Al Ratio on the Nature and Reactivity of HY Zeolite

Oct 7, 2014 - Fourier transform infrared (FTIR), extended X-ray absorption fine structure (EXAFS), X-ray photoelectron spectroscopy (XPS) measurements...
2 downloads 6 Views 3MB Size
Article pubs.acs.org/JPCC

Effect of Si/Al Ratio on the Nature and Reactivity of HY ZeoliteSupported Rhodium Dicarbonyl Complexes Artem Vityuk,† Hristiyan A. Aleksandrov,‡ Georgi N. Vayssilov,‡ Shuguo Ma,† Oleg S. Alexeev,*,† and Michael D. Amiridis*,† †

Department of Chemical Engineering, University of South Carolina, Columbia, South Carolina 29208, United States Faculty of Chemistry and Pharmacy, University of Sofia, Blvd. J. Bauchier 1, BG-1126 Sofia, Bulgaria



S Supporting Information *

ABSTRACT: Fourier transform infrared (FTIR), extended Xray absorption fine structure (EXAFS), X-ray photoelectron spectroscopy (XPS) measurements, and DFT calculations were used to characterize the species formed following reaction of a Rh(CO)2(acac) precursor with dealuminated HY zeolites with different Si/Al ratios. The results indicate the formation of two types of Rh(CO)2 species with characteristic νCO bands at 2117/2053 and 2110/2043 cm−1. Both species are attached to the zeolite framework and have similar structural properties. However, their thermal stabilities are different, and the fraction of each species formed strongly depends on the Si/Al ratio of the zeolite. The carbonyl ligands in both types of these zeolite-bound Rh(CO)2 complexes can react with gas-phase C2H4 to form Rh(CO)(C2H4) species. However, the reaction rate is substantially higher for the Rh(CO)2 complex with the νCO bands at 2117/2053 cm−1, suggesting that the electronic properties of the Rh site affect the reactivity of the carbonyl ligands. The results further indicate that the two types of Rh(CO)2 species are not associated with unreacted or partially reacted Rh(CO)2(acac) complexes or with the formation of a hydrated Rh(CO)2(H2O)x-type species. Instead, the results of the DFT calculations suggest that the differences observed can be attributed to differences in the structure of the binding sites in the dealuminated faujasites framework. achieved.7,9 More specifically, Rh(CO)2(acac) complexes can react with surfaces of dealuminated Y zeolites, leading to the displacement of the acac ligand and the formation of siteisolated, well-defined Rh(CO)2 species anchored inside the zeolite supercage.7 These species are characterized by sharp νCO bands at approximately 2117 and 2053 cm−1, with a fwhm of approximately 6 cm−1. Such narrow νCO bands signify the highly uniform nature of the anchored species formed. On the basis of EXAFS data and DFT calculations, it has been further suggested that these Rh(CO)2 species retain their square-planar geometry upon anchoring, and two oxygen atoms, located in the T4 ring of the zeolite and coordinated to Al3+ cations, represent the binding sites for these species.7 However, in cases where faujasites with lower Si/Al ratios were used as supports for the Rh(CO)2 complexes, the presence of broader νCO bands was observed, consistent with the formation of additional types of rhodium dicarbonyl species. For example, Rode et al.10 have reported two types of Rh(CO)2 species formed on a NaY zeolite (Si/Al = 2.4) with νCO bands at 2111/2045 and 2097/2019 cm−1 and assigned them to species located inside the zeolite pores and on the

1. INTRODUCTION The importance of rhodium carbonyl complexes as catalysts for a variety of industrially relevant liquid-phase reactions is well documented in the literature.1−3 Since the use of solid catalytic materials offers significant operational advantages for many of the same applications, substantial research efforts have been focused on developing heterogeneous analogues for freestanding rhodium complexes. Metal oxides frequently used as catalyst supports have nonuniform surfaces and, therefore, a variety of binding sites with different structural and electronic properties.4,5 As a result, metal complexes grafted on such supports yield catalytic sites that are nonuniform in structure and composition and exhibit different electronic properties, impacting substantially their catalytic performance.6 In this respect, the use of zeolites as supports is more promising since these crystalline materials offer highly ordered arrays of binding sites for metal complexes and, therefore, allow for the preparation of catalytic materials with well-defined and nearly uniform structures. During the past decade, significant progress has been made toward the synthesis and understanding of the structural and catalytic properties of zeolite-supported Rh carbonyl complexes.7,8 When highly dealuminated HY zeolites with a Si/Al ratio of 30 or higher were used as supports, the synthesis of well-defined and site-isolated Rh(CO)2 complexes was © XXXX American Chemical Society

Received: July 26, 2014 Revised: October 6, 2014

A

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

UHP grade) for 2 h, and then evacuation (10−3 Torr) at 400 °C for 16 h. The BET surface area of the resulting material was approximately 100 m2/g. Commercial CBV760, CBV720, and CBV600 dealuminated HY zeolites (Zeolyst International) with Si/Al atomic ratios of 30, 15, and 2.6, respectively, were calcined in flowing O2 at 300 °C for 3 h and then evacuated at 10−3 Torr and 300 °C for 16 h. For simplicity, these zeolite supports are further denoted as HY30, HY15, and HY2.6, respectively. All treated supports were stored and handled in a glovebox (MBraun) filled with dry N2. The residual water and O2 concentrations in the glovebox were kept below 0.1 ppm. 2.2. Preparation of Supported Samples. The syntheses and sample transfers were performed with exclusion of air and moisture in a double-manifold Schlenk line and a N2-filled MBraun glovebox. Supported samples were prepared by slurrying the Rh(CO)2(acac) precursor with the corresponding powder support in n-pentane under N2 for 24 h at room temperature, followed by overnight evacuation at 25 °C to remove the solvent. In each case, the Rh(CO)2(acac) precursor was added in the amount needed to yield samples containing 1 wt % Rh. The Rh weight loading was verified by inductively coupled plasma−mass spectrometry (ICP−MS) analysis (Galbraith Laboratories Inc.). The prepared samples were stored and handled in a glovebox filled with dry N2 to prevent possible contamination and decomposition of supported species. 2.3. Interaction of Supported Rh(CO)2 Complexes with Phosphines. Treatments of supported rhodium dicarbonyl species with phosphines were performed with exclusion of air and moisture in a double-manifold Schlenk line and a N2-filled MBraun glovebox. In a typical experiment, a Schlenk flask was loaded with a dry Rh(CO)2/HY powder sample in the glovebox, sealed, transferred to a stirring plate located outside the glovebox, and filled with n-pentane under N2 flow. The slurry thus formed was stirred for approximately 15 min to ensure homogeneous distribution of the solid material. Pentane solutions of triethylphosphine (P(C2H5)3) and tris(2,4-dimethylphenyl)phosphine (P[C6H3(CH3)2]3) complexes were prepared separately and added to the Rh(CO)2/HY slurry under N2 flow in the amounts needed to achieve Rh/phosphine molar ratios of 1:5. The mixture was then stirred under N2 for approximately 1 h to allow the completion of the reaction between the phosphines and supported Rh(CO)2 species. When this procedure was completed, the solvent was removed, and the solid material was washed several times with pure n-pentane to remove unreacted phosphines. Finally, the solid sample was evacuated overnight at 25 °C to completely remove the solvent and transferred into a glovebox. 2.4. FTIR Spectroscopy. A Nicolet Nexus 470 spectrometer equipped with a MCT-B detector cooled by liquid nitrogen was used to collect spectra with a resolution of 2 cm−1, averaging 64 scans per spectrum. Each powder sample was pressed into a self-supported wafer with a density of approximately 20 mg/cm2 and mounted in a cell connected to a gas distribution manifold. A homemade cell was used, allowing for the in situ treatment of samples at various temperatures and under different gas flowing environments. 2.5. X-ray Absorption Spectroscopy (XAS). XAS spectra were collected at X-ray beamline 4−1 of the Stanford Synchrotron Radiation Laboratory (SSRL), Stanford Linear Accelerator Center, Menlo Park, CA. The storage ring electron

external surface, respectively. Similarly, Maneck et al.9 have observed the νCO bands of NaY (Si/Al = 2.6) and NaX (Si/Al = 1.3) supported Rh(CO)2 complexes at 2118/2053 cm−1 but also identified a second type of Rh(CO)2 species on these supports with characteristic νCO bands at 2099/2020 and 2096/ 2015 cm−1, respectively. These authors suggested that Rh(CO)2 species anchored next to zeolite framework Al atoms that are isolated by one or more than one Si atoms have different fingerprints in the νCO region. Furthermore, Shannon et al.11 have examined the reaction of CO with a Rh exchanged Y zeolite (Si/Al = 2.4) and observed two types of Rh(CO)2 species formed, which were bound differently to the support. This group suggested that the νCO bands at 2101 and 2022 cm−1 represent Rh(CO)2 species bound to two framework oxygen atoms, while the νCO bands at 2116 and 2048 cm−1 represent similar species bound to one framework oxygen and one water molecule. While several other reports appear to support such an assignment,12,13 data reported by Lefebvre et al.14 strongly suggest that coordination of water molecules to zeolite-supported Rh(CO)2 complexes results in a redshift of the corresponding νCO bands, as the νCO bands at 2216/2048, 2101/2022, and 2090/2030 cm−1 were assigned to Rh(CO) 2 (O support ) 2 , Rh(CO) 2 (H 2 O)(O support ), and Rh(CO)2(H2O)2 (or Rh(CO)2(H2O)(OH)) complexes, respectively. None of these reports focused on the reactivity of the different Rh(CO)2 species. The origin of different Rh(CO)2 complexes formed in zeolites with different Si/Al ratios is important for understanding the catalytic properties of these materials. Since several factors (i.e., different binding sites, zeolite acidity, and residual water content) could have a significant impact on the reactivity of zeolite-bound Rh(CO)2 species, we have attempted to approach these issues systematically. In this work we examine the structural and electronic properties of two types of Rh(CO)2 species formed in cages of dealuminated Y zeolites with different Si/Al ratios. FTIR, EXAFS, and XPS measurements, as well as DFT calculations, were used to probe the nature of the anchored complexes formed, their stability, and reactivity in a simple reaction with C2H4. The results obtained suggest that while dealuminated Y zeolites act as macroligands for these metal complexes they have at least two different types of binding sites capable of accommodating the Rh(CO)2 moieties. The fraction of these sites depends on the Si/Al ratio, while the two types of supported Rh(CO)2 complexes formed exhibit significant differences in their reactivity toward C2H4.

2. EXPERIMENTAL METHODS 2.1. Reagents and Materials. Dicarbonylacetylacetonato rhodium(I) Rh(CO)2(acac) (acac = C5H7O2) (Strem, 98% purity), triethylphosphine (Strem, 99% purity), and tris(2,4dimethylphenyl)phosphine (Strem, 98% purity) were used as supplied. n-Pentane (Aldrich, 99% purity) was refluxed under N2 in the presence of Na/benzophenone ketyl to remove traces of moisture and deoxygenated by sparging with dry N2 prior to use. All glassware used during catalyst preparation was previously dried at 120 °C. He and C2H4 (Airgas, UHP grade) were additionally purified prior to their use by passage through oxygen/moisture traps (Agilent) capable of removing traces of O2 and water to 15 and 25 ppb, respectively. The γAl2O3 support was prepared by forming a paste of aluminum oxide C (Degussa) and deionized water, followed by overnight drying at 120 °C, calcination at 400 °C in flowing O2 (Airgas, B

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Table 1. EXAFS Structural Parameters Characterizing Surface Species Formed from the Rh(CO)2(acac) Precursor on Different Supportsa k1-variances (%) support

shell

N

R (Å)

Δσ (Å )

ΔE0 (eV)

ε2v

im.

abs.

HY30b

Rh−Rh Rh−C* Rh−O* Rh−Osupport Rh−Os Rh−Ol Rh−Rh Rh−C* Rh−O* Rh−Osupport Rh−Os Rh−Ol Rh−Rh Rh−C* Rh−O* Rh−Osupport Rh−Os Rh−Ol Rh−Rh Rh−C* Rh−O* Rh−Osupport Rh−Os Rh−Ol

-2.0 2.3

-1.84 2.99

-0.00280 0.00218

-10.0 −7.2

1.4

0.4

0.2

2.3 1.9 -1.9 2.3

2.13 2.75 -1.84 2.99

0.00051 0.00018 -0.00211 0.00258

0.3 −6.2 -10.0 −7.3

1.8

0.6

0.2

2.4 2.0 -1.9 2.2

2.13 2.75 -1.84 2.99

0.00074 0.00105 -0.00203 0.00241

−0.6 −6.1 -10.0 −7.4

1.6

0.5

0.2

2.4 1.9 -1.6 2.1

2.13 2.76 -1.84 2.96

0.00084 0.00039 -0.00519 0.0039

−0.6 −6.4 -10 6

2.1

0.8

0.7

3.3 3

2.13 2.81

0.00755 0.00471

−2.7 −9.1

HY15b

HY3b

Al2O3b

2

2

N, coordination number; R, distance between absorber and backscatterer atoms; Δσ2, Debye−Waller factor relative to the Debye−Waller factor of the reference compound; ΔE0, inner potential correction accounting for the difference in the inner potential between the sample and the reference compound; ε2v , goodness of fit; the superscript * refers to carbonyl ligands. Standard deviations in fits: N ± 20%, R ± 1%, Δσ2 ± 10%, ΔE0 ± 10%. b R-space fit ranges 3.5< k < 15.0 Å−1 and 0.5< r < 3.5 Å; 23 allowed fitting parameters. a

transferability of the phase shifts and backscattering amplitudes for near neighbors in the periodic table has been justified experimentally.16 The parameters used to extract these files from the EXAFS data are reported elsewhere.17 The EXAFS data were extracted from the spectra with the XDAP software developed by XAFS Services International.18 The EXAFS function for each sample was obtained from the X-ray absorption spectrum by a cubic spline background subtraction and normalized by dividing the absorption intensity by the height of the absorption edge. The final normalized EXAFS function for each sample was obtained from an average of six scans. The parameters characterizing both low-Z (O, C) and high-Z (Rh) contributions were determined by multiple-shell fitting with a maximum of 16 free parameters in r (where r is the distance from the absorbing atom, Rh) and k (wave vector) space over the ranges of 0.5 < r < 3.5 Å and 3.5 < k < 15.0 Å−1, respectively, and with application of k1 and k3 weighting of the Fourier transform. The statistically justified number of free parameters (n), estimated from the Nyquist theorem,19,20 n = (2ΔkΔr/π) + 1, where Δk and Δr are the k and r ranges used to fit the data, was approximately 23. The fit was optimized by use of a difference file technique,21,22 with phase- and amplitude-corrected Fourier transforms. The best fit parameters determined for each sample examined are summarized in Table 1, while comparisons of the data and fits in k and r space are shown in Figures S1−S4 (Supporting Information). Standard deviations reported in Table 1 for the various parameters were calculated with the XDAP software, as described elsewhere.23

energy was 3 GeV, and the ring current was in the range of 495−500 mA. XAS measurements were used to characterize the surface species formed after the impregnation of Rh(CO)2(acac) on the support. Prior to these measurements, each powder sample was pressed into a wafer inside a N2-filled glovebox. The sample mass was calculated to give an absorbance of approximately 2.5 at the Rh K absorption edge. After the sample was pressed, it was loaded into an EXAFS cell,15 sealed under N2, and removed from the glovebox. The cell was evacuated at 10−5 Torr and aligned in the X-ray beam. The XAS data were collected at liquid nitrogen temperature in the transmission mode with a Si(220) double-crystal monochromator that was detuned by 30% to minimize effects of higher harmonics in the X-ray beam. Samples were scanned at energies near the Rh K absorption edge (23 220 eV). All spectra were calibrated with respect to Rh foil, the spectrum of which was collected simultaneously. 2.6. Extended X-ray Absorption Fine Structure (EXAFS) Data Analysis. The EXAFS data were analyzed with experimentally determined reference files obtained from EXAFS data characterizing materials of known structure. The Rh−Rh and Rh−Osupport contributions were analyzed with phase shifts and backscattering amplitudes obtained from EXAFS data for Rh foil and Rh2O3, respectively. The Rh−C and Rh−O* contributions (where O* represents carbonyl oxygen) were analyzed with phase shift and backscattering amplitudes obtained from EXAFS data characterizing crystalline Ru3(CO)12, which has only terminal CO ligands. The C

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

computational approach and the experimentally measured (anharmonic) frequency of CO in the gas phase (i.e., 2143 cm−1)

Systematic errors are not included in the calculation of the standard deviations. The values of the goodness of fit (ε2v ) were calculated with the XDAP software as outlined in the Reports on Standards and Criteria in XAFS Spectroscopy.24 The variances in both the imaginary and absolute parts were used to determine the quality of the fit.25 2.7. XPS Measurements. XPS measurements were conducted using a Kratos AXIS Ultra DLD system equipped with a monochromatic Al Kα source. The system energy scale was calibrated using Ag foil with the Ag 3d5/2 binding energy set at 368.21 ± 0.025 eV for the monochromatic Al Kα source operated at 15 keV and 120 W. The pass energy was fixed at 40 eV for the detailed scans. A charge neutralizer was used to compensate for the surface charging during the photoemission. Powder samples (approximately 5 mg) were loaded into an airtight cell in the N2-filled glovebox and transferred without air exposure into the UHV chamber for the XPS analysis. The C 1s signal with a binding energy of 285.0 eV was used as an internal reference for calibration of the Rh 3d5/2 and Rh 3d3/2 binding energy values. All binding energies reported in this work were measured with a precision of ±0.1 eV. XPS data were analyzed by nonlinear curve fitting using the XPSPEAK 4.1 software. In all cases, a linear-type background was subtracted from the spectra, and a curve fit was performed using the minimum number of G/L-type peaks that provides a good fit. In each case the fitting routine was completed when the coefficient of determination (R2) value was 0.98 or higher. 2.8. Computational Method and Models. Periodic DFT calculations were performed with the PW91 exchangecorrelation functional26 using a Vienna ab initio simulation package (VASP).27,28 Ultrasoft pseudopotentials29,30 were used as implemented in the VASP package. Due to the large size of the unit cell (see below), the Brillouin zone was sampled using only the Γ point.31 The valence wave functions were expanded in a plane-wave basis with a cutoff energy of 400 eV. The cubic unit cell of the zeolite framework was optimized for the pure silicate structure with dimensions a = b = c = 24.345 Å.32 To simulate the structure of a highly dealuminated HY zeolite, one Si atom in the unit cell was replaced with Al. The negative charge around the Al site was compensated by the Rh+ ion or its complexes. During the geometry optimization procedure, all the zeolite atoms and the adsorbate species were allowed to relax until the force on each atom was less than 5 × 10−4 eV/pm. The binding energy (BE) of the CO ligands was determined as

ν(C − O)calcd = νcalcd − νcalcd(CO‐gas) + 2143

In this case, the calculated νCO frequencies are corrected for both the anharmonicity (which is 39 cm−1 for gas-phase CO) and the systematic error of the computational method.



RESULTS AND DISCUSSION 1. Different Types of Supported Rh(CO)2 Species. The room temperature infrared spectrum of the Rh(CO)2(acac) precursor dissolved in n-pentane exhibits two strong bands in the νCO region at 2083 and 2014 cm−1 (Figure 1) due to the

Figure 1. Room-temperature FTIR spectrum of the Rh(CO)2(acac) precursor in pentane.

symmetric and asymmetric νCO vibrations of the carbonyl ligands, respectively. These bands are very narrow, with fwhm values of approximately 3.5 cm−1, and each can be fitted with a single component to achieve R2 coefficients of determination on the order of 0.998 or higher. Two very weak bands were also observed in the spectrum at 2065 and 1984 cm−1 (Figure 1), consistent with the symmetric and asymmetric νCO vibrations, respectively, of carbonyl ligands in Rh(CO)(13CO)(acac) complexes, which are present in trace amounts. In contrast to the crystalline form of the Rh(CO)2(acac) precursor, in which the square-planar Rh(CO)2(acac) species are arranged in such a way that the Rh atoms of neighboring molecules form pseudooctahedral structures,33 the spectrum of Figure 1 represents the FTIR fingerprint of isolated and well-defined Rh(CO)2(acac) species in solution. The interaction of acetylacetonate complexes of different metals, including Rh, with γ-Al2O3 and zeolite surfaces has been examined extensively in the past.34−40 From these reports, it is evident that Rh(CO)2(acac) readily reacts with acidic OH groups on these supports, leading to protonation and removal of the acac ligand from the metal complex and the formation of anchored mononuclear Rh(CO)2 species on the surface of the support. Consistent with this type of surface chemistry, the FTIR data shown in Figures 2A−C provide evidence for the formation of such surface species from the Rh(CO)2(acac)

BE[Rh(CO)2+ /Zeo] = E[Rh(CO)2+ /Zeo] − E[Rh+/Zeo] − 2E[CO]

where E[Rh(CO)2+/Zeo] is the energy of the zeolite system together with the metal cation and adsorbed CO molecules in the optimized geometry; E[CO] is the energy of the gas-phase CO; and E[Rh+/Zeo] is the energy of the initial zeolite system containing a bare Rh+ cation. Consistent with this definition, negative values of BE imply a favorable interaction. The vibrational frequencies for periodic models were obtained from a normal-mode analysis where the elements of the Hessian were approximated as finite differences of gradients, displacing each atomic center by 1.5 pm either way along each Cartesian direction. All calculated C−O vibrational frequencies were shifted by the difference of the calculated harmonic frequency of the free CO obtained with the same D

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Figure 2. Room-temperature FTIR spectra in the νCO region of (A) Rh(CO)2/HY30, (B) Rh(CO)2/HY15, (C) Rh(CO)2/HY2.6, and (D) Rh(CO)2/γ-Al2O3 samples (solid lines) and deconvolution results (dotted lines).

Table 2. Deconvolution Parameters of νCO Bands Observed in FTIR Spectra of Various Samples sample Rh(CO)2/HY30

Rh(CO)2/HY15

Rh(CO)2/HY2.6

Rh(CO)2/γ-Al2O3 Rh(CO)2(acac) in pentane a

band position, cm−1

fwhm, cm−1

split (νs−νas),a cm−1

C−Rh−C angle, deg

relative fraction, %

2117 2053 2113 2048 2117 2053 2113 2048 2117 2053 2110 2043 2090 2014 2083 2014

3.6 4.8 4.8 4.8 5.1 7.3 7.7 9.9 7.1 9.3 20.2 28.5 20.0 27.2 3.5 3.4

64 65

96 94

83 17

64 65

94 95

60 40

64 67

99 94

50 50

76

98

-

69

97

-

The “s” and “as” subscripts refer to symmetric and asymmetric vibrations, respectively.

E

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

The positions of the νCO bands assigned to type I Rh(CO)2 species (i.e., 2117 and 2053 cm−1) are identical for all three zeolites used. While the same is true for the type II Rh(CO)2 species (νCO bands at 2113 and 2048 cm−1) formed on the zeolites with Si/Al ratios of 30 and 15, a small shift of these bands to lower frequencies was observed when the zeolite with a Si/Al ratio of 2.6 was used (Table 2). However, this shift is only marginal, and therefore, it appears that the Si/Al ratio in the range of 2.6−30 does not affect the νCO vibrational frequency of HY zeolite-supported rhodium dicarbonyl species. The results summarized in Table 2 for the Rh(CO)2/HY30 sample show that the relative fractions of type I and type II Rh(CO)2 species formed in this case are 83 and 17%, respectively. In the case of the Rh(CO)2/HY15 sample, type I and II species constitute 60 and 40% of Rh dicarbonyls, respectively, while equal fractions of the two species are formed in the Rh(CO)2/HY2.6 sample. These results show that an increase in the number of Al atoms in the zeolite framework results in a larger fraction of type II Rh(CO)2 species being formed. Both types of zeolite-supported Rh(CO)2 species formed in the case of the Rh(CO)2/HY30 sample exhibit a very high degree of uniformity since their νCO bands are very narrow, with fwhm values closely resembling those of Rh(CO)2(acac) complexes in solution (Table 2). At a Si/Al ratio of 15, the width of the νCO bands increased for both types of Rh(CO)2 species, although they still appear to remain relatively uniform with fwhm values below 10 cm−1. In the case of the Rh(CO)2/ HY2.6 sample, however, while the fwhm value for the type I species is still below 10 cm−1, the νCO bands of the type II species are substantially broader (average fwhm of approximately 24 cm−1). This comparison indicates that the type I zeolite-supported Rh(CO)2 species are not affected significantly by the Si/Al ratio, but the type II species are sensitive to the presence of Al in the zeolite framework. In contrast to what was observed with zeolites, when Rh(CO)2 complexes were supported on a γ-Al2O3 surface, the νCO bands of Rh(CO)2 species formed were observed at 2090 and 2014 cm−1 (Figure 2D), resembling those of Rh(CO)2(acac) complexes in n-pentane solution (Table 2). This result suggests that the γ-Al2O3 support does not affect the electronic properties of Rh in the supported Rh(CO)2 species. While the C−Rh−C angle in the γ-Al2O3-supported Rh(CO)2 species is similar to that in the zeolite-bound complexes (Table 2), the νCO bands are wider with an average fwhm of approximately 24 cm−1, indicating some nonuniformity for the γ-Al2O3-supported Rh(CO)2 complexes. Most likely, this can be attributed to the nonuniformity of the anchoring surface alumina sites, in contrast to the uniform nature of the sites in the crystalline zeolite framework. 2. Thermal Stability of Supported Rh(CO)2 Species. To further examine the thermal stability of the supported Rh dicarbonyl species of both types, FTIR spectra were collected during exposure of the Rh(CO)2/HY2.6 sample (i.e., the sample with the largest fraction of type II Rh dicarbonyl species) to different temperatures under the flow of He. During these measurements, the sample was heated at a rate of 3 °C/ min to a desired temperature and held at this temperature until no changes in the νCO region of the spectra were observed. A deconvolution procedure was applied to each resulting spectrum to quantify the relative surface concentration of each type of Rh(CO)2 species and determine the percentage remaining on the surface at each temperature. The results

precursor on HY zeolites dealuminated to various degrees. In all these spectra, the supported Rh(CO)2 species can be identified by two strong bands in the νCO region at 2117 and 2053 cm−1, assigned to the symmetric and asymmetric νCO vibrations of the carbonyl ligands, respectively. Additional very weak bands in the νCO region at 2102 and 2022 cm−1 originate from supported Rh(CO)(13CO) complexes which are present in trace amounts, while those at 2125 and 2065 cm−1 most likely represent a combination of CO stretching and deformation vibrations in these complexes.41 Further analysis of the strong νCO bands shown in Figures 2A−C indicates different degrees of asymmetry at the low frequency side of each band. As the Si/Al ratio decreases, the asymmetry of the νCO bands becomes more apparent, with the spectrum of the Rh(CO)2/HY2.6 sample (Si/Al = 2.6) with the largest content of Al (Figure 2C) clearly indicating that several types of supported Rh(CO)2 species are formed. The deconvolution results shown in Figure 2C confirm this point, as bands corresponding to two different types of Rh dicarbonyl species can be clearly identified. The first type of Rh(CO)2 species is characterized by a pair of strong sharp νCO bands at 2117 and 2053 cm−1, while broader and less intense bands at 2110 and 2043 cm−1 represent the second type of Rh dicarbonyl species formed. The presence of the second set of bands in the spectra of Rh(CO)2/HY30 and Rh(CO)2/HY15 is less apparent (Figures 2A and 2B). However, acceptable fits with R2 coefficients of determination above 0.95 can be obtained for these two spectra only when two components for each νCO band are included in the fit. On the basis of these deconvolution results (Figures 2A and 2B), it is evident that two different types of Rh dicarbonyl species are also present in these samples, although at lower concentrations. The deconvolution results summarized in Table 2 for all zeolite samples examined lead to several conclusions. Regardless of the type of zeolite used as the support, the infrared characteristics for both types of detected Rh(CO)2 species are nearly identical, including not only the position of the bands and the split between symmetric and asymmetric νCO vibration modes but also the C−Rh−C angles calculated from the Isym/Iasym ratios, as reported elsewhere.41 Moreover, these split and angle parameters characteristic of the zeolite-grafted Rh(CO)2 species closely resemble those of the Rh(CO)2(acac) species in solution (Table 2), indicating that the replacement of the acac ligands by the zeolite support does not affect substantially the geometry of the Rh(CO)2 moieties. A comparison of the positions of the νCO bands of zeolitebound Rh(CO)2 species and Rh(CO)2(acac) complexes in solution further indicates that oxygen atoms in the zeolite framework are more electronegative than those in the acac ligand, as the νCO bands of the anchored Rh(CO)2 species are shifted substantially to the high frequency region in the former case. This result is consistent with previous literature reports,42 implying that Rh atoms in zeolite-supported Rh(CO)2 species are more electron deficient than those in free Rh(CO)2(acac) complexes. Such differences in the electronic properties of Rh are expected to have an effect on the chemical properties of the carbonyl ligands, especially when the zeolite support acts as a macroligand for the supported species.43 Finally, the differences in the position of the νCO bands of the two types of Rh(CO)2 complexes observed when zeolites were used as supports (Table 2) indicate that the Rh atoms in these complexes are not identical in terms of their electronic properties, which could have an effect on the reactivity of the carbonyl ligands. F

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

summarized in Figure 3 show that the surface concentration of the type I Rh(CO)2 species with νCO bands at 2117 and 2053

Figure 4. Percent of Rh dicarbonyl species remaining on the surface of Rh(CO)2/γ-Al2O3 during thermal treatments with He at different temperatures. Figure 3. Percent of Rh dicarbonyl species remaining on the surface of Rh(CO)2/HY2.6 during thermal treatments with He at different temperatures: (▲) type I (νCO at 2117/2053 cm−1) and (●) type II (νCO at 2110/2043 cm−1) species.

attached to Rh are 1.83 and 2.99 Å, respectively.33 On the basis of these data, as well as the FTIR results indicating that the C− Rh−C angles in all supported Rh(CO)2 complexes are similar to the angle in the Rh(CO)2(acac) precursor in solution (Table 2), we can infer that a facile substitution of the acac ligand by the support takes place upon anchoring, so that the structure of the Rh(CO)2 moieties remains essentially unchanged. The Rh−support interactions in all samples examined are characterized by the presence of Rh−Os contributions at an average bond distance of approximately 2.14 Å (Table 1). This distance is consistent with those reported in the literature for zeolite- or metal oxide-supported complexes of different transition metals7,37,40 but substantially longer than that between Rh and the two oxygen atoms of the acac ligand in the Rh(CO)2(acac) precursor (i.e., 2.04 Å).33 An average Rh− Os coordination number of approximately 2 was observed for all zeolite-supported samples examined. Similar to the original acac ligand, this result indicates that zeolite supports with different Si/Al ratios are capable of chelating Rh(CO)2 moieties and acting as bidentate ligands. In contrast, when γ-Al2O3 was used as the support, the average Rh−Os coordination number was found to be 3.3 (Table 1). This larger Rh−Os coordination number indicates a more complex binding between the Rh(CO)2 moieties and the γ-Al2O3 surface. This result is consistent with previous literature reports indicating that the {OAl}3 units on the γ-Al2O3 surface are primarily involved in coordination of metal carbonyl complexes16,45,46 and that supported Rh(CO)2 moieties can be located at the hollow sites between three oxygen anions of the support.45 Furthermore, Rh−Ol contributions at distances in the 2.76− 2.81 Å range were also detected in the EXAFS spectra of all samples examined (Table 1). While such contributions have been frequently reported for metal complexes supported on metal oxide surfaces,40 the assignment of neighboring atoms located at such long distances in zeolite structures is less straightforward. For the case of dealuminated Y zeolites, DFT calculations reported elsewhere7 predict coordination of Rh(CO)2 moieties near Al cations of the zeolite framework with expected Rh−Al distances of 2.8 Å. During the fitting of the EXAFS spectra for zeolite-supported Rh complexes, such

cm−1 remains unchanged, as the temperature was increased from 25 to 150 °C. At higher temperatures, removal of the CO ligands was observed, with approximately 20% of the original amount remaining on the surface at 300 °C and complete decarbonylation observed only above 350 °C. In contrast, the type II Rh(CO)2 species with νCO bands at 2110 and 2043 cm−1 were decarbonylated at lower temperatures, with approximately 53 and 10% of the original amount remaining at 100 and 250 °C, respectively (Figure 3). Complete decarbonylation was observed for this type of species at 300 °C. These results indicate that the two types of HY zeolite-supported Rh(CO)2 complexes exhibit substantially different thermal properties, with the type I species being more stable. When similar measurements were performed with the Rh(CO)2/γ-Al2O3 sample, 85% of the original amount of the Rh(CO)2 species remained intact at 100 °C (Figure 4), while at temperatures above 100 °C the removal of the carbonyl ligands had a linear dependence on temperature, with complete decarbonylation observed at 300 °C. 3. Structural Properties of Supported Rh(CO) 2 Species. EXAFS data collected for the Rh(CO)2 complexes supported on γ-Al2O3 and HY zeolites with different Si/Al ratios are summarized in Table 1. Consistent with previous literature reports,7,44 these data provide strong evidence for the formation of site-isolated Rh(CO)2 species in all samples examined. For example, the absence of Rh−Rh contributions in all EXAFS spectra indicates the mononuclear character of the Rh surface species formed. Furthermore, the Rh−C and Rh− O* contributions with average coordination numbers of approximately 2, at average bond distances of 1.83 and 2.97 Å, respectively, confirm that these mononuclear complexes are dicarbonyls. These parameters are consistent with the crystal structure of the Rh(CO)2(acac) precursor in which the Rh−C and Rh−O* bond distances for the two carbonyl ligands G

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

contributions have been often included in the fitting routine, and average coordination numbers and distances obtained for them were found to be in the 0.6−1.3 and 2.74−3.39 Å ranges, respectively.7,44,47 However, it was also reported that such coordination numbers and distances for the Rh−Al contributions were determined with rather low confidence levels. In contrast, structural parameters reported elsewhere for Mo48 and Pt49 clusters, as well as highly dispersed particles, formed in various zeolites do not include metal−Al contributions but report metal−Osupport contributions at distances longer than 2.60 Å, which represent oxygen atoms in the zeolite framework. Consistent with these latter reports, better fits were obtained for all zeolite-supported samples reported herein when Rh−Ol contributions were included in the fit. Regardless of the assignment, it is evident that backscatterers located at such long distances are not bound to Rh by chemical bonds and, therefore, are not expected to have a direct effect on the electronic and chemical properties of the Rh(CO)2 species. While the structural data presented above confirm the formation of well-defined and site-isolated Rh(CO)2 species in the cages of the different dealuminated Y zeolites, they also point out that nearly identical structures are formed in each case and do not distinguish between the two types of Rh(CO)2 species, which are evidently present in these samples based on the FTIR results. It is possible that these two types of Rh(CO)2 species have only marginal differences in their structures, which cannot be resolved in average data provided by EXAFS. Alternatively, one can also suggest that the two types of Rh(CO)2 species have structurally identical binding sites, but the oxygen atoms in these sites exhibit different electronic properties. The presence of such binding sites would explain both the FTIR and EXAFS data discussed above. 4. Electronic Properties of Supported Rh(CO) 2 Species. XPS measurements were used to determine the electronic properties of zeolite- and γ-Al2O3-supported Rh(CO)2 complexes, and the results are summarized in Figure 5 and Table 3. Spectra collected for the Rh(CO)2/HY30 sample show two relatively sharp Rh 3d5/2 and 3d3/2 peaks (fwhm of 2.4 eV) at binding energies of 308.8 and 313.5 eV, respectively.

An almost identical set of Rh 3d peaks was also observed for the Rh(CO)2/HY15 sample. However, in the case of the Rh(CO)2/HY2.6 sample, the Rh 3d5/2 and 3d3/2 peaks were found to be wider (fwhm of 2.7 eV) and the corresponding binding energies lower (i.e., 308.5 and 313.2 eV, respectively). Even broader Rh 3d5/2 and 3d3/2 peaks were observed in the case of the Rh(CO)2/γ-Al2O3 sample at binding energies of 307.9 and 312.8 eV, respectively, matching those of the Rh(CO)2(acac) crystalline precursor (Table 3). Consistent with previous literature reports,45,50−52 the Rh 3d5/2 and 3d3/2 peaks observed in all these cases can be assigned to cationic Rhδ+ (δ ∼ 1) species. However, the measurable difference in the Rh 3d core level binding energies (∼0.3 eV) for the two zeolites with the higher Si/Al ratios indicates differences in the effective overall charge of the Rh cations. To better understand these results, the O 1s region of the XPS spectra was also examined. The signal in this region arises from both support oxygen atoms, as well as the oxygen of the adsorbed carbonyl ligands. However, with only 1 wt % Rh present in all samples examined, the contribution to the O 1s peak from the support strongly dominates the spectrum, and therefore, no attempt was made to deconvolute this region. Results collected for zeolite-supported samples show that the O 1s core level binding energy shifts from 531.8 to 532.8 eV, as the Si/Al ratio increases from 2.6 to 30 and the peaks become sharper (Table 3). In the case of the γ-Al2O3-supported sample, the O 1s peak was found to be relatively wide (fwhm of 2.5 eV) with a binding energy of 530.7 eV, which is substantially lower than that in all Y zeolite-supported samples. Consistent with other literature reports,53,54 this result suggests that the oxygen atoms on the γ-Al2O3 surface are more electron rich as compared to those of the zeolite framework. It is further evident that the Al content in the zeolite framework affects significantly the electronic properties of oxygen atoms associated with Al, as O 1s binding energies increase by approximately 1 eV when the Si/Al ratio increases from 2.6 to 30 (Table 3). Moreover, since these changes are similar to those observed in the Rh 3d core level binding energies, we can conclude that the electron-accepting properties of the oxygen atoms associated with the framework Al affect the electronic properties of the Rh species significantly. This conclusion is consistent with previous literature reports indicating that such oxygen atoms represent binding sites for Rh(CO)2 species.7 Overall, the XPS results presented herein show that both the original acac ligand and the γ-Al2O3 support surface have quite similar electron-withdrawing properties, as the Rh 3d core level binding energies characterizing the Rh(CO)2(acac) precursor and the γ-Al2O3-supported Rh(CO)2 species were found to be nearly identical (Table 3). In contrast, the HY zeolitesupported Rh(CO)2 species exhibit substantially higher Rh 3d binding energies, indicating that the dealuminated zeolites are stronger electron acceptors. This conclusion is further reinforced by the FTIR results showing that the νCO bands of the zeolite-supported Rh(CO)2 species are shifted toward higher frequencies. Therefore, differences in the electronic properties of the oxygen atoms associated with Al sites appear to be the origin for the observed differences in the electronic properties of the Rh(CO)2 species on different supports. However, the XPS analysis also failed to distinguish between two types of zeolite-supported Rh(CO)2 species. The symmetrical shapes of the Rh 3d peaks (Figure 5) do not support potential deconvolution with more than one component, especially when the fwhm parameters of these peaks are in

Figure 5. XPS spectra of the Rh 3d region (solid line) and deconvolution results (dashed line) of (1) Rh(CO)2/HY2.6, (2) Rh(CO)2/HY15, and (3) Rh(CO)2/HY30 samples. H

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Table 3. XPS Data Characterizing Surface Species Formed from the Rh(CO)2(acac) Precursor on Different Supports sample

Rh 3d5/2, eV

fwhm, eV

Rh 3d3/2, eV

fwhm, eV

O 1s, eV

fwhm, eV

Rh(CO)2/HY30 Rh(CO)2/HY15 Rh(CO)2/HY2.6 Rh(CO)2/γ-Al2O3 Rh(CO)2(acac)

308.8 308.8 308.5 307.9 307.8

2.4 2.4 2.7 2.9 3.1

313.5 313.6 313.2 312.8 312.5

2.4 2.4 2.7 2.9 3.1

532.8 532.6 531.8 530.7 -

1.7 1.8 2.4 2.5 -

the range of values typically reported in the literature.51,55 This further suggests that the differences in the electronic properties of the two types of Rh(CO)2 species could be relatively small and difficult to resolve in spectra collected with a conventional XPS equipment. 5. Nature of Supported Rh(CO)2 Species. While differences in the structure and electronic properties of the binding sites in the zeolite framework could be responsible for the presence of two types of Rh(CO)2 species, analysis of previous literature reports suggests that some other factors may also be involved. For example, it has been reported that the νCO bands of Rh(CO)2 species located on external zeolite surfaces are red-shifted when compared to those of encaged complexes.47,56 Since Al atoms in dealuminated zeolites can be isolated from each other by one or more Si atoms, it was also suggested that Rh(CO) 2 complexes anchored to such structurally different Al sites have different fingerprints in the νCO region.9 Several other explanations associated with physisorbed Rh(CO)2(acac) complexes, incomplete removal of acac ligands upon anchoring, and coordination of water molecules directly to Rh(CO)2 species can be also found in the literature.11,14,47 To examine if any of these factors can potentially contribute to the appearance of two types of HY zeolite-supported Rh(CO)2 complexes in our case, additional experiments were performed, as described below. 5.1. Interaction of Zeolite-Supported Rh(CO)2 Complexes with Phosphines. The carbonyl ligands of the Rh(CO)2(acac) precursor can readily react with phosphines in solution yielding partially or fully substituted derivatives, depending on the nature of the phosphine used and the reaction conditions.57 During such substitution reactions, the formal oxidation state of Rh+ does not change, and the Rh−phosphine complexes formed retain the square planar geometry. Since the carbonyl ligands in supported Rh(CO)2 complexes can also react with phosphines, it has been suggested that this reaction can be used to distinguish between surface and encaged Rh(CO)2 species, since sufficiently large phosphines cannot enter the zeolite cages and, therefore, can react only with Rh(CO)2 complexes located on the external surface.10 Therefore, this reaction was used in the current work to determine if any zeolite-supported Rh(CO)2 complexes are located on the zeolite exterior. Experiments were performed with triethylphosphine P(C 2 H 5 ) 3 and tris(2,4-dimethylphenyl)phosphine P[C6H3(CH3)2]3 complexes having molecular diameters of 7.0 and 11.7 Å, respectively.58,59 Between these two complexes, only P(C2H5)3 can fit into the 7.4 Å aperture of the HY zeolites. As expected, when supported Rh(CO)2 complexes were treated with P(C2H5)3, the νCO bands assigned to the dicarbonyl species disappeared, while a new νCO band appeared at 1993 cm−1 in the spectra of all zeolite samples examined (Figure 6 shows an example of these changes for the Rh(CO)2/ HY30 sample). This result confirms the reactivity of zeolitesupported Rh(CO)2 species with phosphines, and the changes observed in the νCO region are consistent with the formation of

Figure 6. Room-temperature FTIR spectra in the νCO region of (1) Rh(CO)2/HY30 and (2) the same sample treated with P(C2H5)3.

Rh(CO)(P(C2H5)3) complexes, which exhibit only one νCO vibration. Such a conversion of surface species resembles closely the reaction of Rh(CO) 2 (acac) with different phosphines in solution to yield Rh(CO)(PR3)(acac) complexes.60 In contrast, when the larger P[C6H3(CH3)2]3 complex was used to treat the zeolite-supported Rh(CO)2 species, no changes in the intensities of the νCO bands of the Rh dicarbonyl species were observed, and no new νCO bands appeared in the spectra, regardless of the zeolite used. This result shows that all Rh(CO)2 complexes are located inside the zeolite pores and, therefore, are inaccessible to this bulky phosphine. Consequently, we can conclude with confidence that neither type of the two zeolite-supported Rh(CO)2 species detected by FTIR represents a Rh(CO)2 complex formed on the external zeolite surface. This conclusion is further reinforced by DFT calculations,7 demonstrating that the majority of energetically preferable binding sites for Rh(CO)2 complexes are located inside the supercages of faujasites. 5.2. Interaction of Rh(CO)2(acac) with Zeolites. Incomplete removal of the (acac) ligand during reaction of the Rh(CO)2(acac) complexes with the HY zeolites could also account for the presence of two sets of νCO bands from Rh(CO)2 species. For example, it was reported previously47 that the reaction of Rh(CO)2(acac) with H-SSZ-42 and HMordenite zeolites does not proceed to completion, and νCO bands from both zeolite-supported Rh(CO)2 species and physisorbed Rh(CO)2(acac) complexes were found to be present in the FTIR spectra of these samples. To further explore this possibility, the HY30 support was initially treated with an excess of acetylacetone/pentane solution at room temperature, corresponding to an acetylaceI

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Table 4. FTIR Bands in the νCO Region Characterizing Surface Species Formed in a Physical Mixture of Crystalline Rh(CO)2(acac) with Dry HY30 Zeolite

tone/Al molar ratio of approximately 1.8. Since acetylacetone forms strong chelate complexes with Al3+ cations,40 the intention was to block the majority of Al3+ sites by acetylacetone, making them unavailable for Rh complexes. After this treatment was completed, the support was washed with pentane to remove any unreacted acetylacetone, and the Rh(CO)2(acac) complex was introduced to the system according to the procedure described in the preparation section. Three pairs of νCO bands can be identified in the FTIR spectrum of this sample (Figure 7): the first one includes

treatment temperature/relative fraction of species (%) νCO band positions, cm−1 2115 2051 2106 2040 2089 2023

25 °C 50 °C 100 °C

150 °C 200 °C

suggested surface species

29

30

45

82

97

supported Rh(CO)2

25

31

30

18

3

partially reacted Rh(CO)2(acac)

46

39

25

-

-

physisorbed Rh(CO)2(acac)

higher temperatures. For example, the fraction of zeolite-bound Rh(CO)2 species increased from 29 to 97%, as the temperature was increased from 25 to 200 °C (Table 4). The fraction of unreacted Rh(CO)2(acac) complexes decreased from 46 to 25% at 100 °C, while no such species were present at higher temperatures. In contrast, the fraction of partially reacted Rh(CO)2(acac) complexes slightly increased at 50 °C, remained nearly unchanged at 100 °C, and declined significantly at even higher temperatures (Table 4). As expected, this pattern demonstrates the intermediate nature of the partially reacted Rh(CO)2(acac) species. Overall, the data presented herein allow us to identify the νCO bands of unreacted and partially reacted Rh(CO)2(acac) complexes and follow their transformation. Furthermore, since the position of the νCO bands of partially reacted Rh(CO)2(acac) complexes (i.e., 2106 and 2040 cm−1) closely resembles that of type II species (i.e., 2113 and 2048 cm−1; Figure 2A), incomplete displacement of the acac ligand from the original complex could reasonably explain the appearance of the type II species in the spectra of the samples examined. However, a further comparison of the thermal properties of partially reacted Rh(CO)2(acac) complexes and type II Rh(CO)2 species suggests that this is not the case, as the latter undergo decarbonylation in the 25−200 °C temperature range (Figure 3), while the former are converted into Rh(CO)2 species bound to the zeolite framework (Table 4). 5.3. Interaction of Zeolite-Supported Rh(CO)2 Complexes with H2O. While aqua complexes of Rh are well-known in solution chemistrywith some of these aqua carbonyl complexes actually catalyzing the water−gas shift reaction61−65limited information related to interactions of H2O molecules with supported Rh carbonyl complexes is available. It has been reported, for example, that water facilitates reductive carbonylation of zeolite- and alumina-supported Rh(CO)2 species at elevated temperatures to yield Rh6(CO)16 or Rh4(CO)12 clusters.66−68 At room temperature, however, this reaction does not proceed at any measurable rate, and Rh(CO)2(H2O)x complexes are formed instead following exposure of Rh(CO)2 to H2O.69,70 To determine the spectroscopic fingerprint of the νCO bands characterizing Rh(CO)2(H2O)x complexes and determine whether such complexes could represent the type II species observed, several additional experiments were performed. In the first set, the HY2.6 and HY30 zeolites were pretreated at 100−400 °C under vacuum to remove any zeolite-trapped water molecules. As expected, the presence of substantial amounts of water is evident in the samples treated at 100 °C, as indicated by infrared bands at approximately 3500 and 1630

Figure 7. Room-temperature FTIR spectra in the νCO region (solid line) and deconvolution results (dashed line) of a sample prepared by impregnation of Rh(CO)2(acac) on the HY30 support pretreated with acetylacetone.

strong bands at 2087 and 2018 cm−1 that can be assigned to physisorbed Rh(CO)2(acac) complexes since the position of these bands is similar to that of Rh(CO)2(acac) in pentane solution (Table 2); the second one includes weaker bands at 2106 and 2040 cm−1 that can be assigned to surface complexes in which the acac ligands are only partially displaced from Rh; and the third one includes minor bands at 2115 and 2051 cm−1 that can be assigned to Rh(CO)2 species supported on the zeolite framework. As expected, deconvolution results further show that physisorbed Rh(CO)2(acac) complexes constitute the majority (approximately 88%) of all surface species formed, while partially reacted Rh(CO)2(acac) and zeolite-bound Rh(CO)2 species are present only in small amounts, as their fractions in the sample do not exceed 9 and 3%, respectively. The same set of bands was observed when crystalline Rh(CO)2(acac) was carefully mixed with the HY30 support, without any solvents being used, to produce a physical mixture containing approximately 5 wt % Rh (Table 4). In this case, however, the fraction of physisorbed Rh(CO)2(acac) complexes was significantly lower (approximately 46%), while greater percentages of partially reacted Rh(CO)2(acac) (approximately 25%) and zeolite-bound Rh(CO)2 species (approximately 29%) were formed. From this result, it is evident that a significant portion of Rh(CO)2(acac) complexes can penetrate into the zeolite pores upon mixing and grinding of these solid materials and can react with the Al sites in the framework. It is further notable that the reaction between Rh(CO)2(acac) complexes and the support accelerates at J

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

cm−1 assigned to stretching and bending vibrations, respectively, of H2O molecules hydrogen-bonded to the zeolite surface (spectra are not shown for brevity). As the treatment temperature was increased to 200 °C, the bands associated with H2O declined in intensity, while relatively weak νOH bands of acidic hydroxyls appeared in the spectra at 3630 and 3565 cm−1. These bands became stronger following treatments in the 300−400 °C temperature range, while the bending vibrations of H2O were no longer present in the spectra. These results are consistent with the progressive desorption of water from the zeolite surface at elevated temperatures, suggesting that the residual water content is different depending on the temperature of the treatment. When the zeolites treated at different temperatures were further impregnated with the Rh(CO)2(acac) precursor, both type I and type II Rh(CO)2 species were detected in all samples examined (Table 5). In all Table 5. FTIR Bands in the νCO Region Characterizing Surface Species Formed Following Adsorption of the Rh(CO)2(acac) Precursor on HY30 and HY2.6 Zeolites Treated under Vacuum at Different Temperatures sample Rh(CO)2/HY30

support treatment conditions vacuum at 100 °C vacuum at 200 °C vacuum at 300 °C vacuum at 400 °C

Rh(CO)2/HY2.6

vacuum at 100 °C vacuum at 200 °C vacuum at 300 °C vacuum at 400 °C

νCO band positions, cm−1

relative fraction, %

2117/2052 2112/2048 2117/2051 2112/2048 2117/2053 2113/2048 2117/2051 2112/2048 2117/2052 2109/2045 2117/2053 2109/2043 2117/2053 2110/2043 2117/2052 2109/2042

85 15 84 16 83 17 79 21 54 46 46 54 50 50 42 58

Figure 8. Room-temperature FTIR spectra in the νCO region (solid line) and deconvolution results (dashed line) of Rh(CO)2/HY30 exposed to H2O/He feeds with different H2O partial pressures: (1) 2.5 Torr, (2) 3.1 Torr, (3) 5.2 Torr, and (4) 8.6 Torr.

be assigned to different Rh(CO)2(H2O)x complexes formed following reaction of the supported Rh(CO)2 species with H2O from the gas phase. The surface concentration of such species depends on the H2O partial pressure. Their formation is further evident from the deformation vibrations of H2O. For example, the δ(H2O) region shown in Figure 9 for the same set of

these cases, however, neither the position of the νCO bands nor the fraction of the species formed changed significantly as a result of the pretreatment temperature. These results suggest that H2O molecules residing in the zeolite pores do not promote the formation of the type II Rh(CO)2 species. In the second set of experiments, the Rh(CO)2/HY30 sample, which had the smallest fraction of type II species originally formed, was exposed to wet He feeds. Significant differences were observed in the FTIR spectra (Figure 8) under these conditions. For example, with 2.5 Torr H2O in the He feed, two sets of νCO bands can be clearly identified in the spectrum at 2117−2053 cm−1 and 2112−2046 cm−1. When the H2O partial pressure was increased to 3.1 Torr, the νCO bands at 2117−2053 cm−1 declined in intensity, while those at 2112− 2046 cm−1 slightly shifted to 2110−2043 cm−1 and increased in intensity. Simultaneously, a third set of νCO bands appeared in the spectrum at 2094−2025 cm−1. The νCO bands at 2110− 2043 cm−1 and 2094−2025 cm−1 continued to shift further toward lower frequencies and grow in intensity as the H2O partial pressure increased, while the set at 2117−2053 cm−1 disappeared at a H2O partial pressure of 8.6 Torr (Figure 8). The νCO bands at 2109−2039 cm−1 and 2090−2022 cm−1 can

Figure 9. Room-temperature FTIR spectra in the δ(H2O) region (solid line) and deconvolution results (dashed line) of Rh(CO)2/ HY30 exposed to H2O/He feeds with different H2O partial pressures: (1) 2.5 Torr, (2) 3.1 Torr, (3) 5.2 Torr, and (4) 8.6 Torr.

spectra exhibits a relatively complex band structure, consistent with the presence of two components at approximately 1630 and 1619 cm−1 with intensities highly dependent on the H2O partial pressure. The first band can be assigned to H2O molecules adsorbed in the zeolite pores and not interacting directly with the Rh complexes, while the second one represents H2O molecules specifically bound to Rh.71 Consistent with such an assignment, the band at 1619 cm−1 K

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Figure 10. Fragment of the faujasite structure containing an Al T atom (A) and a Rh+(CO)2 complex (B). Color coding: Si, yellow; O, red; Al, dark blue; Rh, silver; C, brown.

does not appear in the spectra of Rh(CO)2/HY30 under dry conditions, but it emerges immediately following the introduction of H2O to the He feed. An average number of H2O molecules coordinated to each Rh site can be roughly estimated from the integral intensity of this band, as reported elsewhere.71 Such an estimate obtained from the spectrum collected with 8.6 Torr H2O in the feed indicates an average H2O/Rh ratio of 1.4, indicating the formation of approximately 60% Rh(CO)2(H2O) and 40% Rh(CO)2(H2O)2 complexes under these conditions. These percentages closely resemble the relative ratio of the νCO bands at 2109−2039 cm−1 and 2090− 2022 cm−1 (i.e., 60 and 40%, respectively) in the same spectrum. On the basis of the above and assuming that the molar absorption coefficients of carbonyl ligands in various Rh(CO)2(H2O)x complexes are not substantially different, we can assign the νCO bands at 2109−2039 cm−1 to Rh(CO) 2 (H 2 O) and those at 2090−2022 cm −1 to Rh(CO)2(H2O)2 complexes. When H2O was removed from the feed, the νCO bands assigned to Rh(CO)2(H2O)x species gradually disappeared, while the Rh(CO)2 bands at 2117 and 2053 cm−1 reappeared in the spectrum. Consistent with previous literature reports,70 this result demonstrates that transformations between Rh(CO)2 and Rh(CO)2(H2O)x species are completely reversible. Even though the νCO bands characterizing Rh(CO)2(H2O) complexes overlap with those of type II Rh(CO)2 species, it is evident that these complexes are not related to each other since significant partial pressures of water are required to form Rh(CO)2(H2O) species. 5.4. Modeling of Supported Rh(CO)2 Complexes. To verify if structural differences in the binding sites in the zeolite framework could be responsible for the presence of two types of Rh(CO)2 species, periodic DFT calculations were also performed. The results of these calculations are summarized in Figures 10 and 11 and Tables 6 and 7. The structures are plotted with the VMD software.72 Figure 10A shows a fragment of the faujasite structure that contains an Al T atom. In the optimized structure, a Rh+ charge compensating cation is located approximately in the plane of the Si and Al T atoms of the six-membered ring of the faujasite structure, consistent with earlier reports that are based on simpler models.7 Furthermore, this Rh+ cation interacts with three oxygen atoms, two of which are from an AlO 4−

tetrahedron and another one connected to two Si atoms. When CO ligands are present on the Rh site, the Rh+ cation moves above the ring plain toward the zeolite supercage. Figure 10B shows the location of the Rh+(CO)2 complex in the cavity of the faujasite structure as an example. Since three different pairs of oxygen atoms located at the AlO4− tetrahedron are accessible from the supercage (Figure 10A), the formation of different supported Rh(CO)2 complexes which are attached to these three pairs of oxygen atoms is expected, with the structures shown in Figure 11A−C. For simplicity, the structures shown in these figures are further denoted as “a”, “b”, and “c”, respectively. The main difference between these structures is the location of the oxygen atoms in the zeolite framework that are involved in the anchoring of the Rh species. In the case of Rh(CO)2_a complexes, the oxygen atoms are from two different but coupled four-membered rings, while in the case of Rh(CO)2_b and Rh(CO)2_c complexes both oxygen atoms belong to the same four- and six-membered ring, respectively. The Rh(CO)2_a complex is the most stable one, with a binding energy of −514 kJ/mol (Table 6). The binding energies for Rh(CO)2_b and Rh(CO)2_c complexes were found to be −474 and −487 kJ/mol, respectively, indicating that the stability of supported Rh dicarbonyl species declines in the following order: Rh(CO)2_a > Rh(CO)2_c > Rh(CO)2_b. This stability pattern clearly correlates with the Rh−O distances in these complexes, which are shortest in Rh(CO)2_a and longest in Rh(CO)2_b. The same is true for the C−Rh−C angle which has its smallest value in Rh(CO)2_a and the largest one in Rh(CO)2_b complexes (Table 6). The νCO vibrational frequencies calculated for all three modeled structures are shown in Table 7. The symmetric and asymmetric νCO frequencies of the Rh(CO)2_a and Rh(CO)2_c complexes were calculated to be approximately 2125 and 2057 cm−1, with the difference in the position of the νCO bands for these two structures being only marginal (approximately 2 cm−1). In contrast, the symmetric and asymmetric νCO frequencies of the Rh(CO)2_b complexes are expected to appear at lower wavenumbers (i.e., 2113 and 2050 cm−1, respectively). While DFT calculations predict the formation of three types of zeolite-supported Rh(CO)2 complexes, the spectral fingerprints of the Rh(CO)2_a and Rh(CO)2_c complexes overlap in the νCO region, and it would be impossible to separate these L

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

FTIR spectra. A further comparison of calculated and experimental results suggests that the νCO bands observed in the FTIR spectra at 2117 and 2053 cm−1 can be assigned to the Rh(CO)2_a and Rh(CO)2_c complexes, as the difference between calculated and experimental νCO values is only 4−9 cm−1 in both cases (Table 7). Furthermore, the νCO bands observed in the FTIR spectra at approximately 2113 and 2048 cm−1 can be assigned to the Rh(CO)2_b complexes, as the difference between calculated and experimental νCO values does not exceed 2 cm−1 in this case. Overall, the results of the DFT calculations described above provide strong evidence for the formation of Rh(CO) 2 complexes attached differently to the zeolite framework and exhibiting different stabilities. These theoretical results are consistent with the two types of experimentally observed Rh dicarbonyls exhibiting substantially different thermal properties. 6. Reactivity of Supported Rh(CO)2 Complexes. It has been shown previously that Rh(CO)2(acac) complexes in solution react readily with C2H4 at room temperature to yield Rh(CO)(C2H4)(acac).73 The same facile C2H4−CO substitution reaction has been also documented in the literature for zeolite-supported Rh(CO)2 species.44 Therefore, we have used this reaction as a probe to determine differences in the reactivity of the different types of zeolite-supported Rh(CO)2 complexes. Since the Rh(CO)(C2H4) complexes formed exhibit only one νCO vibration at 2053 cm−1 that overlaps with the asymmetric νCO vibration of the original Rh(CO)2 species,44 changes in intensity of the symmetric νCO bands of the original Rh(CO)2 species can be used to determine the extent of their conversion to Rh(CO)(C2H4). Furthermore, since this substitution reaction is very fast over all Rh(CO)2/ HY samples, the most accurate data can be obtained for the Rh(CO)2/HY2.6 material, which has the largest fraction of type II Rh(CO)2 species formed. Results shown in Figure 12 (curves 1 and 2) for this sample indicate that after 50 s of C2H4 exposure approximately 87 and 50% of type I (band at 2117 cm−1) and type II (band at 2110 cm−1) Rh(CO)2 species, respectively, were converted to Rh(CO)(C2H4). The type I Rh(CO)2 complexes disappeared completely after 300 s of C2H4 exposure, while approximately 16% of the type II complexes still remained intact at that point. These results show that the two types of zeolite-supported Rh(CO)2 complexes are different in terms of their chemical properties, as the carbonyl ligands in the type I Rh(CO)2 species are more reactive toward C2H4. Since the νCO bands characterizing the type II Rh(CO)2 complexes are red-shifted relative to those of the type I species in the FTIR spectra of all samples examined, it is evident that the Rh atoms in the former are more electron rich. This further implies that oxygen atoms in the zeolite framework associated with each type of Rh(CO)2 species are not identical in terms of their electronegativity, which in turn affects the electronic properties of the Rh atoms. Since electron-rich metal sites typically promote CO

Figure 11. Optimized local structures of faujasite-supported (A) Rh(CO)2_a, (B) Rh(CO)2_b, and (C) Rh(CO)2_c complexes. Color coding: Si, yellow; O, red; Al, dark blue; Rh, silver; C, brown.

two complexes from each other in experimental FTIR spectra since the difference in the position of the νCO bands is at or below the resolution of our FTIR measurements. At the same time, the calculated νCO frequencies of the Rh(CO)2_b complexes vary significantly from those of Rh(CO)2_a and Rh(CO)2_c, suggesting that these species can be identified in

Table 6. Binding Energies and Selected Structural Data for Zeolite-Supported Rh(CO)2 Complexes

a

surface structure

BEa (kJ/mol)

O−Ob (pm)

Rh−C (pm)

Rh−O (pm)

C−O (pm)

O−Al−O angle (degrees)

O−Rh−O angle (degrees)

C−Rh−C angle (degrees)

Rh(CO)2_a Rh(CO)2_b Rh(CO)2_c

−514 −474 −487

259 258 258

184, 184 184, 184 184, 184

213, 214 215, 219 214, 217

115.6, 115.6 115.6, 115.7 115.6, 115.7

92.0 92.0 92.5

74.6 73.0 73.6

87.2 90.3 89.5

Binding Energy. bDistance between oxygen atoms interacting with a Rh+ cation. M

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

Table 7. Calculated Vibrational Frequencies for Zeolite-Supported Rh(CO)2 Complexesa surface structure

νs(CO)calc (cm−1)

νas(CO)calc (cm−1)

νs(CO)exp (cm−1)

νas(CO)exp (cm−1)

Δsb (cm−1)

Δasb (cm−1)

Rh(CO)2_a Rh(CO)2_b Rh(CO)2_c

2124 2113 2126

2058 2050 2057

2117 2113 2117

2053 2048 2053

7 0 9

5 2 4

The “s” and “as” subscripts refer to symmetric and asymmetric vibrations, respectively. bDifference between calculated and experimental νCO frequencies (Δ = νcalc − νexp). a

Figure 12. Fractions of the νCO bands remaining in the spectra of Rh(CO)2/HY2.6 (▲, 2110 cm−1; ●, 2117 cm−1) and Rh(CO)2/γAl2O3 (□, 2090 cm−1) samples after exposure to C2H4 as functions of time on stream.

Figure 13. Room-temperature FTIR spectra in the νCO region of (1) Rh(CO)2/HY30 and (2) Rh(CO)2/HY30 pretreated in a 1% NH3/He mixture for 3 min.

framework and influences the electronic properties of Rh indirectly. In any case, the new positions of the νCO bands indicate that the Rh dicarbonyl species became more electron rich following the NH3 treatment. Furthermore, these Rh dicarbonyls were significantly less reactive toward C2H4 than the original Rh(CO)2 complexes (Figure 14). This result is

dissociation but not CO substitution reactions,74,75 we can further infer that variations in electron density of the Rh sites are responsible for the differences observed in their C2H4−CO substitution activities. Similar experiments performed with the Rh(CO)2/γ-Al2O3 sample further support such a hypothesis since the νCO bands of the Rh(CO)2 species were observed in this case at significantly lower frequencies (i.e., 2090 and 2014 cm−1), and their C2H4−CO substitution activity was much lower as well (Figure 12, curve 3). To explore this point further, additional experiments were performed with a Rh(CO)2/HY30 sample pretreated in NH3. NH3 is a strong Lewis base that readily forms complexes with Rh in solution.76,77 Therefore, our intention was to modify the coordination environment of the supported Rh(CO)2 species with NH3 and examine the reactivity of the resulting species with C2H4. Since the fraction of type II Rh(CO)2 species formed in this sample is relatively low (Table 2), no distinction between different types of Rh(CO)2 species was made in this set of measurements. When the Rh(CO)2/HY30 sample was exposed to a 1% NH3/He mixture for 3 min, two significant changes were observed in the νCO region: the νCO bands at 2117 and 2053 cm−1 assigned to Rh(CO)2 species were shifted to 2110 and 2048 cm−1 (Figure 13) and were reduced in intensity. The reduced intensity of these bands could be attributed to either partial decarbonylation of the Rh(CO)2 species or changes in the absorption coefficients of the carbonyl ligands. It is possible that NH3 coordinates directly to the Rh(CO)2 species by replacing the support oxygen atoms at the metal−support interface or alternatively coordinates to Al atoms in the zeolite

Figure 14. Fractions of the νCO bands remaining in the spectra of (1) Rh(CO)2/HY30 (⧫, 2117 cm−1) and (2) Rh(CO)2/HY30 pretreated in a 1% NH3/He mixture (▲, 2110 cm−1) samples after exposure to C2H4 as functions of time on stream. N

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

his assistance during the collection of the EXAFS data, and acknowledge the assistance of the beamline staff at SSRL. The EXAFS data were analyzed with the XDAP software developed by XAFS Services International.18 HAA and GNV acknowledge financial support by the Bulgarian Science Fund (projects DCVP 02/01) and the FP7 program of the European Union (project Beyond Everest), as well as computational resources provided by the Bulgarian Supercomputer Center.

consistent with our previous hypothesis and further confirms that the electronic properties of the Rh atoms in the Rh(CO)2 complexes affect the reactivity of the carbonyl ligands. Furthermore, these results provide strong evidence that the zeolite support acts as a macroligand in the Rh species, inducing significant changes in the reactivity of the supported complexes.



4. CONCLUSIONS The experimental results presented in this manuscript demonstrate the formation of two different types of supported Rh(CO)2 complexes when dealuminated HY zeolites with different Si/Al ratios are used as supports: type I species characterized by νCO bands at 2117 and 2053 cm−1 and type II species characterized by νCO bands at 2110 and 2043 cm−1. The positions of these bands are not affected by the Si/Al ratio of the zeolites used. However, the fraction of each species formed strongly depends on the Si/Al ratio, with more type II species formed on zeolites with lower Si/Al ratios. Our results further show that both types of Rh(CO)2 complexes are located within the zeolite pores and that the Rh atoms in these complexes have a similar coordination environment but slightly different electronic properties since the νCO bands of the type II species appear at lower frequencies. As a result, these two types of zeolite-bound Rh(CO)2 complexes also vary in terms of their chemical properties, with the carbonyl ligands in type I species being more reactive toward C2H4. Experiments conducted with a Rh(CO)2/HY30 sample pretreated in NH3 further reinforce this point and show that the electronic properties of the Rh atoms affect significantly the reactivity of the carbonyl ligands. Our results also indicate that the two types of Rh(CO)2 species formed cannot be linked to unreacted or partially reacted Rh(CO)2(acac) complexes or to the formation of Rh(CO)2(H2O)x species, since both of these complexes have different properties or require quite special conditions for their formation. DFT calculations show that the formation of two types of zeolite-supported Rh(CO)2 complexes can be attributed to differences in the nature of the binding sites in these dealuminated faujasites.



ASSOCIATED CONTENT

S Supporting Information *

Additional Figures 1S−4S. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

(1) Evans, D.; Osborn, J. A.; Wilkinson, G. Hydroformylation of Alkenes by Use of Rhodium Complex Catalysts. J. Chem. Soc. 1968, 3133−3142. (2) Franke, R.; Selent, D.; Börner, A. Applied Hydroformulation. Chem. Rev. 2012, 112, 5675−5732. (3) Harmon, R. E.; Gupta, S. K.; Brown, D. J. Hydrogenation of Organic Compounds Using Homogeneous Catalysts. Chem. Rev. 1973, 73, 21−52. (4) Stakheev, A. Yu.; Kustov, L. M. Effects of the Support on the Morphology and Electronic Properties of Supported Metal Clusters: Modern Concepts and Progress in 1990s. Appl. Catal., A 1999, 188, 3−35. (5) Henrich, V. E. The Surfaces of Metal Oxides. Rep. Prog. Phys. 1985, 48, 1481−1541. (6) Somorjai, G. A.; Park, J. Y. Molecular Factors of Catalytic Selectivity. Angew. Chem., Int. Ed. 2008, 47, 9212−9228. (7) Goellner, J. F.; Gates, B. C.; Vayssilov, G. N.; Rösch, N. Structure and Bonding of a Site-Isolated Transition Metal Complex: Rhodium Dicarbonyl in Highly Dealuminated Zeolite Y. J. Am. Chem. Soc. 2000, 122, 8056−8066. (8) Miessner, H. Surface Chemistry in a Zeolite Matrix. Well-Defined Dinitrogen Complexes of Rhodium Supported on Dealuminated Y Zeolite. J. Am. Chem. Soc. 1994, 116, 11522−11530. (9) Maneck, H. E.; Gutschick, D.; Burkhardt, I.; Luecke, B.; Miessner, H.; Wolf, U. Heterogeneous Carbonylation of Methanol on Rhodium Introduced into Faujasite-Type Zeolites. Catal. Today 1988, 3, 421− 429. (10) Rode, E. J.; Davis, M. E.; Hanson, B. E.; Rösch, N. Propylene Hydroformylation on Rhodium Zeolites X and Y. II. In Situ Fourier Transform Infrared Spectroscopy. J. Catal. 1985, 96, 574−585. (11) Shannon, R. D.; Vedrine, J. C.; Naccache, C.; Lefebvre, F. Rhodium Exchange in Zeolites. J. Catal. 1984, 88, 431−447. (12) Wong, T. T. T.; Stakheev, A. Yu.; Sachtler, W. M. H. Dispersive Oxidation of Rhodium Clusters in Na-Y by the Combined Action of Zeolite Protons and Carbon Monoxide. J. Phys. Chem. 1992, 96, 7733−7740. (13) Miessner, H.; Gutschick, D.; Ewald, H.; Muller, H. The Influence of Support on the Geminal Dicarbonyl Species Rh(CO)2 on Supported Rhodium Catalysts: an IR Spectroscopic Study. J. Mol. Catal. 1986, 36, 359−373. (14) Lefebvre, F.; Gelin, P.; Elleuch, B.; Diab, Y.; Ben Taarit, Y. Metal Carbonyl Compounds Entrapped within Zeolite Cavities. Preparation, Characterization and Catalytic Properties. Stud. Surf. Sci. Catal. 1984, 18, 257−272. (15) Jentoft, R. E.; Deutsch, S. E.; Gates, B. C. Low-Cost, Heated, and/or Cooled Flow-Through Cell for Transmission X-Ray Absorption Spectroscopy. Rev. Sci. Instrum. 1996, 67, 2111−2112. (16) Duivenvoorden, F. B. M.; Koningsberger, D. C.; Uh, Y. S.; Gates, B. C. Structures of Alumina-Supported Osmium Clusters (HOs3(CO)10{OAl}) and complexes (Os(CO)n{OAl}3) (n = 2 or 3) Determined by Extended X-Ray Absorption Fine Structure Spectroscopy. J. Am. Chem. Soc. 1986, 108, 6254−6262. (17) Weber, W. F.; Gates, B. C. Hexarhodium Clusters in NaY Zeolite: Characterization by Infrared and Extended X-Ray Absorption Fine Structure Spectroscopies. J. Phys. Chem. B 1997, 101, 10423− 10434.

AUTHOR INFORMATION

Corresponding Authors

*Tel.: +1 803 7772808. Fax: +1 803 7779502. E-mail: [email protected] (M. D. Amiridis). *Tel.: +1 803 7779914. Fax: +1 803 7778265. E-mail: alexeev@ cec.sc.edu (O. S. Alexeev). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS O.S.A. acknowledges the University of South Carolina for its partial financial support of this work (ASPIRE grant 15510-1229499). XAS data were collected at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences. The authors thank Konstantin Khivantsev at the University of South Carolina for O

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

(41) Miessner, H.; Burkhardt, I.; Gutschick, D.; Zecchina, A.; Morterra, C.; Spoto, G. The Formation of a Well Defined Rhodium Dicarbonyl in Highly Dealuminated Rhodium-Exchanged Zeolite Y by Interaction with Carbon Monoxide. J. Chem. Soc., Faraday Trans. 1989, 85, 2113−2126. (42) Ogino, I.; Gates, B. C. Reactions of Highly Uniform Zeolite Hβ-Supported Rhodium Complexes: Transient Characterization by Infrared and X-ray Absorption Spectroscopies. J. Phys. Chem. C 2010, 114, 8405−8413. (43) Lu, J.; Serna, P.; Gates, B. C. Zeolite- and MgO-Supported Molecular Iridium Complexes: Support and Ligand Effects in Catalysis of Ethene Hydrogenation and H-D Exchange in the Conversion of H2 + D2. ACS Catal. 2011, 1, 1549−1561. (44) Liang, A. J.; Craciun, R.; Chen, M.; Kelly, T. G.; Kletnieks, P. W.; Haw, J. F.; Dixon, D. A.; Gates, B. C. Zeolite-Supported Organorhodium Fragments: Essentially Molecular Surface Chemistry Elucidated with Spectroscopy and Theory. J. Am. Chem. Soc. 2009, 131, 8460−8473. (45) van’t Blik, H. F. J.; van Zon, J. B. A. D.; Huizinga, T.; Vis, J. C.; Koningsberger, D. C.; Prins, R. Structure of Rhodium in an Ultradispersed Rhodium/Alumina Catalyst as Studied by EXAFS and other Techniques. J. Am. Chem. Soc. 1985, 107, 3139−3147. (46) Deutsch, S. E.; Chang, J.-R.; Gates, B. C. Osmium Subcarbonyls on γ-Aalumina: Characterization of the Metal-Support Bonding by Infrared, Nuclear Magnetic Resonance, and X-Ray Absorption Spectroscopies. Langmuir 1993, 9, 1284−1289. (47) Ogino, I.; Chen, C.-Y.; Gates, B. C. Zeolite-Supported Metal Complexes of Rhodium and of Ruthenium: A General Synthesis Method Influenced by Molecular Sieving Effects. Dalton Trans. 2010, 39, 8423−8431. (48) Wang, O.; Sayers, D.; Huang, M.; Yuan, C.; Wei, S. EXAFS Studies of Mo-Y-Zeolite Catalysts, Symposium on catalysis in fuel processing, San-Francisco, 1992. (49) Ji, Y.; van der Eerden, A. M. J.; Koot, V.; Kooyman, P. J.; Meeldijk, J. D.; Weckhuysen, B. M.; Koningsberger, D. C. Influence of Support Ionicity on the Hydrogen Chemisorption of Pt Particles Dispersed in Y Zeolite: Consequences for Pt Particle Size Determination Using the H/M Method. J. Catal. 2005, 234, 376−384. (50) Kawai, M.; Uda, M.; Ichikawa, M. The Electronic State of Supported Rhodium Catalysts and the Selectivity for the Hydrogenation of Carbon Monoxide. J. Phys. Chem. 1985, 89, 1654−1656. (51) Baltanas, M. A.; Onuferko, J. H.; McMillan, S. T.; Katzers, J. R. An Examination of a Rhodium/Magnesia Catalyst Using X-Ray Photoelectron Spectroscopy. J. Phys. Chem. 1987, 91, 3772−3774. (52) Okamoto, Y.; Ishida, N.; Imanaka, T.; Teranishi, S. Active States of Rhodium in Rhodium Exchanged Y Zeolite Catalysts for Hydrogenation of Ethylene and Acetylene and Dimerization of Ethylene Studied with X-Ray Photoelectron Spectroscopy. J. Catal. 1979, 58, 82−94. (53) Okamoto, Y.; Ogawa, M.; Maezawa, A.; Imanaka, T. Electronic Structure of Zeolites Studied by X-Ray Photoelectron Spectroscopy. J. Catal. 1988, 112, 427−436. (54) Shyu, J. Z.; Skopinski, E. T.; Goodwin, J. G.; Sayari, A. Surface Analysis of Dealuminated Y Zeolites by ESCA. Appl. Surf. Sci. 1985, 21, 297−303. (55) Yuan, G.; Chen, Y.; Chen, R. A Novel Copolymer-Bound cisDicarbonylrhodium Complex for the Carbonylation of Methanol to Acetic Acid and Acetic Anhydride. Chin. J. Polym. Sci. 1989, 7, 219− 224. (56) Rode, E. J.; Davis, M. E.; Hanson, B. E. Propylene Hydroformylation on Rhodium Zeolites X and Y. II. In Situ Fourier Transform Infrared Spectroscopy. J. Catal. 1985, 96, 574−585. (57) Trzeciak, A. M.; Ziolkowski, J. J.; Aygen, S.; Van Eldik, R. Reactions of (Acetylacetonato)bis(triphenyl phosphite)rhodium with Dihydrogen, Carbon Monoxide and Olefins. J. Mol. Catal. 1986, 34, 337−343. (58) Otto, S.; Muller, A. J. Cis-dichlorobis(triethylarsine)platinum(II) and Cis-dichlorobis(triethylphosphine)platinum(II). Acta Crystallogr. 2001, 57, 1405−1407.

(18) Vaarkamp, M.; Linders, J. C.; Koningsberger, D. C. A New Method for Parameterization of Phase Shift and Backscattering Amplitude. Physica B 1995, 208−209, 159−160. (19) Stern, E. A. Number of Relevant Independent Points in X-RayAbsorption Fine-Structure Spectra. Phys. Rev. B 1993, 48, 9825−9835. (20) Brigham, E. O. The Fast Fourier Transform; Prentice Hall: Englewood Cliffs, NJ, 1974. (21) Kirlin, P. S.; van Zon, F. B. M.; Koningsberger, D. C.; Gates, B. C. Surface Catalytic Sites Prepared from [HRe(CO) 5] and [H3Re3(CO)12]: Mononuclear, trinuclear, and metallic rhenium catalysts supported on magnesia. J. Phys. Chem. 1990, 94, 8439−8450. (22) van Zon, J. B. A. D.; Koningsberger, D. C.; van’t Blik, H. F. J.; Sayers, D. E. An EXAFS Study of the Structure of the Metal-Support Interface in Highly Dispersed Rhodium/Alumina Catalysts. J. Chem. Phys. 1985, 82, 5742−5754. (23) Vaarkamp, M. Ph. D. Thesis; Eindhoven University, The Netherlands, 1993. (24) Lytle, F. W.; Sayers, D. E.; Stern, E. A. Report of the International Workshop on Standards and Criteria in X-Ray Absorption Spectroscopy. Physica B 1988, 158, 701−722. (25) Koningsberger, D. C.; Mojet, B. L.; van Dorssen, G. E.; Ramaker, D. E. XAFS Spectroscopy; Fundamental Principles and Data Analysis. Top. Catal. 2000, 10, 143−155. (26) Perdew, J. P.; Wang, Y. Accurate and Simple Analytic Representation of the Electron-Gas Correlation Energy. Phys. Rev. B 1992, 45, 13244−13249. (27) Kresse, G.; Hafner, J. Ab Initio Molecular-Dynamics Simulation of the Liquid-Metal-Amorphous-Semiconductor Transition in Germanium. Phys. Rev. B 1994, 49, 14251−14269. (28) Kresse, G.; Furthmüller, J. Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6, 15−50. (29) Vanderbilt, D. Soft Self-Consistent Pseudopotentials in a Generalized Eigenvalue Formalism. Phys. Rev. B 1990, 41, 7892−7895. (30) Kresse, G.; Hafner, J. Norm-Conserving and Ultrasoft Pseudopotentials for First-Row and Transition-Elements. J. Phys.: Condens. Matter 1994, 6, 8245−8257. (31) Jeanvoine, Y.; Angyan, J.; Kresse, G.; Hafner, J. Bronsted Acid Sites in HSAPO-34 and Chabazite: An Ab Initio Structural Study. J. Phys. Chem. B 1998, 102, 5573−5580. (32) Baerlocher, Ch.; McCusker, L. B. Database of Zeolite Structures. http://www.iza-structure.org/databases/ (Accessed 2013). (33) Huq, F.; Skapski, A. C. Refinement of the Crystal Structure of Acetylacetonatodicarbonylrhodium(I). J. Cryst. Mol. Struct. 1974, 4, 411−418. (34) van Veen, J. A. R.; Jonkers, G.; Hesselink, W. H. Interaction of Transition-Metal Acetylacetonates with γ-Alumina Surfaces. J. Chem. Soc., Faraday Trans. 1989, 85, 389−413. (35) Baltes, M.; Collart, O.; Van Der Voort, P.; Vansant, E. F. Synthesis of Supported Transition Metal Oxide Catalysts by the Designed Deposition of Acetylacetonate Complexes. Langmuir 1999, 15, 5841−5845. (36) Van Der Voort, P.; Mitchell, M. B.; Vansant, E. F.; White, M. G. The uses of polynuclear metal complexes to develop designed dispersions of supported metal oxides: Part I. Synthesis and characterization. Interface Sci. 1997, 5, 169−197. (37) Uzun, A.; Bhirud, V. A.; Kletnieks, P. W.; Haw, J. F.; Gates, B. C. A Site-Isolated Iridium Diethylene Complex Supported on Highly Dealuminated Y Zeolite: Synthesis and Characterization. J. Phys. Chem. C 2007, 111, 15064−15073. (38) Ogino, I.; Gates, B. C. Molecular Chemistry in a Zeolite: Genesis of a Zeolite Y-Supported Ruthenium Complex Catalyst. J. Am. Chem. Soc. 2008, 130, 13338−13346. (39) Kenvin, J. C.; White, M. G.; Mitchell, M. B. Preparation and Characterization of Supported Mononuclear Metal Complexes as Model Catalysts. Langmuir 1991, 7, 1198−1205. (40) Guzman, J.; Gates, B. C. Reactions of Au(CH3)2(acac) on γAl2O3: Characterization of the Surface Organic, Organometallic, Metal Oxide, and Metallic Species. Langmuir 2003, 19, 3897−3903. P

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry C

Article

(59) Davis, W. L.; Muller, A. Trans-Di-μ-chlorido-bis{chlorido[tris(3,5-dimethylphenyl)phosphane-κP]palladium(II)} Dichloromethane Monosolvate. Acta Crystallogr. 2012, 68, m1563−m1564. (60) Oswald, A. A.; Hendriksen, D. E.; Kastrup, R. V.; Mozeleski, E. J. Electronic Effects on the Synthesis, Structure, Reactivity, and Selectivity of Rhodium Hydroformylation Catalysts. Adv. Chem. Ser. 1992, 230, 395−418. (61) Gascon, G.; Ortega, M. C.; Suarez, J. D.; Pardey, A. J.; Longo, C. Catalysis of the Water Gas Shift Reaction by Cis-[Rh(CO)2(amine)2](PF6) Complexes in an Aqueous Tetrabutylammonium Hydrogensulfate Solution. React. Kinet. Catal. Lett. 2008, 94, 85−89. (62) Cheng, C.-H.; Hendriksen, D. E.; Eisenberg, R. Homogeneous Catalysis of the Water Gas Shift Reaction Using Rhodium Carbonyl Iodide. J. Am. Chem. Soc. 1977, 99, 2791−2792. (63) Dadci, L.; Elias, H.; Frey, U.; Hornig, A.; Koelle, U.; Merbach, A. E.; Paulust, H.; Schneider, J. S. π-Arene Aqua Complexes of Cobalt, Rhodium, Iridium, and Ruthenium: Preparation, Structure, and Kinetics of Water Exchange and Water Substitution. Inorg. Chem. 1995, 34, 306−315. (64) Kölle, U.; Görissen, R.; Wagner, T. Organometallic Aqua Complexes. Part 3. Olefin Aqua Complexes of Rhodium(I). Chem. Ber. 1995, 128, 911−917. (65) Lincoln, S. F.; Richens, D. T.; Sykes, A. G. In Comprehensive Coordination Chemistry II; McCleverty, J. A., Meyer, T. J., Eds.; Elsevier: Oxford, 2004; Vol. 1, pp 515−555. (66) Gelin, P.; Naccache, C.; Ben Taarit, Y. Coordination Chemistry of Rhodium and Iridium in Constrained Zeolite Cavities: Methanol Carbonylation. Pure Appl. Chem. 1988, 60, 1315−1320. (67) Bergeret, G.; Gallezot, P.; Gelin, P.; Ben Taarit, Y.; Lefebvre, F.; Naccache, C.; Shannon, R. D. Carbon Monoxide-Induced Disintegration of Rhodium Aggregates Supported in Zeolites: In Situ Synthesis of Rhodium Carbonyl Clusters. J. Catal. 1987, 104, 279−287. (68) Smith, A. K.; Huguws, F.; Theolier, A.; Basset, J. M.; Ugo, R.; Zanderighi, G. M.; Bilhou, J. L.; Bilhou-Bougnal, V.; Graydon, W. F. Surface-Supported Metal Cluster Carbonyls. Chemisorption Decomposition and Reactivity of Hexadecacarbonylhexarhodium Supported on Alumina, Silica-Alumina, and Magnesia. Inorg. Chem. 1979, 18, 3104−3112. (69) Robbins, J. L. Rhodium Dicarbonyl Sites on Alumina Surfaces. 1. Preparation and Characterization of a Model System. J. Phys. Chem. 1986, 90, 3381−3386. (70) Ivanova, E.; Hadjiivanov, K. Polycarbonyls of Rh+ Formed After Interaction of CO with Rh-MFI: an FTIR Spectroscopic Study. Phys. Chem. Chem. Phys. 2003, 5, 655−661. (71) Gora-Marek, K. An FTIR Study of the Activation of CO and NO Molecules on Cu+ Sites in CuZSM-5 - The Effect of Coadsorbed Electron Donor Molecules. Vib. Spectrosc. 2012, 58, 104−108. (72) VMD is developed with NIH support by the Theoretical and Computational Biophysics group at the Beckman Institute, University of Illinois at Urbana−Champaign. (73) Moigno, D.; Callejas-Gaspar, B.; Gil-Rubio, J.; Werner, H.; Kiefer, W. The Metal-Carbon Bond in Vinylidene, Carbonyl, Isocyanide and Ethylene Complexes. J. Organomet. Chem. 2002, 661, 181−190. (74) Sette, F.; Stöhr, J.; Kollin, E. B.; Dwyer, D. J.; Gland, J. L.; Robbins, J. L.; Johnson, A. L. Sodium-Induced Bonding and BondLength Changes for Carbon Monoxide on Platinum(111): a NearEdge X-Ray-Absorption Fine-Structure Study. Phys. Rev. Lett. 1985, 54, 935−938. (75) Hocking, R. K.; Hambley, T. W. Database Analysis of Transition Metal Carbonyl Bond Lengths: Insight into the Periodicity of π BackBonding, σ Donation, and the Factors Affecting the Electronic Structure of the TM-C≡O Moiety. Organometallics 2007, 26, 2815− 2823. (76) Khranenko, S. P.; Bykova, E. A.; Alexeyev, A. V.; Tyutyunnik, A. P.; Gromilov, S. A. Complex Salts with Participation of [Rh(NH3)6]3+ Cations. J. Struct. Chem. 2012, 53, 521−526. (77) Evans, R. S.; Hopcus, E. A.; Bordner, J.; Schreiner, A. F. Molecular and Crystal Structures of Halopentaamminerhodium(III)

Complexes. Chloropentaamminerhodium(III)dichloride and Bromopentaamminerhodium(III)dibromide. J. Cryst. Mol. Struct. 1973, 3, 235−245.

Q

dx.doi.org/10.1021/jp507526g | J. Phys. Chem. C XXXX, XXX, XXX−XXX