Continuous Wave and Pulsed Electron Spin Resonance Spectroscopy

Sep 2, 2010 - (26) For comparison, the room temperature ESR spectrum of the as synthesized sample II is likewise illustrated in Figure 2c. ..... Pulse...
0 downloads 13 Views 315KB Size
Download PDF
16630

J. Phys. Chem. C 2010, 114, 16630–16639

Continuous Wave and Pulsed Electron Spin Resonance Spectroscopy of Paramagnetic Framework Cupric Ions in the Zn(II) Doped Porous Coordination Polymer Cu3-xZnx(btc)2 Bettina Jee,† Konrad Eisinger,† Farhana Gul-E-Noor,† Marko Bertmer,† Martin Hartmann,‡ Dieter Himsl,‡ and Andreas Po¨ppl*,† Fakulta¨t fu¨r Physik und Geowissenschaften, UniVersita¨t Leipzig, Linne´straβe 5, D-04103 Leipzig, Germany, and Erlangen Catalysis Resource Center (ECRC), Friedrich-Alexander UniVersita¨t Erlangen Nu¨rnberg, Egerlandstraβe 3, D-91058 Erlangen, Germany ReceiVed: June 28, 2010; ReVised Manuscript ReceiVed: August 18, 2010

In the parent metal-organic framework Cu3(btc)2 material the Cu(II) pairs in the paddle wheel building blocks of the framework give rise to an antiferromagnetic spin state with an electron spin resonance (ESR) silent S ) 0 ground state. The thermally excited S ) 1 state of the Cu(II) pairs can be observed for temperatures above 80 K by ESR spectroscopy but give rise to an exchanged narrowed resonance line preventing the exploration of any structural details in the environment of the paddle wheel units. However, magnetically diluted paramagnetic binuclear Cu-Zn clusters can be formed by substitution of Cu(II) ions by Zn(II) at low doping levels, as already known for zinc-doped copper acetate monohydrate. Indeed, ESR, hyperfine sublevel correlation spectroscopy (HYSCORE) and pulsed electron nuclear double resonance (ENDOR) verify the successful incorporation of zinc ions at cupric ion sites into the framework of the resulting Cu3-xZnx(btc)2 coordination polymer. The formation of such paramagnetic binuclear Cu-Zn paddle wheel building blocks allows the investigation of the interaction between the Cu(II) ions and various adsorbates by advanced pulsed ESR methods with high accuracy. As a first example we present the adsorption of methanol over Cu3-xZnx(btc)2, which was found to coordinate directly to the Cu(II) ions via their open axial binding site. 1. Introduction 1,2

3

Since the pioneering work by Yaghi et al., Fe´rey et al., and others, metal-organic framework (MOFs) compounds have attracted considerable attraction in the past decade. In general, these coordination polymers are composed of two basic structural building units, transition metal ion complexes or metal ion clusters and connecting organic linker molecules, such as di- or tricarboxylic acid. Characteristic features of MOF compounds are a crystalline three-dimensional open framework with large pore diameters and high specific micropore volumes, high metal ion content, and high structural diversity controlled by the use of a variety of different organic linkers and metal ion clusters.4 Cu3(btc)2(H2O)3 · xH2O (btc ) benzene-1,3,5-tricarboxylate), also known as HKUST-1,5,6 is together with MOF-57 one of the best investigated porous MOF materials and among the first commercially available representatives of this new class of organic-inorganic hybrid materials. The network of Cu3(btc)2 belongs to the cubic space group Fm3m.5 Antiferromagnetically coupled Cu(II)2 clusters6 are coordinated by carboxylate groups to form a so-called paddle wheel unit. In this paddle wheel unit, the copper clusters are arranged by four carboxylate groups to a square whereas water molecules weakly bind to the residual axial binding site of the Cu(II) ions in the hydrated Cu3(btc)2(H2O)3 · xH2O material (Figure 1a). The Cu(II) pairs are connected by the btc linker molecules to form a threedimensional porous network (Figure 1b) with interconnected cages having two different pores with approximate diameters of 0.9 and 0.7 nm as measured between the oxygen atoms of * Corresponding author. E-mail: [email protected]. † Universita¨t Leipzig. ‡ Friedrich-Alexander Universita¨t Erlangen Nu¨rnberg.

the carboxylate groups. The axial water molecules can be easily removed from the Cu(II) ions by a moderate heat treatment in vacuum to form structurally well-defined accessible Lewis acid copper sites for catalytic applications.8 Although discovered ten years ago, HKUST-1 is still a material of high interest due to its outstanding properties in gas storage9,10 and separation11 processes, its magnetism,6 and its catalytic activity in cyanosilylation reactions.8 Within this wide variety of possible applications for Cu3(btc)2 and for MOFs in general it is of particular interest to tune the physical and chemical properties of a given material in a certain way while fully maintaining structural features. This can basically be achieved by either manipulating the organic linker molecule12 or by modifying the inorganic building unit, e.g., substituting the metal species to a certain degree.13 As a rare example for the modification of the inorganic building block MOF materials, here we report the synthesis of partially Zn(II) substituted Cu3(btc)2 with the aim to form binuclear Cu-Zn paddle wheel building blocks. The verification of such an isomorphous framework substitution in porous materials is always an experimental challenge if single crystals for diffraction experiments are not available. Therefore, we work at low Zn(II) doping concentrations, which allows us to employ electron spin resonance (ESR) spectroscopic methods for that purpose. In the presence of paramagnetic ions continuous wave (cw) ESR and in particular pulsed ESR14 techniques have proven to be powerful tools to explore such isomorphous framework substitutions involving paramagnetic ions in disordered porous materials.15-17 ESR techniques further provide the opportunity to explore the interaction between paramagnetic framework ions and adsorbed molecules in a detailed unique manner.17,18 However, the antiferromagnetically coupled Cu(II) ion pairs in the paddle wheel units of the pure

10.1021/jp105955w  2010 American Chemical Society Published on Web 09/02/2010

ESR Spectroscopy of Cu3-xZnx(btc)2

J. Phys. Chem. C, Vol. 114, No. 39, 2010 16631

Figure 1. Schematic representation of the structure of Cu3(btc)2(H2O)3 showing (a) the Cu(II) paddle wheel building block and (b) the Cu3(btc)2(H2O)3 network. Only the oxygen atoms of the water molecules binding axially to the Cu(II) ions are shown in (a).

Cu3(btc)2 compound are known to result in an excited S ) 1 electron spin state and an ESR silent S ) 0 ground state.19 In principle, the excited state is sufficiently populated at temperatures above 100 K to allow for a ESR characterization. However, the system is basically magnetically nondiluted with respect to this S ) 1 state of the cupric ion pairs due to their high local concentrations and only an exchanged narrowed resonance line has been observed19 without any further spectroscopic details such as hyperfine (hf) splittings or an anisotropy of the electron Zeeman splitting tensor g. This prevents the exploration of structural details such as adsorbed molecules in the environment of the paddle wheel units. Moreover, the electron spin relaxation times of the S ) 1 state of the cupric ion pairs are much too short to allow for application of high resolution pulsed ESR techniques. Partial substitution of cupric ions by diamagnetic Zn(II) is known to create Cu/Zn paddle wheel units with spin S ) 1/2 for zinc-doped copper acetate monohydrate.20 A similar approach is presented here and cw ESR, pulsed electron nuclear double resonance (ENDOR),21 and hyperfine sublevel correlation spectroscopy (HYSCORE)22 are employed to explore the formation of paramagnetic diluted binuclear Cu/Zn paddle wheel building blocks in such Cu3-xZnx(btc)2 MOF materials at low Zn(II) doping levels. The paramagnetic ground state of the Cu-Zn pairs will allow for a precise characterization of the coordinative environment of the cupric ions by ESR methods including their interaction with adsorbates. As a first example to verify this concept, we are further studying the adsorption of methanol over Cu3-xZnx(btc)2 materials. 2. Experimental Section Synthesis. Zn(II) doped material Cu2.97Zn0.03(btc)2 (I) was synthesized using a slightly modified solvothermal method already established in the literature for Cu3(btc)2.5 In a typical setup, 3.23 mmol of Cu(NO3)2 · 3H2O (0.780 g) and 0.36 mmol of Zn(NO3)2 · 6H2O (0.108 g) were dissolved in 15 mL of a 1:1 mixture of H2O and ethanol and the solution was mixed with 2.0 mmol of trimesic acid (0.420 g) in a Teflon vessel and placed in an autoclave. All chemicals were used as purchased from Aldrich in p.a. grade without further purification. The autoclave was heated in an oven at 398 K for 12 h. The turquoise-blue product was filtered, washed with H2O and ethanol, and airdried (yield: 0.71 g, as synthesized). The parent Cu3(btc)2 (II) was synthesized with a similar procedure. II was extracted with

ethanol for purification.2 Thereafter the samples were dried under vacuum at 393 K for 24 h to obtain dehydrated materials and then kept under dry nitrogen until further use. Methanol adsorbed samples were prepared from dehydrated samples by adsorbing CH3OH at its vapor pressure at room temperature for 1 h. Characterization. The chemical analysis has been performed by atomic absorption spectroscopy (AAS) using a Perkin-Elmer 3300 AAS with external calibration. Powder X-ray diffraction (XRD) patterns were obtained in Debye-Scherrer mode on a STADI-P (STOE) equipped with a linear PSD and a Ge(111) monochromator using Cu KR1 radiation (λ ) 154.060 pm). The samples were placed in 0.5 mm capillaries (No. 14, HILGENBERG). Nitrogen adsorption isotherms were recorded at 77 K on a Micromeritics Gemini II 2370 sorption analyzer. Thermal gravimetric analysis was carried out on a STA 410 (Netzsch) with a heating rate of 10 K/min in a helium flow (75 mL/min) connected to a mass spectrometer. 1H magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra were recorded using a Bruker Avance spectrometer at a magnetic field of 9.4 T. A 4 mm MAS probe was used at a spinning speed of 8 kHz. The samples were placed in sealed glass tubes. The 90° pulse length was 2 µs, and the recycle delay, 30 ms. Spectroscopic Measurements. For ESR measurements the material was placed in quartz glass tubes and heated under vacuum from room temperature to the activation temperature of 393 K over a period of 12 h. The final vacuum was 0.01 Pa. Subsequently, the sample tubes with the dehydrated material I were cooled to room temperature and sealed without further air contact. All X-band cw ESR, two-pulse field-swept electron spin echo (FS-ESE),14 pulsed ENDOR,21 and HYSCORE22 experiments were recorded on a Bruker ELEXYS E580 spectrometer. Pulsed ENDOR experiments at Q-band frequencies were performed using a spectrometer with a home-built microwave (mw) pulse unit23 and a commercial magnet system of a BRUKER EMX cw Q-band instrument. Two-pulse FS ESE and HYSCORE spectra were recorded at X-band at 6 K using nonselective mw pulses of tπ/2 ) 16 and tπ ) 32 ns. To avoid suppression effects in the proton HYSCORE experiments, the spectra were recorded with different pulse delays τ ) 104, 136, and 164 ns at 340.8 mT. A 170 × 170 data matrix was sampled and two-dimensional (2D) Fourier transformed (FT) magnitude spectra were displayed. The simulations of the HYSCORE spectra were

16632

J. Phys. Chem. C, Vol. 114, No. 39, 2010

Jee et al.

calculated in the time domain by exact diagonalization of the spin Hamiltonian. For further information concerning the simulation procedure we refer to an earlier paper.24 For pulsed ENDOR experiments the Davies ENDOR sequence17 was used with mw pulse lengths of tπ/2 ) 100 ns and tπ ) 200 ns with a pulse delay of τ ) 1000 ns between the second and third mw pulse for X-band and tπ/2 ) 90 ns, tπ ) 130 ns, and τ ) 300 ns for Q-band experiments. The length of the radiofrequency (rf) pulses were trf ) 10 µs (X-band) and trf ) 20 µs (Q-band). Spectral simulations of the pulsed ENDOR spectra as well as of the Cu(II) cw ESR powder patterns were done using the EasySpin ESR simulation package.25 3. Results Characterization. The nCu/nZn ratio of sample I was determined as 90:1 by AAS spectroscopy corresponding to the formula Cu2.97Zn0.03(btc)2 (I) for the Zn(II) containing material. As no suitable crystals of I and II for single crystal X-ray analysis were obtained, powder XRD patterns of as synthesized materials were recorded. The XRD data confirmed that the framework structure of I is similar to that of the cubic Cu3(btc)2 parent material II (Fm3jm, Figure S1, Supporting Information). Before the N2 adsorption experiment at 77 K the as synthesized material I was dehydrated at 373 K for 24 h. Adsorption analysis of I (Figure S2, Supporting Information) revealed a specific surface area of 1320 m2 g-1 (BET) and 1678 m2 g-1 (Langmuir) and a total pore volume of 0.6033 cm3 g-1. For further characterization, dehydrated materials of I were rehydrated and thermogravimetric analysis (TG/DTA) of the rehydrated material I showed a total mass loss of 33.1%, which is ascribed to the release of adsorbed water molecules up to a temperature of 573 K (Figure S3, Supporting Information). The thermal stability of I exceeds the values reported in the literature for the pure copper compound5 since decomposition does not occur up to 627 K. Despite the presence of magnetic copper ions, the 1H MAS NMR spectra of dehydrated samples I and II show highly resolved signals (Figure S4, Supporting Information). Both samples show the same signal at 8.3 ppm for the aromatic protons of the framework; therefore, we assume that the small amount of Zn(II) in sample I has no significant effect on the corresponding NMR spectrum. Cw ESR Spectroscopy. X-band cw ESR measurements of the as synthesized and dehydrated samples of I at 298 K (Figure 2a,b) showed the broad isotropic signal of the excited S ) 1 state of the antiferromagnetically coupled Cu(II) pairs of the paddle wheel units known from II.26 For comparison, the room temperature ESR spectrum of the as synthesized sample II is likewise illustrated in Figure 2c. The g value of this isotropic signal of about g ) 2.170 found for I is in the range of g ) 2.164 reported for II.19 The line width of the signal observed for I with ∆Bpp ) 74 mT appears to be somewhat lower than found for II (85 mT). The signal of the antiferromagnetically coupled Cu(II) pairs disappears at lower temperatures (Figure S5, Supporting Information) and the typical anisotropic ESR powder pattern of a single Cu(II) ion species with S ) 1/2 and a well resolved 63,65Cu (nuclear spin ICu ) 3/2) hyperfine (hf) splitting into four lines develops below 80 K for the as synthesized and dehydrated samples of I. Thereby, the line width of the Cu(II) ESR spectrum decreases with decreasing temperature (Figure S5, Supporting Information). Figure 3a,b illustrates the spectra of the as synthesized and dehydrated samples of I at T ) 6 K. In both spectra the slightly different hf splitting due to the two copper isotopes, 63Cu and

Figure 2. X-band cw ESR spectra at T ) 298 K of (a) I as synthesized, (b) I dehydrated, and (c) II as synthesized.

Figure 3. Experimental (solid lines) and simulated (dashed lines) X-band cw ESR spectra at T ) 6 K of (a) I as synthesized, (b) I dehydrated, and (c) II as synthesized. For simulation parameters see Table 1. 65 Cu, can easily be distinguished at the low field gzz edge singularity of the cupric ion ESR powder patterns. In particular, the spectrum of the dehydrated material in Figure 3b exhibits unusual narrow line widths for Cu(II) ESR signals of porous materials, indicating that strain effects in the Zeeman and Cu hf splitting (g and A strain) are almost absent. For the as synthesized sample (Figure 3a) the line width of the four hf signals at the gzz edge singularity increases with rising field due to some correlated g and A strain effects.27 Spectral simulations of the resolved cw Cu(II) ESR powder pattern of the as synthesized and dehydrated samples I at 6 K provided axially symmetric Zeeman and 63Cu hf interaction tensors g and ACu. The principal values of the tensors g and ACu are summarized in Table 1. Correlated g and A strain effects have not been included in the spectral simulations. Furthermore, a comparison of the experimental and simulated spectrum of the dehydrated and methanol adsorbed samples shows the presence of an additional very broad signal that is superimposed on the nicely

ESR Spectroscopy of Cu3-xZnx(btc)2

J. Phys. Chem. C, Vol. 114, No. 39, 2010 16633

TABLE 1: Spin Hamiltonian Parameters at 6 K of Cu(II) Ions in Cu2.97Zn0.03(btc)2 (I) and Cu3(btc)2 (II)a sample

gxx,yy

gzz

Cu Axx,yy (cm-1)

Cu Azz (cm-1)

Idehydr Ias synth I · nCH3OH IIas synth

2.046 2.058 2.060 2.0874

2.279 2.337 2.336 2.381

0.0032 0.0015 0.0015 not resolved

0.0190 0.0158 0.0158 0.0152

a

Cu Errors: ∆gii ) (0.002, ∆Axx,yy ) (0.0004 cm-1.

Figure 4. Experimental (solid lines) and simulated (dashed lines) X-band cw ESR spectra at T ) 6 K of (a) I dehydrated, (b) I after adsorption of CH3OH, and (c) I after reactivation at 373 K in vacuo. For simulation parameters see Table 1.

resolved Cu(II) ESR powder pattern (Figure 4). We tentatively assign this broad spectral feature to agglomerated cupric ion species. The as synthesized sample II displays likewise the characteristic cw ESR spectrum of a cupric ion species having S ) 1/2 (Figure 3c) with resolved Cu hf interaction at T ) 6 K, but with low intensity and distinctly different g and ACu tensors (Table 1). However, the line widths and the related strain effects appear to be considerably larger in comparison with the Zn substituted sample I. Therefore, we assign this weak signal to a small amount of octahedrally coordinated extra-framework Cu2+ ions as a side product from synthesis. Adsorption of methanol at 298 K on dehydrated samples I gives rise to the low temperature cw ESR spectrum shown in Figure 4b. Again, the line widths of the Cu(II) ESR signals are sufficiently small to resolve the hf splitting from both copper 63 Cu and 65Cu isotopes at the gzz spectral position; however, the increase of the line widths of the four Cu(II) hf signals with rising field indicates the presence of some correlated g and A strain, as already observed for the as synthesized material of I. The spectrum of the dehydrated material prior adsorption is presented for comparison in Figure 4a. We have to note that at room temperature sample I with adsorbed methanol shows likewise the broad isotropic signal at g ) 2.170 of the S ) 1 state of the antiferromagnetically coupled Cu(II) pairs from the paddle wheel units (Figure S6, Supporting Information). The Cu(II) spin Hamiltonian parameters of the methanol adsorbed sample as derived from spectral simulations are presented in Table 1. The obtained principal values of the tensors g and ACu of the cupric ions differ significantly from dehydrated materials, indicating a change in the Cu(II) coordination geometry upon adsorption of CH3OH. Reactivating of the sample at 373 K

results in the spectrum at 6 K illustrated in Figure 4c. The spectrum is almost identical to that of the dehydrated material I prior methanol adsorption and verifies that the adsorption/ desorption of methanol is reversible. FS ESE Spectroscopy. Two pulse FS ESE spectra of the Zn substituted materials Cu2.97Zn0.03(btc)2 (I) were taken at X-band frequency at 6 K for comparison with the cw ESR spectra. The spectra show also nicely resolved Cu(II) powder patterns (Figure S7, Supporting Information) that can be described by the same sets of spin Hamiltonian parameters as the corresponding cw ESR spectra in Figures 3 and 4. Therefore, we may assume that the same cupric ion species are being investigated by cw and pulsed ESR methods. It is worthwhile to note that the additional broad spectral feature in the cw ESR spectrum of the methanol adsorbed samples I could not be observed in the FS ESE spectrum. This supports its assignment to Cu(II) agglomerates or clusters, as these are expected to have short electron spin relaxation times preventing their detection by ESE experiments. Pulsed ENDOR Spectroscopy. The spectral resolution of the cw ESR and FS ESE experiments are in general not sufficient to resolve weak hf interactions between the Cu(II) ions and proton nuclei in their close environment. Therefore, orientation-selective14 pulsed ENDOR spectroscopy was employed for a dehydrated sample of I to probe the coordination environment and the incorporation side of the Cu(II) ions in the Zn substituted material Cu2.97Zn0.03(btc)2. Pulsed ENDOR experiments allow the measurement of such weak ligand hf couplings, which are usually hidden within the line width of the ESR powder pattern. If the ESR spectrum is anisotropic, both the principal values and the orientation of the principal axes of the proton hf coupling tensor AH with respect to the complex coordinate frame, given, e.g., in the case of Cu(II) complexes by the tensors g and ACu, can be determined by orientation-selective ENDOR spectroscopy. In this approach, pulsed ENDOR spectra are recorded at several observer positions across the Cu(II) ESR powder pattern and in that way the 1 H hf couplings are probed along various directions of the external magnetic field with respect to the g tensor frame. Figure 5 illustrates a set of orientation-selective X-band Davies ENDOR spectra of dehydrated material I. The observer field positions where the ENDOR measurements were taken are indicated by arrows in the corresponding FS ESE spectrum (Figure S7b, Supporting Information). The spectra indicate that only protons with weak hf couplings not exceeding 2.5 MHz are interacting with the cupric ions in the samples of I. A close inspection of the spectra reveals that 1H (IH ) 1/2) ENDOR signals from three distinct protons can be distinguished. Orientation-selective Davies ENDOR spectra recorded at Q-band frequency indicate likewise the presence of two of the three distinct protons in the environment of the Cu(II) ions (Figure S8, Supporting Information). A typical decomposition of the experimental X-band pulsed ENDOR spectra into the three subspectra of these protons H1, H2, and H3 is shown for the spectrum recorded at 342.1 mT at the gxx,yy spectral region of the Cu(II) ESR powder pattern (Figure 6). The most striking feature in the ENDOR spectra is the intense doublet of proton H1 with a splitting of 1.22 MHz observed particularly well for irradiation close to the gzz and gxx,yy spectral positions. It indicates that the hf coupling tensor AH1 of proton H1 is almost coaxial to the g tensor; the splitting of 1.22 MHz corresponds then approximately to the principal value AH1 xx,yy. The pronounced shoulders with a splitting of about 2.4 MHz in the ENDOR spectrum taken at the gxx,yy spectral position (Figure 6) define

16634

J. Phys. Chem. C, Vol. 114, No. 39, 2010

Figure 5. Experimental (solid lines) and simulated (dashed lines) orientation-selective X-band 1H Davies ENDOR spectra of Cu(II) species in dehydrated sample I at 6 K at (a) 278.0 mT, (b) 292.2 mT, (c) 342.1 mT, and (d) 346.1 mT. The simulated traces show the sum of the computed spectra of the three protons H1, H2, and H3. For simulation parameters see Table 2.

Figure 6. Experimental (a) and sum of simulated spectra (b) of protons H1 (c), H2 (d), and H3 (e) at 342.1 mT. For simulation parameters see Table 2.

the AH1 zz principal value. The ENDOR doublet of proton H1 with 1.22 MHz splitting provides a characteristic spectral feature of the nuclear environment of the cupric ions in Cu2.97Zn0.03(btc)2 for further comparisons of the Cu(II) coordination site in other samples. The complete hf coupling tensors AH1,H2,H3 of all three interacting protons have been obtained from spectral simulations of the orientation-selective Davies ENDOR spectra in Figure 5. Table 2 summarizes the principal values of the tensors AH1,H2,H3, the Euler angles β defining the angle between the z principal axes of the tensors g and AH1,H2,H3, and the isotropic H and T⊥H computed from and dipolar hf coupling constants Aiso H1,H2,H3 14 . The determined the principal values of the tensors A tensors are axially symmetric within the given errors of the H and T⊥H indicate that analysis, and the isotropic values of Aiso only distant protons located several bonds away from the Cu(II) ions interact with the metal ions. We have to note that in the central part of the proton ENDOR spectra at the 1H Larmor

Jee et al. frequency νH even more distant protons with smaller hf interactions than those of H1, H2, or H3 contribute to the ENDOR spectrum but have not been included in the simulation procedure. Furthermore, our spectral simulations do not take into account effects caused by the finite excitation width of the mw pulses. Both simplifications in the computations explain the differences in the intensities of the simulated and experimental spectra in particular for signals close to νH, whereas the line positions are satisfactorily reproduced by the simulations of the pulsed ENDOR powder spectra. Pulsed ENDOR measurements have also been performed for the Cu2.97Zn0.03(btc)2 sample (I) after exposure to methanol. Figure 7 displays the Davies ENDOR spectra recorded at the gzz and gxx,yy spectral positions. Again the characteristic doublet of proton H1 with its splitting of approximately 1.22 MHz is observed for both spectra, indicating that the environment of the cupric ions is partially conserved upon methanol adsorption. Simulated 1H ENDOR spectra using the parameters of protons H1, H2, and H3 as deduced from dehydrated sample I in Table 2 are shown for comparison. But obviously the spectra show additional 1H ENDOR signals with a huge splitting of about 7 MHz (proton H4) only observed for measurement at the gzz position and smaller splitting of 3 MHz (proton H5) and 0.75 MHz (proton H6) that are not taken into account by the simulations. We assign these new signals to pronounced anisotropic 1H hf interactions of the Cu(II) ions with protons from adsorbed CH3OH molecules. HYSCORE Spectroscopy. As pulsed ESR methods based on electron spin echo envelope modulation techniques are particularly well suited for the study of highly anisotropic proton hf couplings,14 we employed orientation-selective HYSCORE spectroscopy at X-band frequency to elucidate the interaction of the Cu(II) ions with methanol for the Cu2.97Zn0.03(btc)2 sample (I). For comparison, HYSCORE spectra were also recorded for the dehydrated material of I prior methanol adsorption. Selected spectra for both samples taken at the gzz and gxx,yy spectral positions are illustrated together with simulated spectra in Figure 8 and clearly demonstrate the effect of the methanol adsorption on the local Cu(II) environment. Here we present spectra that were sampled with pulse delays τ selected in such a way as to enhance modulations from protons with large hf interactions (for HYSCORE spectra recorded with other pulse delays τ see Figure S9, Supporting Information). Whereas cross peak ridges from 13C nuclei (IC ) 1/2) are present in all spectra at the corresponding Larmor frequency, νC ) 3.0-3.6 MHz depending on the observer position, intense 1H cross peak ridges from protons (Larmor frequency νH ) 11.9-14.5 MHz) with substantial anisotropic hf couplings are only detected in the spectra recorded after exposure to methanol (Figures 8e,f) in accordance with the ENDOR results. The spectrum of the dehydrated sample I (Figures 8b) taken at 340.8 mT (gxx,yy position) displays only two narrow proton cross peaks located close to Larmor frequency at (15.3, 13.9) and (13.9, 15.3) MHz. Their maximum frequency spread corresponds to an hf coupling of about 2.4 MHz, and we assign them to proton H1 whose hf tensor AH1 has been determined already by ENDOR spectroscopy (Table 2). H1 provides only faint cross peaks at (12.5, 11.4) and (11.4, 12.5) MHz in the spectrum measured at 280.0 mT (gzz position). Spectral simulations of the HYSCORE correlation features of H1 using the parameters derived from the ENDOR experiments support this assignment (Figures 8c,d). The correlation features of the other protons H2 and H3 are suppressed by the chosen pulse delays τ in the spectra illustrated in Figure 8a,b but give rise to intense

ESR Spectroscopy of Cu3-xZnx(btc)2

J. Phys. Chem. C, Vol. 114, No. 39, 2010 16635

TABLE 2: 1H and 13C hf Coupling Parameters of Cu(II) Ions and Cu · · · H Distances Derived from Presented ENDOR/ HYSCORE and Published XRD Data in Dehydrated Cu2.97Zn0.03(btc)2 (I) and in I after CH3OH Adsorptiona sample I I I I I I

dehydr dehydr dehydr dehydr nCH3OH nCH3OH

nuclei

H,C Axx,yy (MHz)

H,C Azz (MHz)

β (deg)

H,C Aiso (MHz)

TH,C ⊥ (MHz)

rCu · · · H (Å) ENDOR

rCu · · · H (Å) XRD

H1 H2 H3 C1 H4 C1

-1.30 -0.50 -0.17 1.395 -3.30 1.05

2.36 0.9 0.34 4.64 7.38 3.65

70 62 80 70 17 70

-0.08 -0.03 0 2.475 0.26 1.92

1.22 0.47 0.17 1.08 3.56 0.87

4.03 5.55 7.38

4.30 5.40 7.38

2.82

H a Errors: ∆AiiH ) (0.02 MHz, ∆β ) (10°, ∆Aiso ) (0.02 MHz, ∆T⊥H ) (0.02 MHz, ∆rCu · · · H ) (0.05 Å for H1, H2, ∆AiiH ) (0.04 MHz, H H ∆β ) (30°, ∆Aiso ) (0.04 MHz, ∆T⊥H ) (0.04 MHz, ∆rCu · · · H ) (0.6 Å for H3, and ∆AiiH ) (0.05 MHz, ∆β ) (10°, ∆Aiso ) (0.05 MHz, C ∆T⊥H ) (0.05 MHz, ∆rCu · · · H ) (0.05 Å for H4, ∆AiiC ) (0.05 MHz, ∆β ) (20°, ∆Aiso ) (0.05 MHz, ∆TC⊥ ) (0.05 MHz for C1.

Figure 7. Experimental (solid lines) and simulated (dashed lines) orientation-selective X-band 1H Davies ENDOR spectra of Cu(II) species in sample I after adsorption of CH3OH at 6 K. The spectra were recorded at (a) 280.0 mT and (b) 338.9 mT corresponding to the gzz and gxx,yy spectral positions of the Cu(II) ESR powder pattern. The simulated spectra show the sum of the computed spectra for the four protons H1, H2, H3, and H4. For simulation parameters see Table 2.

correlation peaks at the 1H Larmor frequency in the spectra recorded with shorter τ values at both field positions, 340.8 and 280.0 mT (Figure S9, Supporting Information). The HYSCORE spectrum of sample I after adsorption of CH3OH recorded at 278.0 mT corresponding to the gzz spectral position (Figure 8e) shows two pairs of pronounced cross peak ridges, a very intense pair at (9.5, 14.7) and (14.7, 9.5) MHz and a second somewhat less intense pair at (10.8, 13.1) and (13.1, 10.8) MHz. We assign the former correlation feature to proton H4 and the latter to H5 already observed in the pulsed ENDOR spectrum at the gzz position (Figure 7a). The HYSCORE spectrum in Figure 8f taken at 338.9 mT (gxx,yy position) is dominated by the intense cross peaks of H3 at (13.3, 15.8) and (15.8, 13.3) MHz. The smaller splitting of the cross peaks for the gxx,yy position indicates a smaller hf coupling of proton H4 along the gxx,yy plane and consequently an orientation of its hf interaction tensor AH4 in such a way that its z principal H4 axis belonging to the largest principal value Azz and the z axis of the Cu(II) g tensor form an angle of less than 45°. Indeed, simulations of the HYSCORE spectra of proton H3 (Figure 8g,h) support this interpretation and provide angle β ) 17° between the two z axes of the tensors AH4 and g. To account for the width of the cross peak ridges of proton H4 in the HYSCORE

spectrum measured at the gzz position (Figure 8e), a Gaussian distribution of the angle β with a distribution width ∆β12 ) 14° was assumed in the simulations. The principal values of AH4 (Table 2) indicate the substantial anisotropic hf coupling of H4 whereas the isotropic hf interaction is small. Further HYSCORE spectra of sample I after adsorption of CH3OH recorded at the intermediate field positions 288.5 and 321.0 mT of the Cu(II) ESR powder pattern are presented in Figure S10 in the Supporting Information together with the corresponding simulations for proton H4. These two HYSCORE spectra at the intermediate field positions display also weak cross peaks of proton H5 (Figure S10c,e, Supporting Information), whereas they are not clearly resolved in the spectrum taken at 338.9 mT (Figure 8f). However, the appearance of weak shoulders at both the side of the cross peak ridges of proton H4 indicate a superposition with additional less intense cross peaks at (12.4, 16.7) and (16.7, 12.4) MHz and at (13.9, 15.2) and (15.2, 13.9) MHz from two other protons. We tentatively assign the former pair of cross peaks to H5. The positions of the inner shoulders at (13.9, 15.2) and (15.2, 13.9) MHz suggest that they belong to the proton H3 as they are comparable with its cross peak positions observed in sample I before exposure to methanol (Figure 8b). Because of the poorly resolved cross peak ridges of proton H5, only a rough estimate for its hfc parameters, AHxx,yy H ) -2.0 ( 0.2 MHz, Azz ) 3.9 ( 0.2 MHz, β ) 40 ( 20°, could be deduced from spectral simulations of the HYSCORE spectra (Figure S10, Supporting Information). We further attempted to derive a rough estimate for the 13C (IC ) 1/2) hf couplings of the cupric ions in both samples from an analysis of the 13C cross peak ridges observed in their HYSCORE spectra. Here, the maximum curvature of the cross peaks appears in the spectra measured at the gxx,yy position. This indicates that the angle β between the z axes of the tensors AC and g is larger than 45°. The angle β and the dipolar hf coupling parameter TC⊥ can be estimated from their maximum shift ∆νsmax along the ν1 ) -ν2 frequency axis at the 13C Larmor frequency towards higher frequency in the HYSCORE spectra.28 The precise hf coupling parameters of the observed 13C nucleus C1 for the two samples of I before and after adsorption of methanol are determined by spectral simulations (Figures 8b,c,g,h) and are summarized in Table 2. Finally, we note that only relative signs of the principal values of the hf coupling tensors in Tables 1 and 2 were obtained by the experiments but their absolute signs could not be determined. 4. Discussions Cu(II) Substitution by Zn(II) in Cu2.97Zn0.03(btc)2. The results of the structural characterization techniques XRD, N2 adsorption, and 1H NMR show that the structure of the zinc doped sample I, Cu2.97Zn0.03(btc)2, corresponds basically to that

16636

J. Phys. Chem. C, Vol. 114, No. 39, 2010

Jee et al.

Figure 8. X-band HYSCORE spectra of Cu(II) species in the dehydrated sample I and after adsorption of CH3OH at 6 K: experimental spectra of dehydrated sample I recorded (a) at 280.0 mT (gzz position) with τ ) 166 ns and (b) at 340.8 mT (gxx,yy position) with τ ) 134 ns; simulated spectra of dehydrated sample I (c) at 280.0 mT with τ ) 166 ns and (d) at 340.8 mT with τ ) 134 ns; experimental spectra of sample I after adsorption of CH3OH recorded (e) at 278.0 mT (gzz position) with τ ) 166 ns and (f) at 338.9 mT (gxx,yy position) with τ ) 134 ns; simulated spectra of CH3OH adsorbed sample I (g) at 278.0 mT with τ ) 166 ns and (h) at 338.9 mT with τ ) 134 ns. For the 1H and 13C simulation parameters see Table 2.

ESR Spectroscopy of Cu3-xZnx(btc)2 of the parent Cu3(btc)2 material, sample II. Likewise, the thermal stability of the Cu2.97Zn0.03(btc)2 framework was found to be comparable to that of Cu3(btc)2, as evidenced by the TG/DTA measurements. Therefore, it is justified to claim that doping of HKUST-1 by a small amount of Zn(II) conserves the framework structure of the MOF material and its stability. This result is in accordance with the room temperature cw ESR experiments. They provide clear evidence for the presence of antiferromagnetically coupled Cu(II) pairs in the paddle wheel units as the major magnetic species in Cu2.97Zn0.03(btc)2, as indicated by the characteristic intense ESR signal of their excited S ) 1 state. The isotropic nature of this ESR signal has been interpreted in terms of a spin exchange between the S ) 1 states of neighbored Cu(II) pairs in the paddle wheel units across the btc linker molecules for the parent Cu3(btc)2 materials19 and an almost isotropic ESR signal is likewise observed for Cu2.97Zn0.03(btc)2. It is natural to assume that the low dilution of the magnetic MOF framework by diamagnetic Zn(II) ions used here will not be sufficient to interrupt these spin exchange paths. However, we may speculate that the zinc incorporation might lead to minor modulations of the spin exchange processes and in that way might account for small changes in the line width and line shape of the exchange narrowed signal of the S ) 1 state of the cupric ion pairs. Whereas the parent and Zn substituted MOF materials behave similarly at room temperature, their ESR characteristics are completely different at low temperatures. Both as synthesized samples exhibit ESR signals of isolated paramagnetic S ) 1/2 Cu(II) species but with very distinct spin Hamiltonian parameters, indicating different coordination environments of the cupric ions. The Cu(II) spin Hamiltonian parameters usually provide some rough guide to the overall coordination geometry of the metal ion based on extensive studies of known cupric Cu increases from approximately ion compounds. In general, Azz 0.007 cm-1 for tetrahedral symmetry, through elongated distorted octahedral and square-pyramidal symmetry, to approximately 0.018 cm-1 for square-planar symmetry, whereas gzz decreases from 2.516 to 2.245 for this sequence of coordination symmetries.29 The as synthesized sample II, Cu3(btc)2(H2O)3 · xH2O, shows a weak ESR signal of Cu(II) ions, those spin Hamiltonian parameters (Table 1) are typical for an elongated octahedral coordination. They have recently been assigned to extraframework cupric ion species that are accommodated as [Cu(H2O)6]2+ complexes in the pore system of Cu3(btc)2(H2O)3 · xH2O.26 The Cu(II) parameters of the Zn containing as synthesized sample I, Cu2.97Zn0.03(btc)2 (H2O)3 · xH2O, suggest a square pyramidal geometry. Such a coordination environment is actually expected for a cupric ion in the paddle wheel unit of the as synthesized material where four oxygen atoms from the carboxylate groups forming the paddle wheels and an axial water or ethanol ligand is coordinating to the metal ion.5 Similar parameters but with a slightly orthorhombic g tensor have been reported for ZnCuAc4 · 2H2O,20 which can be regarded as the molecular analog of I, further supporting the assignment of the observed S ) 1/2 EPR Cu(II) signal to mixed Cu/Zn paddle wheel units. The axial ligand can easily be removed by dehydration in vacuum.8 Indeed, the decrease of the principal Cu observed for the dehydrated value gzz and the increase of Azz sample of I (Table 1) indicates a transition from the square pyramidal toward a square planar coordination geometry when the axial ligand is removed from the cupric ion. Therefore, the appearance of a S ) 1/2 spin state of the Cu(II) ions at low

J. Phys. Chem. C, Vol. 114, No. 39, 2010 16637 temperatures for the zinc substituted material I together with the coordination geometries of the cupric ions found by cw ESR is in accordance with their assignment to mixed Cu/Zn paddle wheel units where magnetic Cu(II) ions have been substituted by diamagnetic Zn(II) ions. Furthermore, g and A strain effects are drastically reduced for the ESR signals of the monomeric Cu(II) species in sample I in comparison with the signals of the extra-framework Cu(II) ions in the parent sample II. This suggest a considerably more uniform local environment for these cupric ion species in I in accordance with the assumption of their incorporation at well-defined framework sites in mixed Cu/Zn paddle wheels. Further evidence for the formation of binuclear Cu/Zn paddle wheels in sample I comes from the pulsed ENDOR measurements that probe the nuclear spin environment around the Cu(II) ions and allow a direct comparison with atomic positions of selected framework constituents as obtained from crystallographic data.5 Unfortunately, no ENDOR responses from neighboring Zn(II) ions in the paddle wheel unit could be detected, presumably due the low natural abundance of 4.11% of the only magnetic zinc nuclei 67Zn (IZn ) 5/2). However, three axially symmetric hf coupling tensors with surrounding protons, labeled H1, H2, H3 in Table 2 could be determined from the orientation-selective ENDOR spectra of the dehydrated sample Cu2.97Zn0.03(btc)2 (I). Their isotropic hf interaction parameters H are small, and therefore, the unpaired electron spin density Aiso at the protons is negligible. Consequently, it is justified to evaluate the distances rCu · · · H between the Cu(II) ion and the protons from the dipolar hf coupling parameter T⊥H,C using a simple point-dipole approximation.14 In Table 2 the distances rCu · · · H computed from the T⊥H,C values are compared with the distances between a cupric ion in the paddle wheel and its closest proton positions taken from the XRD data.5 These are three protons from the neighboring btc ligands labeled by H1, H2, and H3 in the schematic structure of the Cu/Zn paddle wheel unit displayed in Figure 9a. Equivalent protons from the four coordinating btc linkers cannot be distinguished in the orientation-selective ENDOR spectra because of the C4h symmetry of the Cu/Zn paddle wheel unit. The Cu · · · H distances of the protons H1, H2, and H3 derived from the ENDOR results agree very well within the experimental accuracy with the distances between the Cu(II) ions in the paddle wheels and the three proton positions of the next neighbored btc linkers obtained from the crystallographic data. Furthermore, the z axes of the tensors AH1,H2,H3 must point along the vector joining the cupric ion and the respective proton within the limit of the point-dipole approximation. Here the orientation of the z axis of the tensors AH1,H2,H3 is defined by the Euler angle β (Table 2) between the z axes of the tensors AH1,H2,H3 and the Cu(II) g or ACu tensor. It is natural to assume that the z axes of the tensors g and ACu are directed along the C4 axis of the paddle wheel unit defined by a vector pointing from the Zn to the Cu ion. From the crystallographic data5 we can compute the angles 81°, 52°, and 80° between the C4 axis and the vectors pointing from the cupric ion toward the positions of protons H1, H2, and H3, respectively. These angles are in good agreement with the Euler angles β of the proton hf coupling tensors AH1,H2,H3 (Table 2). We may conclude that the results of the 1H pulsed ENDOR experiments confirm likewise the formation of mixed binuclear Cu/Zn paddle wheels in Cu2.97Zn0.03(btc)2 and further show that their structure including the next neighbored btc ligands has not significantly been altered by the Zn substitution. Methanol Adsorption in Cu2.97Zn0.03(btc)2. Upon adsorption of CH3OH on the dehydrated Cu2.97Zn0.03(btc)2 sample (I) the

16638

J. Phys. Chem. C, Vol. 114, No. 39, 2010

Jee et al.

Figure 9. Schematic representation of the structure of (a) a binuclear Cu/Zn paddle wheel unit and (b) of the CH3OH adsorption complex in Cu2.97Zn0.03(btc)2 (I). The protons H1, H2, H3, and H4 observed in the ENDOR and HYSCORE spectra are indicated in the drawings.

tensors g and ACu of the Cu(II) ions change significantly, indicating the expected change in the coordination geometry of the copper center. The increase of the principal value gzz and the decrease of Azz indicate a transition from a square planar back toward a square pyramidal geometry when CH3OH donor molecules are coordinating to the free axial binding sites of the Cu(II) ions in the Cu/Zn paddle wheel units. The somewhat larger g and A strain effects observed for the sample exposed to CH3OH in comparison with the dehydrated material indicates that the coordination of the adsorbate leads to a slightly enhanced structural variation at the Cu(II) center presumably due to small differences in the orientation of the adsorbed methanol molecules. The adsorption process is reversible as the original Cu(II) ESR spectrum of the Cu/Zn paddle wheels in the dehydrated material can be restored after evacuation the samples at elevated temperatures. Pulsed ENDOR and HYSCORE spectroscopy confirm the proposed coordination of the methanol at the free axial Cu(II) binding site and provide further information about the geometry of the adsorption complex. The spectra of the Cu2.97Zn0.03(btc)2 sample (I) after adsorption of CH3OH reveal the presence of an additional proton H4 with substantial anisotropic hf coupling H . Small isotropic proton ligand hf and small value of Aiso interactions of less than 1 MHz are indicative of an axial ligand30 as the overlap between the ligand atomic orbitals and the spin bearing 3dxy Cu(II) atomic orbital29 is negligible. The T⊥H value of H4 translates into a Cu · · · H4 distance of rCu · · · H4 ) 2.82 Å (Table 2). Therefore, we assign H4 to the hydroxyl proton of an axially coordinating CH3OH that binds via its oxygen to the cupric ion of the Cu/Zn paddle wheel unit (Figure 9b). The found Euler angle β ) 17° of the tensor AH4 proves likewise that the methanol coordinates to the free axial binding site of the metal ion. We have to note that comparable 1H hf coupling parameters have been obtained for water molecules axially coordinating to Cu(II) centers in zinc-doped copper acetate monohydrate.30 The other proton H5 observed in the HYSCORE spectra with smaller hf coupling values of AHxx,yy ) -2.0 MHz H ) 3.9 MHz must then be assigned to one of the three and Azz methyl group protons of the adsorbed CH3OH molecule. Besides the 1H proton hf coupling tensors the HYSCORE experiments provided also information about the hf interaction of the Cu(II) ions in the Cu/Zn paddle wheels with 13C nuclei. Although a 13C hf coupling tensor could be determined for both dehydrated and methanol adsorbed Cu2.97Zn0.03(btc)2 (I) samples (Table 2) a reliable assignment to a specific carbon site in the framework is not straightforward. Using the point-dipole approximation, the T⊥C parameter provides a Cu · · · C distance to

the next neighboring carbon atom of about 2.4 Å, which is too small in comparison with the distance 2.8 Å taken from the crystal structure.5 Therefore, we must consider spin delocalization effects across the paddle wheel that prevent a solid interpretation of the 13C hf couplings by such a simplified approach. We have to refer to quantum chemical calculations of the binuclear Cu/Zn paddle wheel unit that are beyond the scope of this work and currently in progress. However, the decrease in ACiso upon coordination of the CH3OH donor molecule to the Cu(II) ion via its open axial binding site indicates already the significant influence of the axial ligand on the details of the spin density distribution and electronic structure of the Cu/Zn paddle wheel unit. 5. Conclusions The presented cw and pulsed ESR results showed that we have successfully substituted paramagnetic Cu(II) ions in the paddle wheel units of Cu3(btc)2 by diamagnetic Zn(II) ions. Although this work has concentrated on low Zn doping levels of up to about 1%, it confirms that the synthesis of mixed metal HKUST-1 structures is in principle feasible. The ESR spectroscopic methods clearly prove the presence of magnetically diluted binuclear Cu/Zn paddle wheel units that give rise to a paramagnetic behavior of the framework at low temperatures. Otherwise, the majority of the paddle wheels are still formed by the antiferromagnetically coupled Cu(II) pairs that determine the magnetic properties of the Cu2.97Zn0.03(btc)2 at temperatures above 100 K. We note that Cu2.97Zn0.03(btc)2 is one of the rare examples where the magnetic framework properties can be tuned by an isomorphous framework substitution of paramagnetic ion by a diamagnetic ion. In addition, our approach creates an excellent ESR sensitive probe by incorporation of a suitable paramagnetic metal species into the framework. These spin probes can be exploited to investigate the interactions of adsorbate molecules with the MOF framework in a detailed manner for example, as outlined here for the case of methanol. Pulsed ENDOR and HYSCORE spectroscopy applicable only for magnetically diluted frameworks such as Cu2.97Zn0.03(btc)2 give precious information on the coordination of the adsorbate molecules that is not available for the parent undoped Cu3(btc)2 material. Acknowledgment. We thank the DFG in the frame of the Priority Program 1362 for financial support. Supporting Information Available: Powder X-ray diffraction patterns and 1H MAS NMR spectra of I and II, N2

ESR Spectroscopy of Cu3-xZnx(btc)2 adsorption isotherms, TG/DTA data, X-band ESR temperature dependence, X-band ESR spectra after removal of CH3OH adsorbates, FS ESE spectra, Q-band ENDOR and HYSCORE spectra of I. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Yaghi, O. M.; Davids, C. E.; Li, G. M.; Li, H. L. J. Am. Chem. Soc. 1991, 119, 2861. (2) Li, H. L.; Eddaoudi, M.; O’Keeffe, M.; Yaghi, O. M. Nature 1999, 402, 276. (3) Fe´rey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F.; Dutour, J.; Surble´, S.; Margiolaki, I. Science 2005, 309, 2040. (4) Kitagawi, S.; Kitaura, R.; Nore, S. Angew. Chem., Int. Ed. 2004, 43, 2334. (5) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P. H.; Orpen, A. G.; Williams, I. D. Science 1999, 283, 1148. (6) Zhang, X. X.; Chui, S. S.-Y.; Williams, I. D. J. Appl. Phys. 2000, 87, 6007. (7) Li, H.; Eddaoudi, M.; O’Keefe, M.; Yaghi, O. M. Nature 1999, 402, 276. (8) Schlichte, K.; Kratzke, T.; Kaskel, S. Microporous Mesoporous Mater. 2004, 73, 81. (9) Prestipino, C.; Regli, L.; Vitillo, J. G.; Bonino, F.; Damin, A.; Lamberti, C.; Zecchina, A.; Solari, P. L.; Kongshaug, K. O.; Bordiga, S. Chem. Mater. 2006, 18, 1337. (10) Mueller, U.; Schubert, M.; Teich, F.; Puetter, H.; Schierle-Arndt, K.; Pastre´, J. J. Mater. Chem. 2006, 16, 626. (11) Hartmann, M.; Kunz, S.; Himsl, D.; Tangermann, O.; Ernst, S.; Wagener, A. Langmuir 2008, 24, 8634.

J. Phys. Chem. C, Vol. 114, No. 39, 2010 16639 (12) Wang, Z.; Cohen, S. M. Chem. Soc. ReV. 2009, 38, 1315. (13) Zeng, M.-H.; Wang, B.; Wang, X.-Y.; Zhang, W.-X.; Chen, X.M.; Gao, S. Inorg. Chem. 2006, 45, 7069. (14) Schweiger, A.; Jeschke, G. Principles of pulse electron paramagnetic resonance; Oxford University Press: Oxford, U.K., 2001. (15) Goldfarb, D.; Bernardo, M.; Strohmaier, K. G.; Vaughan, D. E. W.; Thomann, H. J. Am. Chem. Soc. 1994, 116, 6344. (16) Hartmann, M.; Kevan, L. Chem. ReV. 1999, 99, 635. (17) Weckhuysen, B. M.; Heidler, R.; Schoonheydt, R. A. Mol. SieVes 2004, 4, 295. (18) Kevan, L. Acc. Chem. Res. 1987, 20, 1. (19) Po¨ppl, A.; Kunz, S.; Himsl, D.; Hartmann, M. J. Phys. Chem. C 2008, 112, 2678. (20) Kokoszka, G. F.; Allen, H. C., Jr. J. Chem. Phys. 1965, 42, 3693. (21) Davies, E. R. Phys. Lett. A 1974, 47A, 1. (22) Ho¨fer, P.; Grupp, A.; Nebenfu¨hr, H.; Mehring, M. Chem. Phys. Lett. 1986, 132, 279. (23) Hoentsch, J.; Rosentzweig, Yu.; Ko¨hler, K.; Gutjahr, M.; Po¨ppl, A.; Vo¨lkel, G.; Bo¨ttcher, R. Appl. Magn. Reson. 2003, 25, 249. (24) Po¨ppl, A.; Hartmann, M.; Bo¨hlmann, W.; Bo¨ttcher, R. J. Phys. Chem. A 1998, 102, 3599. (25) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42. (26) Jee, B.; Himsl, D.; Icker, M.; Hartmann, M.; Po¨ppl, A. Chem. Ing. Tech. 2010, 82, 1025. (27) Carl, J. C.; Larsen, S. C. J. Phys. Chem. B 2000, 104, 6568. (28) Po¨ppl, A.; Kevan, L. J. Phys. Chem. 1996, 100, 3387. (29) Hathaway, B. J.; Billings, D. E. Coord. Chem. ReV. 1970, 5, 143. (b) Tominaga, H.; Ono, Y.; Keii, T. J. Catal. 1975, 40, 107. (30) Kita, S.; Uchida, K.; Miyamoto, R.; Iwaizumi, M. Bull. Chem. Soc. Jpn. 1991, 64, 3324.

JP105955W