Catalytic Oxidation of Methanol to Formaldehyde by Mass-Selected

Apr 17, 2014 - Catalytic Oxidation of Methanol to Formaldehyde by Mass-Selected Vanadium Oxide Clusters Supported on a TiO2(110) Surface. Scott P. Pri...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/JPCA

Catalytic Oxidation of Methanol to Formaldehyde by Mass-Selected Vanadium Oxide Clusters Supported on a TiO2(110) Surface Scott P. Price, Xiao Tong, Claron Ridge, Hunter L. Neilson, Joshua W. Buffon, Jeremy Robins, Horia Metiu, Michael T. Bowers, and Steven K. Buratto* Department of Chemistry and Biochemistry, University of California, Santa Barbara, California 93106-9510, United States S Supporting Information *

ABSTRACT: We report the results of a systematic study of the catalytic activity of massselected vanadium oxide clusters deposited on rutile TiO2 surfaces under ultrahigh vacuum (UHV) conditions. Our results show that supported V, VO, and VO2 clusters are not catalytically active for the oxidative dehydrogenation of methanol to formaldehyde but can be made catalytically active by postoxidation. In addition, we found that the postoxidized VO/TiO2 produces the most formaldehyde. Scanning tunneling microscopy (STM) imaging of the postoxidized VO/TiO2 reveals isolated clusters with height and width indicative of VO3 bound to the TiO2 surface. Our results are consistent with previous density functional theory (DFT) calculations that predict that VO3 will be produced by postoxidation of VO and that VO3/TiO2 is an active catalyst.

■

■

INTRODUCTION Titania-supported monolayers or submonolayers of vanadium oxide are highly versatile catalysts that are useful in a wide range of reactions, including the oxidation of SO2 to SO3, selective oxidation of organic compounds, and reduction of NOx (in the presence of NH3).1−3 One of the simplest of these reactions is the catalytic oxidation of methanol to produce formaldehyde and water. For this reason, a significant effort has focused on studying this reaction on powder4 and single-crystalline5−11 supports. From these experiments and theoretical investigations,12−15 it has been suggested that the O−H bond in methanol dissociates upon adsorption to the cluster, forming a methoxide (presumably bound to the V atom) and a hydroxyl (resulting from the H atom binding to an oxygen atom in vanadia). Scission of a C−H bond in the methoxide results in formation of formaldehyde and water. The vanadia catalyst can be regenerated by heating in oxygen to reoxidize the cluster.9 Due to the composition dependence on the catalytic activity of these vanadia clusters, knowledge of the correct vanadia species responsible for methanol oxidation is critical in understanding and explaining the catalytic properties of VOx/TiO2. Unfortunately, as a result of the wide variety of cluster sizes and vanadium oxidation states that have been observed in these studies, the exact composition of catalytically active vanadia remains unknown. Here, we report the results of a systematic study into the catalytic activity of composition-selected vanadium oxide clusters (V, VO, and VO2) deposited on a rutile TiO2(110) surface. We reacted both the vanadia-decorated surfaces and postoxidized versions of these surfaces with methanol and observed whether or not formaldehyde is formed. STM was used to characterize the potential active site(s). © XXXX American Chemical Society

EXPERIMENTAL SECTION Single-crystal TiO2(110) samples were purchased from Commercial Crystal Laboratories, and atomically flat surfaces were prepared by several cycles of argon ion bombardment (1 kV, 20 min) and annealing (970 K, 11 min) in an ultrahigh vacuum (UHV) chamber with a base pressure below 1 × 10−10 Torr. Samples were checked for cleanliness using STM and then heated to 1000K for 10 s immediately before depositing clusters to remove surface hydroxyls. Mass-selected vanadium and vanadium oxide clusters were deposited onto these surfaces using a home-built apparatus that has been described previously.16,17 Briefly, a rotating, translating vanadium rod is exposed to the beam from a frequencydoubled, pulsed yttrium aluminum garnet (YAG) laser, resulting in ablation of the vanadium. Vanadium or vanadium oxide clusters are generated by introducing a pulse of pure argon or a 4/1 mixture of Ar/O2, respectively, into the ion source chamber. Positive ions are extracted, and ion optics are used to focus and steer the resulting ion beam. Mass selection is achieved using a magnetic-field analyzer prior to deposition on the TiO2(110) substrate, which is housed in an UHV chamber with a base pressure of ∼3 × 10−10 Torr. A typical mass spectrum of the vanadia species generated from our cluster source is shown in Figure 1. It should be noted that the minimum ion current for deposition of 0.01 ML in 4 h is 0.1 nA. Special Issue: A. W. Castleman, Jr. Festschrift Received: January 31, 2014 Revised: April 16, 2014

A

dx.doi.org/10.1021/jp5011378 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

= 18; however, we were unable to observe H2O desorption due to the high background pressure of water in our apparatus.

■

RESULTS AND DISCUSSION Formaldehyde TPR spectra recorded from bare (black line), Vdecorated (orange line), VO-decorated (blue line), and VO2decorated (red line) vacuum-annealed TiO2(110) surfaces are shown in Figure 2a. None of these surfaces produced a

Figure 1. Mass spectrum of the vanadia species generated from our cluster source where the x coordinates indicate the number of V atoms present in the cluster and the y coordinates the number of O atoms.

A positive bias applied to the sample during deposition decelerates the clusters, allowing them to be soft-landed with an incident kinetic energy between 0 and 3 eV/atom. An upperbound estimate of the cluster coverage is determined by measuring the ion current on the surface during the deposition. We estimate that ∼0.03 ML of size-selected clusters was deposited in these experiments, where 1 ML is defined as one vanadium atom per TiO2 unit cell. Scanning tunneling microscopy (STM) characterization suggested coverage of only ∼0.01 ML. STM characterization was done around the outer edges of the TiO2 sample because terraces in this region are the largest and most conducive to STM imaging at atomic resolution. The fact that our ion beam is focused on the center of the sample along with the relatively small area of the sample imaged explains the discrepancy between estimated and observed coverage. STM experiments were carried out at room temperature using a sample bias of +1.3 to +2.1 V and a constant tunneling current of 0.2 nA. Formaldehyde temperature-programmed reaction (TPR) experiments were carried out using a residual gas analyzer quadrupole mass spectrometer (Stanford Research Systems RGA 200 QMS (quadrupole mass spectrometer)). Clusterdecorated surfaces were exposed to 70 L of methanol at room temperature by introducing methanol into the chamber through a calibrated leak valve. These conditions for methanol exposure are equivalent to those used in the work of Wang and Madix for formaldehyde TPR.5,6 Samples were then heated at a rate of 0.6 K/s while recording the QMS signals at m/z = 29 amu to detect formaldehyde. We monitored m/z = 29 amu because it has a higher signal-to-noise ratio than the parent peak in our QMS. Methanol is also expected to desorb from the sample at the same temperature as formaldehyde;17−20 fragmentation of this methanol in the QMS will lead to additional signal at m/z = 29 that does not correspond to formaldehyde.19 To correct for this, we introduced pure methanol into the chamber without the titania sample present and calculated the ratio of the QMS signal at m/z = 29 to the signal at m/z = 31 (Supporting Information Figure 1a). During the TPD experiment, we determined the contribution of methanol to the signal at m/z = 29 amu by multiplying the signal at m/z = 31 by the cracking ratio. We then subtracted this contribution from the signal measured at m/z = 29 amu to determine the amount of formaldehyde produced (Supporting Information Figure 1). Because water is also believed to be a product of this reaction, we monitored the QMS signal at m/z

Figure 2. (a) Formaldehyde TPR spectra from bare (black line), Vdecorated (orange line), VO-decorated (blue line), and VO2-decorated (red line) TiO2(110). (b) Formaldehyde TPR spectra from bare (black line), V-decorated (orange line), VO-decorated (blue line), and VO2-decorated (red line) TiO2(110) surfaces following oxidation in 2 × 10−6 Torr of O2 at 540 K for 1 h. All spectra are plotted on the same y scale.

measurable amount of formaldehyde. The absence of formaldehyde production during methanol desorption from vacuum-annealed bare TiO2(110) is consistent with reports from Henderson and co-workers19 as well as Wong and coworkers.8−10 In an effort to generate catalytically active species, bare, V-, VO-, and VO2-decorated titania surfaces were annealed at 540 K for 1 h in 2 × 10−6 Torr of O2. These conditions are similar to those used by Wong and co-workers for preparing vanadia particles on TiO2.9 Formaldehyde TPR spectra (Figure 2b) show that formaldehyde is generated from each of these surfaces at ∼600 K, with the largest amount produced from the postoxidized VO-decorated surface (blue line). The presence of a formaldehyde TPR peak from the oxidized, bare surface (black line) is consistent with previous reports and has been attributed to a disproportionation reaction between two methoxy species stranded on the oxidized surface, resulting in the formation and simultaneous desorption of methanol and formaldehyde.5,6,8−10,19 To compare the amount of formaldehyde produced from these surfaces, the area under each TPR peak was integrated and normalized to oxidized VO/TiO2. Using the oxidized, bare TiO2 surface (normalized area = 0.32) as a baseline, it is clear that postoxidized, V-decorated TiO2 (orange line, normalized area = 0.48) exhibits some catalytic activity, although much less than the postoxidized VOdecorated surface. This observation implies that metallic vanadium atoms are more difficult to fully oxidize than VO B

dx.doi.org/10.1021/jp5011378 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

these features and cannot say how the reconstruction affects the VOx clusters. For type 2 terraces, we observe a mixture of round and elongated bright spots (Figure 3b) as well as larger features elongated perpendicular to the [001] direction that are likely the beginning of rosette formation or vanadia cluster aggregation. We assign the round spots as oxygen adatoms based on their resemblance to features that have been observed previously on bare TiO2 surfaces.22−24 To assign the elongated bright spots of Figure 3b, we examined line scans in the [001] direction along the clusters (Figure 3c). As shown in Figure 3c, these features have a length of 13 Å and height of 2.3 Å and are presumed to be the postoxidized VO clusters. In our previous work, we reported that mass-selected VO and VO2 clusters deposited on TiO2(110) exhibit characteristic dimensions in STM imaging; VO clusters are 10 Å long and 1.7 Å tall, while VO2 clusters are 14 Å long and 1.7 Å tall.17 Due to the fact that VO and VO2 are both coplanar with the surface, we attribute the increased height of these elongated spots to the presence of an out-of-plane, or vanadyl, oxygen. The presence of this vanadyl oxygen along with the fact that the length of this feature is very similar to VO2 (previously described as two adjacent 5c-Ti bound O atoms bridged by a V atom positioned in the upper three-fold hollow site) leads us to assign these features to VO3. A structural model for a VO3 cluster on the TiO2 support is shown in Figure 3d. Our model is identical to the lowest-energy DFT structure for VO3 on TiO2 that was presented in the work of Shapovalov and co-workers.13,17 Importantly, the authors also predicted that VO3 is the most thermodynamically stable composition of vanadia formed under postoxidation conditions similar to those used in this work.13 While we cannot say with absolute certainty that our active catalyst is VO3, our data implies that the presence of a vanadyl oxygen in the vanadia cluster is necessary for methanol oxidation. After performing formaldehyde TPR on postoxidized VO/ TiO2, we exposed the same surface to 70 L of methanol as soon as it cooled to room temperature. The resulting TPR spectrum (Figure 4a) shows that no detectable amount of formaldehyde

molecules, which is consistent with previous reports showing a mixture of several vanadium oxidation states resulting from postoxidation of evaporatively deposited vanadium at pressures below 10−3 Torr of O2.9 In contrast, oxidized VO2/TiO2 (red line, normalized area = 0.31) produces roughly the same amount of formaldehyde as the oxidized bare surface. Therefore, VO2 is not oxidized under these conditions to form an active catalyst. To gain structural information about the highly active oxidized VO-decorated surface, we carried out room-temperature STM to examine a VO-decorated titania surface that was annealed for 10 min at 600 K and a pressure of 2 × 10−6 Torr of O2. Oxidation of these surfaces for 1 h (to recreate identical conditions to those used for TPR experiments) produces a surface that is covered with TiO2 rosettes, making atomicresolution imaging unattainable. However, no large, agglomerated clusters were observed on these surfaces, implying that the vanadia did not sinter during this additional annealing. A large-area STM image (Figure 3a) exhibits several terraces that

Figure 3. (a) Large-area and (b) high-resolution STM images of the VO-decorated TiO2(110) surface following oxidation in 2 × 10−6 Torr of O2 at 600 K for 10 min. Numbers 1 and 2 indicate the rosette and characteristic atomic structure with vanadia clusters, respectively. Features marked with a circle are oxygen adatoms, while features marked with a square are vanadia clusters. (c) Line scan in the [001] direction along a vanadia cluster. (d) Proposed structural model for a VO3 cluster supported on the TiO2 surface. This model is identical to the lowest-energy DFT structure for VO3 on TiO2, which was calculated in the work of Shapovalov and co-workers.13

can be categorized as belonging to one of two types, (1) those that are covered in ordered bright patches and (2) those that contain bright protrusions on a background of alternating light and dark rows, the characteristic atomic-resolution structure for TiO2(110).23 On the basis of their size and structure, we assign the bright patches on type 1 terraces as TiO2 rosettes, which have been previously observed for titania surfaces that have been annealed in oxygen.21 Because the formation of the rosettes requires a reconstruction of the surface atoms, it is possible, even likely, that any VO clusters in this region of the surface become incorporated in the rosette features. However, we are unable to deduce the identity of individual atoms in

Figure 4. (a) Second run formaldehyde TPR spectrum from the oxidized VO-decorated TiO2 surface. (b) Formaldehyde TPR spectrum from the same surface following reoxidation in 2 × 10−6 Torr of O2 at 540 K for 1 h. All spectra are plotted on the same y scale as those shown in Figure 1. C

dx.doi.org/10.1021/jp5011378 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

■

was produced during this “second run”. This result is in agreement with previous studies that show that vanadiadecorated TiO2 is rendered inactive after carrying out a single TPR experiment in UHV and has been attributed to reduction of the VO3 cluster during methanol oxidation.7−9 Immediately following the second run TPR experiment, the same surface was reoxidized by annealing at 540 K for 1 h in 2 × 10−6 Torr of O2. Formaldehyde was produced from the reoxidized surface, as shown in the TPR spectrum of Figure 4b. The y scale of Figure 4 is identical to that used in Figure 1 to illustrate that the amount of formaldehyde generated from the reoxidized surface is smaller than that observed in the first cycle (TPR peak area = 0.60 normalized to the first TPR run from oxidized VO/TiO2). However, it is important to note that the yield is higher than that observed for the oxidized bare surface (normalized peak area = 0.32). This observation implies that the active catalyst can be at least partially regenerated by annealing in oxygen and is consistent with previous reports involving surfaces prepared without size selection.7−9 It is expected that VO3 is reduced to VO2 in the first TPR run, and subsequent annealing in O2 regenerates the active VO3 catalyst. Interestingly, we have shown above that deposited size-selected VO2 clusters cannot be oxidized to form VO3. This result suggests that the structures of “as-prepared” VO2 and VO2 generated from the reduction of VO3 are different. Furthermore, oxidation of the as-prepared structure must be kinetically hindered. We have previously shown that deposited, size-selected VO2 clusters have a planar structure, with the vanadium atom located between a five-fold-coordinated titanium (5c-Ti) atom row and a bridging oxygen row, and both oxygen atoms are bound to neighboring 5c-Ti atoms.17 Our STM results suggest that the VO3 cluster has a similar structure to VO2, with the third oxygen directed out of the plane of the surface.13 In order for the VO2 cluster resulting from methanol oxidation to have a different structure from as-prepared VO2, one of the planar oxygen atoms must be consumed in the formation of water. This leads to a VO2 cluster with one oxygen atom bound to a 5c-Ti atom and the other oxygen atom directed out of the plane of the surface. However, our second run TPR data show that only about 60% of the VO3 catalyst is regenerated, suggesting that after the initial TPR run, we likely produce a distribution of VO2 structures, where the out-of-plane structure is slightly favored relative to the planar structure. This suggests the existence of multiple pathways of methanol dehydrogenation by VO3.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +1-805-893-3393. Fax: +1-805-893-4120. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (CHE-1152229). Graduate student support was funded, in part, by the NSF PIRE ECCI program (OISE-0968399).

■

REFERENCES

(1) Wachs, I. E. Recent Conceptual Advances in the Catalysis Science of Mixed Metal Oxide Catalytic Materials. Catal. Today 2005, 100, 79−94. (2) Weckhuysen, B. M.; Keller, D. E. Chemistry, Spectroscopy and the Role of Supported Vanadium Oxides in Heterogeneous Catalysis. Catal. Today 2003, 78, 25−46. (3) Baldychev, I.; Vohs, J. M.; Gorte, R. J. The Effect of Support on Redox Properties and Methanol-Oxidation Activity of Vanadia Catalysts. Appl. Catal., A 2011, 391, 86−91. (4) Deo, G.; Wachs, I. E. Reactivity of Supported Vanadium Oxide Catalysts: The Partial Oxidation of Methanol. J. Catal. 1994, 146, 323−334. (5) Wang, Q.; Madix, R. J. Partial Oxidation of Methanol to Formaldehyde on a Model Supported Monolayer Vanadia Catalyst: Vanadia on TiO2(110). Surf. Sci. 2002, 496, 51−63. (6) Wang, Q.; Madix, R. J. Preparation and Reactions of V2O5 Supported on TiO2(110). Surf. Sci. 2001, 474, L213−L216. (7) Biener, J.; Bäumer, M.; Madix, R. J. A Synchrotron Study of the Deposition of Vanadia on TiO2(110). Surf. Sci. 1999, 432, 178−188. (8) Wong, G. S.; Kragten, D. D.; Vohs, J. M. TemperatureProgrammed Desorption Study of the Oxidation of Methanol to Formaldehyde on TiO2(110)-Supported Vanadia Monolayers. Surf. Sci. 2000, 452, L293−L297. (9) Wong, G. S.; Kragten, D. D.; Vohs, J. M. The Oxidation of Methanol to Formaldehyde on TiO2(110)-Supported Vanadia Films. J. Phys. Chem. B 2001, 105, 1366−1373. (10) Wong, G. S.; Concepcion, M. R.; Vohs, J. M. Reactivity of Monolayer V2O5 Films on TiO2(110) Produced via the Oxidation of Vapor-Deposited Vanadium. Surf. Sci. 2003, 526, 211−218. (11) Artiglia, L.; Agnoli, S.; Vittadini, A.; Verdini, A.; Cossaro, A.; Floreano, L.; Granozzi, G. Atomic Structure and Special Reactivity toward Methanol Oxidation of Vanadia Nanoclusters on TiO2(110). J. Am. Chem. Soc. 2013, 135, 17331−17338. (12) Kim, H. Y.; Lee, H. M.; Metiu, H. Oxidative Dehydrogenation of Methanol to Formaldehyde by a Vanadium Oxide Cluster Supported on Rutile TiO2(110): Which Oxygen Is Involved? J. Phys. Chem. C 2010, 114, 13736−13738. (13) Shapovalov, V.; Metiu, H. VOx (x = 1−4) Submonolayers Supported on Rutile TiO2(110) and CeO2(111) Surfaces: The Structure, the Charge of the Atoms, the XPS Spectrum, and the Equilibrium Composition in the Presence of Oxygen. J. Phys. Chem. C 2007, 111, 14179−14188. (14) Döbler, J.; Pritzsche, M.; Sauer, J. Oxidation of Methanol to Formaldehyde on Supported Vanadium Oxide Catalysts Compared to Gas Phase Molecules. J. Am. Chem. Soc. 2005, 127, 10861−10868. (15) Goodrow, A.; Bell, A. T. A Theoretical Investigation of the Selective Oxidation of Methanol to Formaldehyde on Isolated Vanadate Species Supported on Titania. J. Phys. Chem. C 2008, 112, 13204−13214. (16) Kemper, P.; Kolmakov, A.; Tong, X.; Lilach, Y.; Benz, L.; Manard, M.; Metiu, H.; Buratto, S. K.; Bowers, M. T. Formation, Deposition and Examination of Size Selected Metal Clusters on Semiconductor Surfaces: An Experimental Setup. Int. J. Mass Spectrom. 2006, 254, 202−209.

■

CONCLUSIONS The ability to deposit size-selected clusters has allowed us to definitively show that titania-supported V, VO, and VO2 are not catalytically active toward methanol oxidation. Postoxidation of supported V and VO, however, result in catalytically active species, with postoxidized VO/TiO2 producing the most formaldehyde. STM results show, for the first time, that these catalytically active species are isolated clusters and allow us to determine that the active catalyst is VO3.

■

Article

ASSOCIATED CONTENT

S Supporting Information *

the ratio of the QMS signal at m/z = 29 to the signal at m/z = 31. This material is available free of charge via the Internet at http://pubs.acs.org. D

dx.doi.org/10.1021/jp5011378 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

(17) Price, S. P.; Tong, X.; Ridge, C.; Shapovalov, V.; Hu, Z.; Kemper, P.; Metiu, H.; Bowers, M. T.; Buratto, S. K. STM Characterization of Size-Selected V1, V2, VO, and VO2 Clusters on a TiO2(110)-(1 × 1) Surface at Room Temperature. Surf. Sci. 2011, 605, 972−976. (18) Bates, S. P.; Gillan, M. J.; Kresse, G. Adsorption of Methanol on TiO2(110): A First-Principles Investigation. J. Phys. Chem. B 1998, 102, 2017−2026. (19) Henderson, M. A.; Otero-Tapia, S.; Castro, M. E. The Chemistry of Methanol on the TiO2(110) Surface: The Influence of Vacancies and Coadsorbed Species. Faraday Discuss. 1999, 114, 313− 329. (20) Farfan-Arribas, E.; Madix, R. J. Different Binding Sites for Methanol Dehydrogenation and Deoxygenation on Stoichiometric and Defective TiO2(110) Surfaces. Surf. Sci. 2003, 544, 241−260. (21) Li, M.; Hebenstreit, W.; Gross, L.; Diebold, U.; Henderson, M. A.; Jennison, D. R.; Schultz, P. A.; Sears, M. P. Oxygen-Induced Restructuring of the TiO2(110) Surface: A Comprehensive Study. Surf. Sci. 1999, 437, 173−190. (22) Matthiesen, J.; Wendt, S.; Hansen, J. Ø.; Madsen, G. K. H.; Lira, E.; Galliker, P.; Vestergaard, E. K.; Schaub, R.; Lægsgaard, E.; Hammer, B.; et al. Observation of All the Intermediate Steps of a Chemical Reaction on an Oxide Surface by Scanning Tunneling Microscopy. ACS Nano 2009, 3, 517−526. (23) Diebold, U. The Surface Science of Titanium Dioxide. Surf. Sci. Rep. 2003, 48, 53−229. (24) Lin, X.; Wang, Z.-T.; Lyubinetsky, I.; Kay, B. D.; Dohnálek, Z. Interaction of CO2 with Oxygen Adatoms on Rutile TiO2(110). Phys. Chem. Chem. Phys. 2013, 15, 6190−6195.

E

dx.doi.org/10.1021/jp5011378 | J. Phys. Chem. A XXXX, XXX, XXX−XXX