Strong Metal–Support Interactions Enhance the ... - ACS Publications

Letter pubs.acs.org/acscatalysis

Strong Metal−Support Interactions Enhance the Pairwise Selectivity of Parahydrogen Addition over Ir/TiO2 Evan W. Zhao,† Haibin Zheng,‡ Kaylee Ludden,† Yan Xin,# Helena E. Hagelin-Weaver,*,‡ and Clifford R. Bowers*,† †

Department of Chemistry and ‡Department of Chemical Engineering, University of Florida, Gainesville, Florida 32611, United States # National High Magnetic Field Laboratory, Florida State University, Tallahassee, Florida 32310, United States S Supporting Information *

ABSTRACT: The effects of strong metal−support interactions (SMSI) on the pairwise selectivity of propene hydrogenation over metal-oxide-supported Ir nanoparticles were investigated using parahydrogen-enhanced NMR spectroscopy. A ∼20-fold increase in the pairwise selectivity was observed following a reduction treatment of the Ir/TiO2 catalyst at 500 °C. Consistent with SMSI, the effects could be completely reversed by oxidation followed by rereduction at 200 °C. Noninteracting supports, such as Al2O3 and SiO2, did not show this behavior. X-ray photoelectron spectroscopy reveals partial reduction of the TiO2 support, and STEM data reveal flattening of Ir particles after high-temperature reduction. The presence of chloride ions during activation was found to further promote pairwise selectivity but only for the Ir/TiO2 catalyst. The results are interpreted in terms of the electronic and possible geometric blocking effects associated with SMSI. KEYWORDS: heterogeneous catalysis, iridium nanoparticles, hydrogenation, NMR spectroscopy, parahydrogen induced polarization

P

approach of the metal atom toward the reduced cation of the support, facilitating electron transfer from the d-orbitals of the reduced cation to the metal. For TiO2, SMSI can promote migration of TiOx over the surface of the metal, resulting in encapsulation and a geometrical blocking effect for adsorbates28,29,31 which can significantly affect the dissociative chemisorption of H2, thereby reducing the activity of hydrogenation catalysis.32−36 Hence, SMSI might also be expected to have an effect on the pairwise selectivity in hydrogenation reactions. Indeed, a recent report demonstrated the effect of H2 reduction of Pd/TiO2 at 500 °C on the pairwise selectivity of 1,3-butadiene hydrogenation over this catalyst.37 Unfortunately, both the overall activity and the pairwise selectivity were reduced with the production of a Pdδ+ charge state. Here we examine the role of SMSI on the pairwise selective addition of parahydrogen to propene over Ir nanoparticles supported on three different metal oxides: TiO2, a reducible support that is well-known to induce SMSI, and the nonreducible oxides SiO2 and Al2O3. The supported Ir nanoparticle catalysts were synthesized by two different protocols, as described in the Supporting Information (SI). The deposition−precipitation method using

arahydrogen-induced polarization by heterogeneous catalysis (hetPHIP) is a promising method for continuous production of hyperpolarized fluids for NMR spectroscopy and imaging with potential biomedical applications.1−3 Ease of separation of the reaction product from the catalyst in hetPHIP is a key advantage over PHIP by homogeneous catalysis.3−12 However, hetPHIP yields comparatively low NMR signal enhancement factors due to a low pairwise selectivity of parahydrogen transfer.13−15 Loss of singlet proton spin-order of parahydrogen occurs due to rapid diffusion of H atoms formed by dissociative chemisorption on the metal surface.16 Despite substantial effort to increase the pairwise selectivity through optimization of reaction conditions17−24 and testing of novel catalysts,13,25 the pairwise addition over supported noble metals has been limited to only a few percent of the total conversion. In the present work, the effects of strong metal−support interactions26−31 (SMSI) on the pairwise selectivity of propene hydrogenation over metal-oxide-supported Ir nanoparticles is investigated using parahydrogen-enhanced NMR spectroscopy. We demonstrate that H2 reduction pretreatment of TiO2 supported Ir at 500 °C can yield up to a ∼20-fold increase in pairwise selectivity. The results are interpreted in terms of the electronic and possible geometric blocking effects associated with SMSI. In SMSI, a term introduced by Tauster,26 cations of the support are reduced in the presence of H2 at elevated temperatures. The concomitant loss of oxygen anions facilitates © XXXX American Chemical Society

Received: November 20, 2015 Revised: December 28, 2015

974

DOI: 10.1021/acscatal.5b02632 ACS Catal. 2016, 6, 974−978

Letter

ACS Catalysis NaOH and an Ir precursor was used to prepare catalysts with low chlorine content on TiO2 and Al2O3 supports. To investigate the effect of Cl− ions, a modified impregnation method was used in which a chloride containing Ir precursor was impregnated onto TiO2 and SiO2 supports (Ir/TiO2−Cl and Ir/SiO2−Cl). The nominal loading of Ir metal was 0.2 atom % (0.5 wt %) on all of the supports. The actual Ir loading was quantified for the Ir/TiO2 catalyst by energy-dispersive Xray spectroscopy (EDS) and was measured to be 0.2−0.3 atom % (see Figure S4, SI). Prior to running hydrogenation reactions, the supported Ir catalysts were oxidized in air at 400 °C and then reduced in pure H2 at temperatures ranging from 200 to 500 °C. Scanning transmission electron microscopy high-angle annular dark field (STEM-HAADF) images of the Ir/TiO2 sample were collected after the freshly synthesized catalyst was reduced at 200 °C (Figure 1a,c and Figures S1a,c in the SI) or

Figure 2. ALTADENA (bottom) and normal (top) 1H NMR spectra of PE hydrogenation products obtained using Ir/TiO2 after reduction in H2 at (a) 200 °C, (b) 500 °C, and (c) oxidation in air at 400 °C followed by reduction in H2 at 200 °C. Sixteen transients were accumulated for all spectra except for the normal spectrum in B (64 transients). Flow rate: 120 mL/min H2; 20 mL/min PE; 160 mL/min N2. The vertical scale is the same for all spectra.

The hydrogenation products formed at low magnetic field (ca. 5 mT) were “adiabatically” transported to 9.4 T for NMR detection. At both reduction temperatures, propane (PA) exhibits enhanced NMR signals with characteristic ALTADENA intensity pattern. Note the emission phase (downward) peaks of the PA−CH3 in the ALTEDENA spectra due to the effect of population inversion across the NMR transition.14 These spectra can be compared to the thermally polarized spectra acquired using n-H2 rather than p-H2 in Figure 2 (top). The thermally polarized PA signals are proportional to total conversion resulting from pairwise + nonpairwise addition. Although the high-temperature (500 °C) H2 pretreatment decreased the total conversion to PA by a factor of 44, the PHIP signal decreased by only a factor of 2.2. Because the PHIP NMR signal enhancement is the ratio between the PHIP and thermally polarized signals (see SI), a 20-fold boost of the PA signal enhancement (and pairwise selectivity) was induced. This is opposite to the trend reported for Pd/TiO2,37 where reduction treatment at 500 °C had the effect of decreasing the pairwise selectivity. A hallmark of SMSI is the reversibility of the effects. Changes in the activity and pairwise selectivity due to particle sintering/ aggregation, on the other hand, are expected to be irreversible. In the original studies demonstrating SMSI on metal/TiO2 systems, the SMSI were shown to be fully or partially reversible by oxidation treatment followed by mild reduction.26 Therefore, we hypothesize that if the observed changes in the pairwise selectivity are truly related to SMSI, they too should be reversible. The spectra in Figure 2c were obtained by recycling the catalyst specimen by reoxidation treatment in air at 400 °C followed by reduction at 200 °C in pure H2. Indeed, both the pairwise selectivity and the overall activity have been restored to nearly the same values as the initial ones (compare Figure 2a,c), consistent with SMSI. The pairwise selectivity increased from 0.2% to 3.9% after reduction at 500 °C and then decreased back to 0.5% after reoxidation and then rereduction at 200 °C, whereas overall conversion decreased from 21.9% to 0.5% and then increased back to 17.7%. The slightly lower conversion can be attributed to minor aggregation and sintering which reduces the surface area of the metal, consistent with

Figure 1. High-resolution STEM-HAADF images at two magnifications of the Ir/TiO2 catalyst (precipitation−deposition synthesis), oxidized at 400 °C, then reduced in H2 at 200 °C (a,c) or 500 °C (b,d). Flat-shaped Ir particles are indicated by the red arrows.

500 °C (Figure 1b,d and Figure S1b,d). Minor agglomeration of smaller clusters and isolated Ir atoms seems to have occurred after reduction at 500 °C, but since this is well below the Tammann temperature of Ir (1110 °C),38 sintering or agglomeration should be minimal. However, the metal nanoparticles appear to be flatter (evidence of SMSI), and in many cases the Ir atoms appear to adopt the lattice of the underlying TiO2 (see Figure 1d). The measured particle size distributions indicate slightly larger average diameter for the higher-temperature reduction (0.7 ± 0.2 nm vs 0.5 ± 0.2 nm, see Figure S3). No trace of a thin amorphous TiOx layer on top of the Ir nanoparticle can be seen in our STEM annular brightfield (STEM-ABF) images (Figure S5) in which there is neither delocalization effect nor Fresnel contrast, as in the conventional high-resolution TEM images. It is claimed that such a layer can be seen in ref 31.39 Figure 2a,b present the 400 MHz proton NMR spectra of the reactor effluent, acquired in the ALTADENA-PHIP mode (bottom spectra), following reductive pretreatment of the Ir/ TiO2 catalyst in pure H2 at 200 °C and 500 °C, respectively. 975

DOI: 10.1021/acscatal.5b02632 ACS Catal. 2016, 6, 974−978

Letter

ACS Catalysis slightly larger average particle size observed after the 500 °C reduction. A similar recovery was observed for Ir/TiO2−Cl (see Table 1) except for a slight increase of overall conversion, possibly due to loss of Cl− after the oxidation−reduction cycle.40 Table 1. Percent Conversion (Conv) and Pairwise Selectivity (PS) of Ir Nanoparticles on Various Supports reduction temperature support material TiO2−Cl TiO2 Al2O3 SiO2−Cl a

conversion or pairwise selectivity

200 °C

500 °C

200 °Ca

Conv PS Conv PS Conv PS Conv PS

26.2 ± 1.52 0.4 ± 0.01 21.9 ± 1.25 0.2 ± 0.01 20.1 ± 0.57 0.2 ± 0.01 19.9 ± 1.60 0.1 ± 0.01

1.9 ± 0.17 5.3 ± 0.22 0.5 ± 0.04 3.9 ± 0.15 37.2 ± 1.17 0.8 ± 0.01 14.9 ± 1.78 0.3 ± 0.01

33.2 ± 5.69 0.3 ± 0.02 17.7 ± 1.0 0.4 ± 0.01 40.2 ± 2.56 0.3 ± 0.01 17.4 ± 1.75 0.1 ± 0.01

Figure 3. (a) Signal enhancement factor and pairwise selectivity of PE hydrogenation as a function of reduction temperature; (b) Total conversion vs reduction temperature. Black solid circles: Ir/TiO2−Cl. Hollow black circles: Ir/TiO2. Blue squares: Ir/Al2O3. Green triangles: Ir/SiO2−Cl.

Reoxidation and rereduction.

Since SMSI can be induced only on a reducible transition metal oxide, control experiments were performed using Ir deposited onto two nonreducible inert supports, SiO2 and Al2O3.27 The NMR spectra are presented in Figure S9 of the SI, and the results are summarized in Table 1. For Ir nanoparticles on these supports, increasing the reduction temperature from 200 °C to 500 °C resulted in the following changes in the pairwise selectivity: 0.2% to 0.8% for Ir/Al2O3 and 0.1% to 0.3% for Ir/SiO2. The total conversion increased from 20.1% to 37.2% over Ir/Al2O3 and decreased slightly for Ir/SiO2. The decrease in conversion over the Ir/SiO2 could be due to sintering of the Ir particles at the high reduction temperature, although it is also possible that the chloride ions in this sample migrated to the surface during the reductive treatment and blocked some of the active sites. After oxidation and another reduction treatment at 200 °C, some of this chlorine has likely been removed. However, compared to the behavior of the Ir/ TiO2 catalyst, the changes in activity and pairwise selectivity are much smaller and consistent with a noninteracting support. The behavior of the Ir/Al2O3 catalyst is different from both the SiO2 and TiO2 supported catalysts. The high-temperature reduction significantly increased the conversion, but unlike the Ir/TiO2 results, the subsequent low-temperature reduction of the Ir/Al2O3 catalyst did not reverse the increased activity (within the experimental uncertainty). The increase in activity with reduction temperature suggests that a significant fraction of the IrOx on the surface of this catalyst is not reduced at 200 °C. Because the “strength” of SMSI depends on reduction temperature, a series of reduction pretreatments were performed at temperatures ranging from 200 °C to 500 °C. After each pretreatment, hydrogenation reactions were performed at 150 °C, a temperature lower than the lowest reduction temperature, to ensure no further SMSI effect is induced during the hydrogenation reaction. After catalyst reduction, NMR spectra of the product gas stream were acquired using a reactant mixture containing propene and 50% para-enriched H2. Figure 3 presents the PHIP signal enhancement, pairwise selectivity, and total conversion as a function of reduction temperature for all of the catalysts. For hydro-

genation over the Ir/TiO2 catalyst, the pairwise selectivity monotonically increased from 0.2% to 3.9% (a 20-fold increase!) upon increasing the reduction temperature from 200 to 500 °C. For Ir/TiO2−Cl, a sharp increase from 0.5% to 4.4% was observed after reduction at 350 °C, which further increased to 5.3% after reduction at 500 °C. Overall conversion (catalytic activity) exhibited the opposite trend, decreasing sharply from 21.9% to 0.5% over Ir/TiO2 and from 26.2% to 1.9% over Ir/TiO2−Cl (Figure 3B). These results further demonstrate that the high-temperature H2-reduction treatment is beneficial for increasing the pairwise selectivity of propene hydrogenation over Ir/TiO2. The effect on the pairwise selectivity is further enhanced by the presence of chloride. The behavior of the activity with reduction temperature, i.e. the decrease in propene conversion after reduction at 500 °C and recovery of activity with an oxidation and low temperature reduction treatment, is consistent with the expected behavior of SMSI.26,27 In an attempt to prove SMSI effects, the catalysts were subjected to H2 and CO chemisorption analysis after reduction at 200 °C and 500 °C. Due to the very low metal loading (below 0.2 atom %), the typical decrease in CO chemisorption for SMSI effects was not observed. Instead, a small increase in the amount of CO adsorbed was measured for the Ir/TiO2 catalyst (from 11 to 17 μL/g) after reduction at 500 °C, while the H2 chemisorption was too small to be detected. Hydrogen temperature-programmed reduction (TPR) analysis of the Ir/TiO2 catalyst and the TiO2 support without Ir (see Figure 4A) revealed that the increase in CO adsorption is due to a higher degree of reduction at 500 °C, consistent with previous studies.41 The reduction peak at around 135 °C for Ir/TiO2 is most likely due to reduction of IrOx to Ir metal, while the reduction peaks between 300 and 400 °C are likely due to reduction of IrOx that interacts strongly with the TiO2 support and/or reduction of TiO2 near Ir particles. The TiO2 support itself, when no Ir is present, is reduced at temperatures in excess of 400 °C. TPR measurements on the Ir/Al2O3 catalyst reveal that at low Ir loading, the IrOx on the surface of the support is difficult to reduce (Figure 4B). Therefore, at a reduction temperature of 976

DOI: 10.1021/acscatal.5b02632 ACS Catal. 2016, 6, 974−978

Letter

ACS Catalysis

or result in a spreading of Ir on the support to maximize IrTiOx interactions (geometrical effect of SMSI). The charge transfer increases the number of metallic and possibly negatively charged Ir atoms at the interface.27 Based on all of our observations and the literature, there are several plausible models that can account for the correlation between H2 reduction temperature and pairwise selectivity. (1). Geometric Effect. Loss of hydrogenation activity could be due to a geometric blocking effect resulting from migration of TiOx across the metal surface. Even though this layer is not visible by STEM, the existence of one or two monolayers, which would be difficult to image, cannot be ruled out. In this model, the active metal surface area, and hence the number of metal atoms available to split molecular H2 in nonpairwise addition by the Horiuti−Polanyi (HP) mechanism, would vary as ∼ r2, where r, the radius of exposed Ir0 metal, decreases with increasing H2 reduction temperature. If the active sites for pairwise addition reside along the Ir−TiOx interface on the periphery of the exposed particle surface, the number of such sites would vary as ∼ r. Hence, the relative increase in active sites for pairwise addition due to migration of TiOx would vary as ∼1/r. (2). Electronic Effect. According to Tauster,27 Ti3+ → Ir charge transfer results in polar-covalent bond formation that can suppress dissociative chemisorption of H2 on the metal, implying a short lifetime on the surface. While this would tend to decrease the total conversion, the weakened adsorbate− surface interaction may favor pairwise transfer involving direct/ concerted addition of surface-bound molecular H2 to the substrate via a four membered ring transition state. In this purely electronic SMSI model, geometric blocking is not essential to explain the results. To summarize, we find that SMSI in the Ir/TiO2 system produce a dramatic boost of the PHIP signal enhancement factor, in contrast to the negative effect of SMSI reported recently for Pd/TiO2.37 This is brought about by a disproportionate decrease in the nonpairwise addition relative to pairwise addition. The contrasting results for Pd and Ir suggest that the effect of SMSI on pairwise selectivity depends on the specific catalyst system due to different electronic and/or geometric SMSI effects. The ambiguity in the oxidation state determination for our low-loading Ir/TiO2 catalyst prohibits a more specific conclusion. Moreover, a broader survey of other supported metals is necessary to establish a general pattern on the effect of SMSI on pairwise selectivity. Furthermore, Cl− ions were found to promote pairwise selectivity. Chlorine on the surface is expected to block Ir sites, which could reduce the diffusion of hydrogen on the surface. The effect of chlorine may also be electronic. Direct addition of molecular H2 to the substrate in a concerted process would preserve the singlet spin order inherent to para-H2. It has been suggested that H2 binds to low coordination metal atoms located at edges or kink atoms on the metal cluster.45 The demonstration of a substantial positive effect of SMSI over Ir/TiO2 can be explained if such sites occur at the interface between the metal and oxide support. Alternatively, the weakening of the adsorbate−surface interaction by charge transfer and ionic bond formation could serve this role. Disentangling these effects, and consideration of possible kinetic/diffusion effects on the pairwise selectivity, remains a formidable challenge for future studies, which is relevant not only to the fundamental understanding of SMSI in catalytic processes but also to engineering hydrogenation

Figure 4. Hydrogen TPR of (a) bare TiO2 support and Ir/TiO2 catalyst. (b) Freshly prepared and recycled (after 500 °C reduction) Ir/Al2O3.

200 °C, only a small fraction of the IrOx on the Al2O3 support is reduced, which also explains why no CO adsorption was detected on this catalyst. Higher reduction temperatures are needed to reduce the IrOx on the Al2O3 support, which explains the increase in conversion with reduction temperature. In contrast to the Ir/TiO2 catalyst, where the effects of the 500 °C reduction were reversible, the Ir/Al2O3 catalyst has been permanently altered by the high-temperature reduction. After oxidation following reduction at 500 °C, the appearance of the TPR curve of the “re-cycled” Ir/Al2O3 looks very different from the original TPR curve. Most of the Ir on the surface is now reduced at a lower temperature, consistent with a larger particle size due to sintering after the high-temperature reduction. This was not observed for the Ir/TiO2 catalyst, where all effects were reversible. The Ir/TiO2 catalyst was analyzed by X-ray photoelectron spectroscopy (XPS) to probe changes in oxidation state associated with SMSI.36,37,42 XPS was performed on three Ir/ TiO2 samples following the same pretreatments as in the PHIP experiments: oxidation at 400 °C only and oxidation at 400 °C followed by reduction at 200 °C or 500 °C. Unfortunately, the Ir 4f peaks (the most intense Ir peaks) overlap with the Ti 3s peak of the support. This, together with the very low Ir loading, make it difficult to determine the oxidation state of Ir. Nevertheless, peak fitting using the PHI MultiPak43 software (see SI) suggests that the relative amount of Ir0 (or Ir+) increases as reduction temperature increases, in accord with the published in situ XPS characterization of Ir/TiO2 catalysts with a higher Ir loading33 and consistent with the TPR data. Furthermore, the TiO2 support is partially reduced at a reduction temperature of 500 °C. This is evident as a slight broadening of the low-energy side of the Ti 2p3/2 peak obtained from the Ir/TiO2 catalyst reduced at 500 °C compared to the catalyst reduced at 200 °C (Figure S7). The broadening is consistent with reduction of some Ti4+ ions to Ti3+.44 In SMSI, the removal of oxygen anions is needed to allow the metal atom and the surface cation of the support to be close enough for charge transfer and bonding to occur, and this can promote the migration of TiOx over the Ir nanoparticle surface 977

DOI: 10.1021/acscatal.5b02632 ACS Catal. 2016, 6, 974−978

Letter

ACS Catalysis

(14) Zhou, R.; Zhao, E. W.; Cheng, W.; Neal, L. M.; Zheng, H.; Quiñones, R. E.; Hagelin-Weaver, H. E.; Bowers, C. R. J. Am. Chem. Soc. 2015, 137, 1938−1946. (15) Reineri, F.; Boi, T.; Aime, S. Nat. Commun. 2015, 6, 5858. (16) Horiuti, I.; Polanyi, M. Trans. Faraday Soc. 1934, 30, 1164− 1172. (17) Barskiy, D. A.; Kovtunov, K. V.; Primo, A.; Corma, A.; Kaptein, R.; Koptyug, I. V. ChemCatChem 2012, 4, 2031−2035. (18) Salnikov, O.; Kovtunov, K.; Barskiy, D.; Bukhtiyarov, V.; Kaptein, R.; Koptyug, I. Appl. Magn. Reson. 2013, 44, 279−288. (19) Salnikov, O.; Barskiy, D.; Burueva, D.; Gulyaeva, Y.; Balzhinimaev, B.; Kovtunov, K.; Koptyug, I. Appl. Magn. Reson. 2014, 45, 1051−1061. (20) Kovtunov, K. V.; Barskiy, D. A.; Salnikov, O. G.; Khudorozhkov, A. K.; Bukhtiyarov, V. I.; Prosvirin, I. P.; Koptyug, I. V. Chem. Commun. 2014, 50, 875−878. (21) Balu, A. M.; Duckett, S. B.; Luque, R. Dalton T 2009, 26, 5074− 5076. (22) Glöggler, S.; Grunfeld, A. M.; Ertas, Y. N.; McCormick, J.; Wagner, S.; Schleker, P. P. M.; Bouchard, L.-S. Angew. Chem., Int. Ed. 2015, 54, 2452−2456. (23) Sharma, R.; Bouchard, L.-S. Sci. Rep. 2012, 2, 277. (24) Corma, A.; Salnikov, O. G.; Barskiy, D. A.; Kovtunov, K. V.; Koptyug, I. V. Chem. - Eur. J. 2015, 21, 7012−7015. (25) Zhao, E. W.; Zheng, H.; Zhou, R.; Hagelin-Weaver, H. E.; Bowers, C. R. Angew. Chem., Int. Ed. 2015, 54, 14270−14275. (26) Tauster, S. J.; Fung, S. C.; Garten, R. L. J. Am. Chem. Soc. 1978, 100, 170−175. (27) Tauster, S. J.; Fung, S. C.; Baker, R. T. K.; Horsley, J. A. Science 1981, 211, 1121−1125. (28) Belton, D. N.; Sun, Y. M.; White, J. M. J. Am. Chem. Soc. 1984, 106, 3059−3060. (29) Pesty, F.; Steinrück, H.-P.; Madey, T. E. Surf. Sci. 1995, 339, 83−95. (30) Tauster, S. J. Acc. Chem. Res. 1987, 20, 389−394. (31) Hernández-Cristóbal, O.; Arenas-Alatorre, J.; Díaz, G.; Bahena, D.; Yacamán, M. J. J. Phys. Chem. C 2015, 119, 11672−11678. (32) Liu, J. ChemCatChem 2011, 3, 934−948. (33) Reyes, P.; Salinas, D.; Campos, C.; Oportus, M.; Murcia, J.; Rojas, H.; Borda, G.; Fierro, J. L. G. Quim. Nova 2010, 33, 777−780. (34) Rojas, H.; Borda, G.; Reyes, P.; Martinez, J. J.; Valencia, J.; Fierro, J. L. G. Catal. Today 2008, 133-135, 699−705. (35) Mériaudeau, P.; Ellestad, O. H.; Dufaux, M.; Naccache, C. J. Catal. 1982, 75, 243−250. (36) Haller, G. L.; Resasco, D. E. Adv. Catal. 1989, 36, 173−235. (37) Kovtunov, K. V.; Barskiy, D. A.; Salnikov, O. G.; Burueva, D. B.; Khudorozhkov, A. K.; Bukhtiyarov, A. V.; Prosvirin, I. P.; Gerasimov, E. Y.; Bukhtiyarov, V. I.; Koptyug, I. V. ChemCatChem 2015, 7, 2581− 2584. (38) Baker, R. T. K.; Sherwood, R. D. J. Catal. 1980, 61, 378−389. (39) The image is under-focused, and the dark outline is probably Fresnel fringes. Second, the HRTEM image from field emission TEM has the issue of delocalization, where there is contrast outside of the sample in the vacuum close to the edge of the sample, and this can be mistakenly recognized as some sample contrast. (40) Orita, H.; Naito, S.; Tamaru, K. J. Phys. Chem. 1985, 89, 3066− 3069. (41) Hernandez-Cristobal, O.; Diaz, G.; Gomez-Cortes, A. Ind. Eng. Chem. Res. 2014, 53, 10097−10104. (42) Park, C.; Baker, R. T. K. J. Phys. Chem. B 2000, 104, 4418−4424. (43) Product information for PHI VersaProbe II, Scanning XPS Microprobe. https://www.phi.com/surface-analysis-equipment/ versaprobe.html. (44) Conesa, J. C.; Soria, J. J. Phys. Chem. 1982, 86, 1392−1395. (45) Zhivonitko, V. V.; Kovtunov, K. V.; Beck, I. E.; Ayupov, A. B.; Bukhtiyarov, V. I.; Koptyug, I. V. J. Phys. Chem. C 2011, 115, 13386− 13391.

catalysts with significantly higher pairwise selectivity. The increased efficiency of production of hyperpolarized gases and liquids will be of great utility for biomedical magnetic resonance imaging of the lung and other tissues.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.5b02632. Details of the catalyst synthesis, procedures, calculations, XPS, supplemental XPS and NMR spectra are provided in the Supporting Information (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]fl.edu. *E-mail: [email protected]fl.edu. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank W. Cheng for catalyst synthesis, S. McCrea for assistance with the TPR, and E. Lambers for assistance with the XPS characterization. This work was supported by startup funds from the University of Florida (H.E.H.-W) and NSF grant CHE-1507230. Part of the work was performed at the TEM facility at the National High Magnetic Field Laboratory, which is supported by NSF Cooperative Agreement DMR1157490 and the State of Florida.

■

REFERENCES

(1) Kovtunov, K.; Zhivonitko, V.; Skovpin, I.; Barskiy, D.; Koptyug, I. In Hyperpolarization Methods in NMR Spectroscopy; Kuhn, L. T., Ed.; Springer: Berlin Heidelberg, 2013; Vol. 338, pp 123−180. (2) Kovtunov, K. V.; Truong, M. L.; Barskiy, D. A.; Koptyug, I. V.; Coffey, A. M.; Waddell, K. W.; Chekmenev, E. Y. Chem. - Eur. J. 2014, 20, 14629−14632. (3) Chekmenev, E. Y.; Hövener, J.; Norton, V. A.; Harris, K.; Batchelder, L. S.; Bhattacharya, P.; Ross, B. D.; Weitekamp, D. P. J. Am. Chem. Soc. 2008, 130, 4212−4213. (4) Bowers, C. R.; Weitekamp, D. P. Phys. Rev. Lett. 1986, 57, 2645− 2648. (5) Bowers, C. R.; Weitekamp, D. P. J. Am. Chem. Soc. 1987, 109, 5541−5542. (6) Natterer, J.; Bargon, J. Prog. Nucl. Magn. Reson. Spectrosc. 1997, 31, 293−315. (7) Bowers, C. R. In Encyclopedia of Nuclear Magnetic Resonance: Supplementary Vol.; Grant, D. M., Harris, R. K., Eds.; John Wiley and Sons, Ltd.: Chichester, West Sussex, England, 2002; Vol. 9, pp 750− 770. (8) Kovtunov, K. V.; Beck, I. E.; Bukhtiyarov, V. I.; Koptyug, I. V. Angew. Chem., Int. Ed. 2008, 47, 1492−1495. (9) Adams, R. W.; Aguilar, J. A.; Atkinson, K. D.; Cowley, M. J.; Elliott, P. I. P.; Duckett, S. B.; Green, G. G. R.; Khazal, I. G.; LópezSerrano, J.; Williamson, D. C. Science 2009, 323, 1708−1711. (10) Kovtunov, K. V.; Zhivonitko, V. V.; Skovpin, I. V.; Barskiy, D. A.; Salnikov, O. G.; Koptyug, I. V. J. Phys. Chem. C 2013, 117, 22887− 22893. (11) Shi, F.; Coffey, A. M.; Waddell, K. W.; Chekmenev, E. Y.; Goodson, B. M. Angew. Chem., Int. Ed. 2014, 53, 7495−7498. (12) Shi, F.; Coffey, A. M.; Waddell, K. W.; Chekmenev, E. Y.; Goodson, B. M. J. Phys. Chem. C 2015, 119, 7525−7533. (13) Barskiy, D. A.; Salnikov, O. G.; Kovtunov, K. V.; Koptyug, I. V. J. Phys. Chem. A 2015, 119, 996−1006. 978

DOI: 10.1021/acscatal.5b02632 ACS Catal. 2016, 6, 974−978