomer shift (δ), quadrupole splitting (∆), and line width (Γ). The values of δ are reported relative to sodium nitroprusside. H2 adsorption isotherms up to 7600 Torr (10 atm) were recorded using an ASAP 2050 analyzer (from Micromeritics). Sample tubes of known weight were loaded with an appropriate amount of sample, ∼50 mg, and sealed using TranSeal. Previous

to H2 adsorption, the samples were degassed on the ASAP analyzer using a heating rate of 5 °C/min and then maintained at the dehydration temperature indicated by the TG curve until to obtain a stable outgas rate below 1 µmHg. This process usually requires of 24 h of degassing. The degassed sample and sample tube were weighed and then transferred back to the analyzer (with TranSeal to prevent exposure of the sample to air). After volume measurement with He, the degassing was continued for 24 h at 60 °C in the sample port. Measurements were performed with a liquid N2 bath. 3. Results and Discussion A. Materials Behavior on Heating. The TG curves for the series Cu3[M(CN)6]2 · yH2O with M ) Fe, Co, Ir, has already been discussed,9 and here only a summary of their features is provided. These curves are available in the Supporting Information. The solids become anhydrous from about 80 °C. In Table 2 the estimated hydration degree and dehydration temperature are reported. Both coordinated and weakly bonded water molecules abandon the solid at relatively low temperature in a practically continuous process. Once the first hydrogen bonded water molecules are removed, the remaining ones enhance their mutual interactions and a higher temperature to allow their liberation is required. This explains the observed continuous weight loss for the TG curves. The dehydration temperature, which corresponds to the weight loss below 80 °C, shows a slight dependence on the inner metal, according to the following order: Ir < Co < Fe. This was attributed to the effective charge on the copper atom. The charge subtraction from the inner atom (Ir, Co, Fe) via π-back-bonding follows the inverse order, Ir > Co > Fe. The removed charge is located on the N end of the CN group and partially donated to the copper atom, reducing its polarizing power and weakening the metal interaction with the coordinated water molecules. This has been well documented from XRD, Mo¨ssbauer, magnetic, and TG data.24 An analogous

Structure Porous Copper PB Analogues

J. Phys. Chem. C, Vol. 114, No. 11, 2010 5045

Figure 1. XRD powder pattern for Cu3[Fe(CN)6]2 · xH2O fitted according to the Pm3jm structural model. Inset: The same pattern fitted using the Fm3jm space group. The pattern has a set of weak diffraction peaks (indicated by arrows) that cannot be fitted using the Fm3jm model. In the Inset the Bragg positions for the Fm3jm and Pm3jm models are also indicated.

effect related to π-back-bonding interaction has been observed for zinc hexacyanometallates(II) with Fe, Ru, and Os as the inner metal, where the dehydration temperature follows the order Os < Ru < Fe.10 The IR spectra for the studied series of copper PB analogues have also been reported.9,24 The frequency values for their main IR absorption bands have been included in Table 2. B. Crystal Structure and Vacancy Distribution. The materials under study were found to be cubic (Pm3jm space group). The XRD powder patterns of this series of compounds show several weak peaks that cannot be assigned within the Fm3jm space group (Figure 1). The structural model (Pm3jm space group) has already been observed from neutron diffraction data for PB samples recrystallized in concentrated HCl solution.22 However, no structural studies for PB analogues in that model are available. Figure 2 shows the XRD powder patterns, experimental and calculated, and the difference profiles obtained from the Rietvelt refinement, for the studied series of copper PB analogues according to the Pm3jm structural model. In Table 1 the determined cell parameters are summarized. As indicated above, the cubic Fm3jm structural model where the crystal structure of PB analogues is usually refined supposes a random vacancy distribution.2–6,19,20 When the vacancies are at least partially ordered, the structural model necessarily corresponds to a primitive space group. The cubic symmetry for the unit cell is dictated by the equivalence of the -T-N-C-M-CN-T- chains in all three directions for the three-dimensional structure, but without restrictions related to the vacancy distribution. For the studied series of copper hexacyanometallates(III), the evidence of a nonrandom vacancy distribution is favored by a good sample crystallinity (narrow diffraction peaks) where also relatively weak reflections can be detected. In the Pm3jm model there are two positions for the copper atom (1a and 3c) and also two positions for the inner metal (1b and 3d). The ligand configuration for the copper sites (1a and 3c) is determined by the local availability of vacancies (the occupation factors for 1b and 3d sites). For a random distribution of vacancies the occupation for 1b and 3d sites is 2/3 and the structure is described by the Fm3jm structural model where the two copper sites are equivalent, with a coordination environment formed by four N ends plus two water molecules, CuN4(H2O)2. For a nonrandom vacancy distribution the 1b and 3d sites have different occupation factors and this leads to also different coordination environments for the copper atom. The refined atomic positions and the calculated bond distances and angles, thermal factors, and occupation factors are available in the

Figure 2. Experimental XRD powder patterns and the calculated and difference profiles obtained from the Rietveld refinement of the crystal structure for the studied materials.

Supporting Information. This structural model has three positions for N and O atoms (6e, 6f, and 12h). The 6e sites belong to the coordination environment for the copper atom located at the 1a site; N and O atoms in 6f and 12h sites form the coordination sphere for copper atoms found at 3c sites. From the calculated occupation factors (see the Supporting Information), the coordination environments for the two types of copper atoms result: CuN2(H2O)4 (site 1a) and CuN5(H2O) (site 3c). In the Pm3jm model the unit cell contains a formula unit (Z ) 1) of the studied compound versus Z ) 1.33 for a random vacancy distribution (Fm3jm). This indicates that for a nonrandom vacancy distribution the porous framework has greater available free volume. This is a relevant feature for gas storage studies, for instance (discussed below). Figure 3 shows the coordination environments for the involved metals in the Pm3jm structure. The unit cell is represented in Figure 4. Within PB analogues the lowest thermal stability is observed for copper.9 Such behavior can be attributed to the abovediscussed vacancy distribution and to the related higher porosity for the Pm3jm crystal structure. C. Mo¨ssbauer Spectra. The assignment of the Pm3jm structural model for copper hexacyanometallates(III) is supported by Mo¨ssbauer spectroscopy. The Mo¨ssbauer spectra of

5046

J. Phys. Chem. C, Vol. 114, No. 11, 2010

Jime´nez-Gallegos et al.

Figure 5. Mo¨ssbauer spectra at room temperature for (a) Fe3[Co(CN)6]2 · yH2O and (b) Cu3-xFex[Co(CN)6]2 · yH2O. These spectra were fitted with two quadrupole splitting doublets.

TABLE 3: Mo¨ssbauer Parameters at Room Temperature for the Materials under Study composition Figure 3. Coordination environments for the metals in Cu3[M(CN)6]2 · yH2O with M ) Fe, Co, Ir. The structure contains two sites for the copper atoms, 3c and 1a. These sites have one (a) and four (b) water molecules, respectively, in the coordination environment for the copper atom.

Fe3[Co(CN)6]2 · yH2O

δa ∆ Γ (mm/s) (mm/s) (mm/s)

1.41 1.40 Cu3-xFex[Co(CN)6]2 · yH2O 1.45 1.43 Cu3[Fe(CN)6]2 · 10H2O 0.10

1.09 1.90 1.27 2.19 0.54

0.45 0.50 0.54 0.39 0.31

assignment high spin Fe(2+) high spin Fe(2+) high spin Fe(2+) high spin Fe(2+) low spin Fe(III)

a

The values of δ are reported relative to sodium nitroprusside. Fitting errors in the values of δ, ∆, and Γ remain below 0.01 mm/s.

Figure 4. Atomic packing within the cubic unit cell for Cu3[M(CN)6]2 · yH2O with M ) Fe, Co, Ir. Indicated are the vacancies of the building unit, [M(CN)6]. There are two structural positions for each of the involved metals (M and Cu).

the mixed compositions, Cu3-xFex[Co(CN)6]2 · yH2O, show two quadrupole splitting doublets of high spin Fe(2+) (Figure 5 and Table 3). That spectrum corresponds to two structural sites for iron and also for copper. The preparation of PB analogues from mixed metals leads to the formation of solid solutions within

the same framework, where the involved metals are randomly distributed at the cavity surface.9 These doublets have different quadrupole splitting values (Table 3), indicating the existence of two different coordination environments for the iron atom (sites 1a and 3c). The estimated isomer shift values are slightly different because the water molecule and the CN group have different bonding properties. The observed quadrupole splitting doublets show a certain asymmetry in their peak intensities. This was attributed to the existence of vibrational anisotropy for the iron atoms in 1a and 3c sites.25 This is a well-known effect in Mo¨ssbauer spectroscopy.26 The iron atoms are located at the surface of the cavity with mixed coordination spheres, FeN2(H2O)4 (site 1a) and FeN5(H2O) (site 3c). The vibrational anisotropy is related to the configurations of these ligands and their bonding properties. For comparison, in Figure 5 the Mo¨ssbauer spectrum for Fe3[Co(CN)6]2 · yH2O was also included. Mo¨ssbauer spectra of Fe3[M(CN)6]2 · yH2O with M ) Cr, Mn, Co, Rh, Ir are also formed by two Fe(2+) doublets.27 The reported magnetic properties and Mo¨ssbauer spectrum for copper within hexacyanoferrate(III)24 find explanation in the crystal structure obtained for this compound. The coordination of the copper atom to five CN groups at 3c sites is responsible for a pronounced charge subtraction, via π-back-donation, from the iron atom linked at the C end. This enables a strong ferromagnetic interaction between iron and copper atoms, detected as a large value for the exchange integral (J), the largest one within that series of molecular magnets.24 Regarding the Mo¨ssbauer spectrum, a strong charge subtraction at the N end

Structure Porous Copper PB Analogues

J. Phys. Chem. C, Vol. 114, No. 11, 2010 5047 for the material related to a nonrandom vacancy distribution could also be contributing to the H2 storage at high pressures but not at low pressures where the determining factor is the availability of specific interaction sites. 4. Conclusions

Figure 6. Excess H2 adsorption isotherms for T3[Co(CN)6]2 with T ) Co, Ni, Cu, Cd.

induces an increase for the electron density removal from the iron atom at the C end. This reduces the 3d electron shielding effect on the s-electron density at the iron nucleus, and a low isomer shift value is detected, the lowest one within divalent transition metal hexacyanoferrates(III).21 In Table 3 the Mo¨ssbauer parameters for the studied copper hexacyanoferrate(III) sample are reported. D. High Pressure H2 Adsorption Isotherms. From H2 adsorption in the T3[Co(CN)6]2 series of PB analogues up to 1 atm for T ) Cu the highest H2 storage capacity has been reported.2 For H2 storage in copper PB analogues relatively high adsorption heat values have also been observed which is ascribed to the H2 coordination interaction to the copper atoms at the surface of cavities related to a high electron density accumulation on the metal center (Cu).9 In PB analogues the copper atom shows a marked ability to subtract charge from the CN ligands to adopt an effective valence close to Cu(+).24 Figure 6 shows the recorded H2 adsorption isotherms up to 10 atm in representative compositions of porous PB analogues. These isotherms correspond to the excess adsorption, which is the amount of adsorbate molecules that are retained (adsorbed) by interaction with the surface. The maximum of the excess adsorption isotherm occurs at the pressure where the gas densities at the sample pore and the bulk gas are increasing at the same rate, so that a pressure increase has no effect on the amount adsorbed.28 Above that point the gas density in the sample pores saturates while the bulk gas density keeps increasing, resulting in the observed negative gain for the excess amount adsorbed. From this fact, the maximum adsorbed amount is a sensor for the available adsorption sites for the H2 molecule. For the recorded isotherms the maximum adsorption corresponds to 6.8, 4.8, 4.8, and 4.4 H2 molecules per formula unit for Cu, Ni, Co, and Cd, respectively. The H2 adsorption capacity for Cu is about 40% higher than the value observed for the remaining compositions. This agrees with the above-discussed structural difference between porous frameworks related to nonrandom and random vacancy distributions. For copper atoms in PB analogues with a crystal structure based on the Pm3jm model, the specific H2 interaction with the metal must be particularly favorable for the 3c sites. In these sites the copper atom is interacting with five CN groups and, as a consequence, with a high availability of electron density to be subtracted from their 5σ orbitals. This leads to a relatively high concentration of electron density on the copper atom, to favor the probable H2 coordination interaction to the metal (Cu). This could explain the relatively high H2 adsorption heat values observed for copper in PB analogues.2,4 The higher porosity

The high H2 storage capacity observed for copper within PB analogues finds explanation in their refined crystal structures using the Pm3jm space. The primitive space group for the structure of these compositions is related to a nonrandom vacancy distribution. This structural model has two sites for the copper atom. The local configuration of ligands around the copper atom for these two sites was estimated from the refined occupation factors. Deviations from the average coordination, T(NC)4(OH2)2, corresponding to the usually accepted Fm3jm model for PB analogues were observed. The structural refinement in the Pm3jm model is supported by the recorded Mo¨ssbauer spectra for Cu3-xFex[Co(CN)6]2 · yH2O, where two quadrupole splitting doublets were identified. These doublets have different quadrupole splitting values related to the existence of different coordination environments for the iron atom. According to the Pm3jm model, the structure for the studied porous solids has two open metal sites of quite different electronic densities and configurations of ligands for a specific interaction with the H2 molecule. In addition, their porous framework has a higher available free volume for the H2 adsorption when it is compared with the supposed cavity volume for a random vacancy distribution. The high pressure H2 adsorption isotherms also support the refined structural model. For copper the limit H2 adsorption capacity is about 40% higher than the value observed for other metals within a given series of PB analogues. Structural information for the studied materials has been deposited at ICSD Fachinformationszentrum Karlsruhe (FIZ) ([email protected]). CSD Numbers and Compounds: 419439, Cu3[Fe(CN)6]2 · 10H2O; 419440, Cu3[Co(CN)6]2 · 10H2O; 419441, Cu3[Ir(CN)6]2 · 10.5H2O. Acknowledgment. The access to Laboratorio Nacional de Luz Sı´ncrotron (LNLS) at Campinas, Brazil, is greatly recognized. Partial support from ICyTDF PIFUTP08-158 and CONACyT SEP-2007-61541 projects is acknowledged. The authors thank C. P. Krap for the H2 adsorption isotherm recording. Supporting Information Available: TG curves (the dehydration region only), tables of refined structural parameters, and EPR spectra for the studied samples. This information is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Verdaguer, M.; Bleuzen, A.; Marvaud, V.; Vaissenmann, J.; Seuleiman, M.; Despanches, C.; Scuiller, A.; Train, C.; Garde, R.; Gelly, G.; Lomenech, C.; Rosenman, I.; Veillet, P.; Carter, C.; Villain, C. F. Coord. Chem. ReV. 1999, 190-192, 1023. (2) Kaye, S. S.; Long, J. R. J. Am. Chem. Soc. 2005, 127, 6506. (3) Chapman, K. W.; Southon, P. D.; Weeks, C. L.; Kepert, C. J. Chem. Commun. 2005, 3322. (4) Hartman, M. R.; Peterson, V. K.; Liu, Y.; Kaye, S. S.; Long, J. R. Chem. Mater. 2006, 18, 3221. (5) Kaye, S. S.; Long, J. R. Catal. Today 2007, 120, 311. (6) Natesakhawat, S.; Culp, J. T.; Matranga, C.; Bockrath, B. J. Phys. Chem. C 2007, 111, 1055. (7) Kaye, S. S.; Long, J. R. Chem. Commun. 2007, 4486. (8) Reguera, L.; Balmaseda, J.; del Castillo, L. F.; Reguera, E. J. Phys. Chem. C 2008, 112, 5589. (9) Reguera, L.; Krap, C. P.; Balmaseda, J.; Reguera, E. J. Phys. Chem. C 2008, 112, 15893.

5048

J. Phys. Chem. C, Vol. 114, No. 11, 2010

´ vila, M.; Reguera, E. J. (10) Reguera, L.; Balmaseda, J.; Krap, C. P.; A Phys. Chem. C 2008, 112, 17443. ´ vila, M.; Reguera, L.; Rodrı´guez-Herna´ndez, J.; Balmaseda, J.; (11) A Reguera, E. J. Solid State Chem. 2008, 181, 2899. (12) Chen, B.; Eddaoudi, M.; Reineke, T. M.; Kampf, J. W.; O’Keeffe, M.; Yaghi, O. M. J. Am. Chem. Soc. 2000, 122, 11559. (13) Forster, P. M.; Eckert, J.; Chang, J.-S.; Park, S.-E.; Fe´rey, G.; Cheetham, A. K. J. Am. Chem. Soc. 2003, 125, 1309. (14) Schlichte, K.; Kratzke, T.; Kaskel, S. Microporous Mesoporous Mater. 2004, 73, 81. (15) Rowsell, J. L. C.; Yaghi, O. M. Angew. Chem., Int. Ed. 2005, 44, 4670. (16) Chen, B.; Ockwig, N. W.; Millward, A. R.; Contreras, D. S.; Yaghi, O. M. Angew. Chem., Int. Ed. 2005, 44, 4745. (17) Dinca, M.; Dailly, A.; Liu, Y.; Brown, C. M.; Neumann, D. A.; Long, J. R. J. Am. Chem. Soc. 2006, 128, 16876. (18) Liu, Y.; Kabbour, H.; Brown, C. M.; Neumann, D. A.; Ahn, C. C. Langmuir 2008, 24, 4772. (19) Ludi, A.; Gudel, H. U. Struct. Bonding (Berlin) 1973, 14, 1.

Jime´nez-Gallegos et al. (20) Herren, F.; Fisher, P.; Ludi, A.; Halg, W. Inorg. Chem. 1980, 19, 956. (21) Martı´nez-Garcı´a, R.; Knobel, M.; Reguera, E. J. Phys. Chem. B 2006, 110, 7296. (22) Buser, H. J.; Schwarzenbach, D.; Petter, W.; Ludi, A. Inorg. Chem. 1977, 16, 2704. (23) Rodrı´guez-Carvajal, J. The FullProf Program; Institute Leon Brillouin: Saclay, France, 2000. (24) Reguera, E.; Rodrı´guez-Herna´ndez, J.; Champi, A.; Duque, J. G.; Granado, E.; Rettori, C. Z. Phys. Chem. 2006, 220, 1609. (25) Reguera, E.; Yee-Madeira, H.; Demeshko, S.; Eckold, G.; Jime´nezGallegos, J. Z. Phys. Chem. 2009, 223, 701. (26) Chemical Applications of Mo¨ssbauer Spectroscopy; Goldanskii, V. I., Herber, R. H., Eds.; Academic Press: New York, 1968. (27) Reguera, E.; Ferna´ndez, J.; Nu´n˜ez, L.; Yee-Maderia, H. Transition Met. Chem 1999, 24, 163. (28) Zhou, L. In Adsorption: Theory, Modeling and Analyzing; Toth, J., Ed.; Marcel Dekker, Inc.: New York, 2001; pp 211-250.

JP910544J