ARTICLE pubs.acs.org/JPCA
Gas Phase Synthesis of Au Clusters Deposited on Titanium Oxide Clusters and Their Reactivity with CO Molecules Hidenori Himeno, Ken Miyajima, Tomokazu Yasuike,† and Fumitaka Mafune* Department of Basic Science, Graduate School of Arts and Sciences, The University of Tokyo, Komaba, Meguro-ku, Tokyo 153-8902, Japan ABSTRACT: Titanium oxide clusters were formed in the gas phase by the laser ablation of a Ti rod in the presence of oxygen in a He gas. Not only stoichiometric but also nonstoichiometric titanium oxide clusters, TinO2n+x+ (n = 122 and x = 13), were formed. The content of oxygen atoms depends strongly on a partial pressure of oxygen. Gold clusters, Aum (m = 14), were generated by the laser ablation, which were then deposited on TinO2n+x clusters. The formation of AumTinO2n+x+ follows electron transfer from Aum to TinO2n+x+. The reactivity of AumTinO2n+x+ cluster ions with CO was examined for different m, n, and x by the mass spectrometry. It was found that Aum on TinO2n-1+ are less reactive than those on the other TinO2n+x+ (x = 0 and 1). In addition, the reactivity is highest when Aum (m = 1 and 3) is on the stoichiometric titanium oxide (x = 0), whereas the reactivity is also high when Au2 is on the oxygen-rich titanium oxide (x = 1). The reactivity was found to relate to geometrical structures of AumTinO2n+x+, which were studied by density functional calculations.
’ INTRODUCTION Since the discovery of catalytic activity of small gold clusters, the field of the catalysis of gold nanoclusters and nanoparticles has attracted the attention of many scientists.14 In the studies on gas-phase clusters, Cox et al. examined the chemical reactivity of gold clusters toward D2, CH4, and O2 molecules and found pronounced size dependences.5 After this pioneering work, the reactivity69 and geometric structure10,11 of free gold clusters have been major topics in the field of gas-phase cluster science. For the reaction with CO, Au clusters exhibit different reactivity depending on the charge state. Kimble et al. studied the reactivity of Au2,3Om m = 15 with CO experimentally and theoretically.12 In addition, many efforts have been made to probe the reactivity of size-selected Au clusters deposited on the surface. Heiz et al. studied the influence of surface defects on the reactivity of the smallest active cluster, Au8, on the MgO surface.1316 They calculated the excess charge on the cluster and showed the mechanism for the CO oxidation reaction. They also proposed that the key step for CO oxidation by the Au8 cluster on the MgO surface involves an O2 molecule bonded on the periphery of the interfacial layer of the cluster. Hence, it is very important to consider the effects on both the gold cluster and the substrate material. Inspired by the surface experiments, we had an idea to prepare both gold clusters and a cluster of substrate material in the gas phase, because defect-rich (nonstoichiometric) substrate clusters are known to be formed in the gas phase in a controlled manner. Specifically, we used titanium oxide clusters, which have been previously prepared and are well investigated. Gas-phase titanium r 2011 American Chemical Society
oxide clusters were prepared by Yu and Freas.17 They produced TinOm+ clusters by sputtering surface-oxidized titanium foil and titanium dioxide powder, respectively. The most prominent cluster ions have the stoichiometry TinO2n-d+ n = 18, d = 04 and n = 17, d = 13 for each case. Laser vaporization of a metal target in the presence of oxygen1824 and titanium compounds (TiO2, Ti(SO4)2 and PbTiO3)25,26 is also an effective way to produce metal oxide clusters. Guo et al. produced titanium oxide clusters by laser vaporization of Ti metal rods in high-pressure He gas, which contains a trace amount of O2.18 They investigated the reaction of oxygen-poor and oxygen-rich clusters with O2 in the drift tube, elucidating that (TiO2)m+ and TiO3(TiO2)m+ are quite inert and less reactive toward O2. Chen et al. produced (TiO)n(TiO2)m+ clusters via a laser vaporization of Ti metal with water vapor diluted by helium gas.18 They proposed that the TiO unit promotes the growth of the cluster, since this unit acts as an active site to incoming TiO2 units during the gas aggregation. The formation of neutral TinOm clusters has been investigated by Bernstein’s group extensively.20,22 Matsuda et al. measured the mass spectra of 118 nm single-photon ionized TinOm cluster distribution while varying the oxygen concentration of the expansion gas.22 Oxygen-deficient clusters are more abundantly formed when the O2 concentration is low enough. They concluded that the stable neutral clusters are TinO2n and TinO2n+1 under saturated O2 conditions (g2%). Received: March 5, 2011 Revised: September 1, 2011 Published: September 01, 2011 11479
dx.doi.org/10.1021/jp202125g | J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A
Figure 1. Experimental apparatus used in the present study. Two metal target rods are set in a block with two channels.
Geometric and electronic structure of small stoichiometric and oxygen-deficient TinO2n+x clusters have been extensively studied by theoretical calculation.2735 Yu and Freas investigated the energy-optimized geometrical structure of cluster ions by a simple pair-potential ionic model.17 They found that the lowest energy TinO2n+ isomers have pendant and terminal oxygen atoms, and the most abundant clusters have the lowest calculated energy per atom. They explained the predominant formation of TinO2n+x+ with x = 1, in terms of the feasible elimination of pendant oxygen atoms by collisional activation within the cluster source. Hagfeldt et al. performed ab initio calculations on neutral and singly positive charged TinO2n+x clusters, with n = 13 and x = 0, 1.27 They estimated that the ionization energies for the TinO2n clusters were significantly higher than those of the oxygen-deficient TinO2n1 clusters. Deng et al. investigated the ionization energies, Ei’s, of various niobium oxide clusters and indicated that oxygen-rich oxide clusters generally have higher Ei’s than metal-rich (oxide-poor) oxide clusters.36 In the present study, both gold clusters and titanium oxide clusters were prepared in the gas phase, and then the gold clusters were attached to titanium oxide clusters. We investigated the mechanism of deposition by mass spectrometry. In addition, the chemical reaction of the clusters with CO was investigated and found to be affected by the number of oxygen atoms for each TinO2n+x cluster.
’ EXPERIMENTAL SECTION Figure 1 shows the experimental setup used in the present study. The apparatus consists of a cluster ion source of AumTinO2n+x+, a temperature-controlled reaction cell, and a reflectron-equipped timeof-flight mass spectrometer. For the formation of AumTinO2n+x+ cluster ions, titanium oxide clusters and gold clusters were prepared in different gas flows of cylindrical channels (6 mm in diameter, respectively). Titanium oxide clusters were prepared by laser ablation of a Ti (99.9%; Nilaco Corp.) rod in the gas mixture O2 and He (>99.99995 vol%). The mixing ratio of O2 to He was prepared at 102 ∼ 103 ppm. To avoid the accidental degeneracy between 48Ti and three 16O, we used 18O2 gas (99.1 atom %; Taiyo-Nippon Sanso Corp.). Thus cluster composition of TinOm can be assigned with no ambiguity. Actually, a Ti rod was set downstream of a supersonic source from a solenoid pulsed valve (General Valve). The Ti rod was irradiated with focused laser pulses (∼10 mJ/pulse) at 532 nm from a Quanta Ray GCR-130 Nd:YAG laser for generating the plasma. The evaporated Ti atoms reacted with O2 and were also cooled
ARTICLE
in a cylindrical channel by the He gas from the valve, forming neutral and charged titanium oxide clusters. Also, gold clusters were prepared by laser ablation of an Au rod in the other cylindrical channel. The rod was irradiated with focused laser pulses (∼10 mJ/pulse) at 532 nm from a Continuum Surelite II Nd:YAG laser. Both clusters encountered at the end of channel (distance of rods to end is 17.5 mm), forming AumTinO2n+x+ cluster ions. After the formation, the clusters entered a reaction cell (2 mm in diameter 45 mm long) filled with a gas mixture of reactant and He, and an extension tube 4 mm in diameter and 120 mm long with a resistive heater. He gas was used to remove excess energy from the clusters, generated upon attachment of reactant to the clusters. In the present study, we used CO as a reactant gas. The concentration of CO gas was carefully optimized to minimize the unwanted overlapping of mass peaks in the mass spectrum due to multiple CO attachment. After the reaction cell, the cluster ions were introduced into a differentially pumped chamber through a skimmer. The native ions were accelerated orthogonally by the pulsed electric field for the TOF mass analysis, after they entered between a repeller and the first electrode. The ions gained a kinetic energy of ∼3.4 keV in the acceleration region. After traveling in a 1 m field-free region, the ions were reversed by the reflectron and were detected by a Hamamatsu double-microchannel plate detector. Signals from the detector were amplified by a Stanford SR445A 350 MHz preamplifier and were digitized using an oscilloscope (LeCroy LT344L). Averaged TOF spectra (typically 1000 sweeps) are sent to a personal computer for analysis via a general purpose interface bus (GPIB). The mass resolution, m/Δm, exceeds 1000. The geometry of the clusters was optimized by using the density functional theory with the B3LYP functional.37 The employed basis sets were 6-31G(d)38 for carbon and oxygen and the SBKJC basis set with its associated effective core potential39,40 for titanium and gold. All calculations were performed with the GAMESS quantum chemistry program package.41 The initial geometry of Ti6O12+ was prepared on the basis of most stable neutral Ti6O12 obtained by Qu and Kroes.32 We confirmed that this isomer is also the most stable in the corresponding cationic species. The initial geometries of Ti6O13+ or AuTi6O12+ are generated by attaching additional oxygen or gold to all titanium and oxygen sites of Ti6O12+. For Au2Ti6O12+, we considered two initial geometries obtained by attaching one gold atom to one of the gold and the dangling oxygen atoms of the most stable AuTi6O12+. The initial structures of AuTi6O13+ are obtained from the most stable AuTi6O12+ and Ti6O13+ by attaching one oxygen and one gold atom to all the available sites, respectively. In addition, the initial structures of Au2Ti6O13+ are similarly prepared from AuTi6O13+ and Au2Ti6O12+. For all the species, the lowest and second-lowest spin states were considered.
’ RESULTS Figure 2 shows a mass spectrum of positively charged titanium oxide clusters. As Ti is composed of five stable isotopes including 48 Ti (73.8%), there appear many mass peaks for each cluster ion. We are able to assign mass peaks uniquely by using 18O2 and taking the isotope ratios of the Ti atom into consideration. The mass peaks are assignable to TinO2n+x+ (n = 122 and x = 1, 0, 1, 2). The abundance of stoichiometric and nonstoichiometric titanium oxide cluster ions depends strongly on the O2 gas pressure, and TinO2n+x+ (x = 0, 1) were dominantly formed in 11480
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A
Figure 2. Mass spectrum of titanium oxide cluster ions formed by the laser ablation of the Ti metal rod in the presence of O2 in a He gas at room temperature. Tin18Ok+ clusters are designated by (n, k).
Figure 3. Mass spectra of (a) Ti718Ok+, (b) Aum+, and (c) AumTin18Ok+ at room temperature. All intensities are comparable except for Au3+ whose intensity is reduced to 1/5. Mass spectral peaks of AumTin18Ok+ are composed of several isotopic compositions so that labels are located near the highest peak. Calculated isotope pattern of Tin18Ok+ clusters is superimposed in (a). The peaks marked by asterisks are assigned to AumTin(16Oj18Okj)+ cluster isotopologues.
our present condition. No appreciable change in the abundance and size distribution of cluster ions were observed, when the fundamental light of the pulsed Nd:YAG laser was used instead of the second harmonic for the laser ablation.23 Figure 3 shows mass spectra of (a) Ti7O14+x+, (b) Au3+, and (c) AumTinO2n+x+ in the mass range of m/z 555635. Here, spectrum (a) was obtained with the laser for ablation of a titanium metal rod only, while the spectrum (b) was obtained with the laser for ablation of a gold metal rod only. Although the whole mass spectrum is not shown here, Au+ was most abundantly formed as cations of gold clusters in the gas phase, and the intensities of the gold cluster ions decreased with the cluster size. As shown in spectrum (c), it is evident that AumTinO2n+x+ were formed when both two lasers were operated. In addition, there are several points to stress: intensities of TinO2n+x+ (x = 0, 1) decrease as AumTinO2n+x+ are formed, but the intensity of Au3+ increases drastically. Close examination of the mass spectrum of AumTinO2n+x+ leads us to conclude that the propensity to form AumTinO2n+x+
ARTICLE
Figure 4. Relative abundances of AumTin18O2n+x+ clusters as functions of Au, Ti, and O atoms at room temperature. Each element in a map corresponds to each cluster ion, whose stoichiometry is shown by m, n, and x, the number of atoms involved in the cluster. Color scales represents the relative abundance normalized within each layer of m.
Figure 5. Mass spectra of AumTin18O2n+x+ clusters before and after reaction with a CO gas at room temperature. Blue and red lines represent before and after the reaction, respectively.
depends on the number of gold atoms, m and an excess number of the oxygen atom, x, while it does not depend significantly on a number of titanium atoms, n. Figure 4 shows the typical abundance map for AumTinO2n+x+ clusters. Interference due to the overlapping of TOF peaks was carefully examined. It was found from the abundance map, cluster ions with x = 1 are most abundant for TinO2n+x+, those with x = 0 are most abundant for Au1TinO2n+x+, and then those with x = 1 are most abundant for Au2TinO2n+x+. In the mass range of m/z 500800, the following clusters are prominent among the other stoichiometries: Ti6O13+, Ti7O15+, Ti8O17+, Ti9O19+, Au1Ti4O8+, Au1Ti5O10+, Au2Ti2O5+, Au2Ti3O7+, and Au2Ti4O9+. The reactivity of AumTinO2n+x+ with CO was measured. Figure 5 shows mass spectra of AumTinO2n+x+ with and without CO in the reaction gas cell located downstream of the cluster source. After the reaction with CO, the intensities of parent cluster ions, such as Au2Ti2O5+ and Au1Ti5O10+, decrease, whereas there appear CO-attached cluster ions, such as Au2Ti2O5CO+ and Au1Ti5O10CO+ in the mass spectrum. The mass peaks of CO-attached cluster ions do not appear to be increased appreciably after the reaction, because each peak is contributed by both a CO-attached cluster ion and a parent cluster ion, which 11481
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A
ARTICLE
Figure 7. Lowest energy isomers of Ti6O12+, Ti6O13+, Au1Ti6O12+, Au1Ti6O13+, Au2Ti6O12+, and Au2Ti6O13+. Yellow, green, and red balls correspond to Au, Ti, and O atoms, respectively. Arrows indicates the position of CO attachment.
Figure 6. Temperature dependence of mass spectra and the abundance of selected Tin18O2n+x+ clusters.
should decrease in its intensity. It is also likely that the nascent CO-attached cluster ions dissociate for dissipation of the available energy generated by the CO attachment.
’ DISCUSSIONS Formation of titanium oxide clusters. The titanium oxide cluster ions, TinO2n+x+, (n = 115 and x = 13) were formed by the laser ablation of the Ti metal rod in the presence of O2 in a He gas. A Ti atom and a Ti ion isolated in the gas phase should react with O2, and the produced building blocks should aggregate into TinO2n+x+. Formation of an oxygen-deficient cluster, TinO2n1+ has been intensively studied: a TiO unit promotes growth of the cluster by attaching a TiO2 unit, since this unit acts as an active site.20,2325 In this relation, Velegrakis et al. confirmed recently that the core ion is TiO+, because TiO+ is dominantly formed as a fragment ion as a result of multiphoton-dissociation of TinO2n+x+ (n = 23 and x = 1, 0) at a wavelength of 308 nm.24 In the present study, we focused our attention on TinO2n+1+ with one excess O atom, which are dominantly formed when the partial pressure of O2 is high. In addition, TinO2n+1+ were more richly formed for small n (n < 13), whereas TinO2n+0,-1+ became dominant for large n (n > 13). The formation of the clusters has already been reported by several research groups,1719 but the question to be answered is “How does the oxygen atom/ molecule attach to the cluster? Is the O2 molecule weakly bound to TinO2n-1+ forming TinO2n+1+?”. In order to identify the binding character of the excess oxygen, we measured the thermal dissociation of TinO2n+x+: nascent TinO2n+x+ formed in the
cluster source were introduced into a heated tube, and the cluster ions were then mass-analyzed. Figure 6a shows the intensities of Ti6O12+x+ after the tube as a function of the temperature (333413 K). The intensity of Ti6O12+ decreases gradually with the temperature. By contrast, the intensity of Ti2O3+ increases with the temperature as shown in Figure 6b. These findings suggest that Ti6O12+x+ dissociate in the heated tube. It is important to emphasize that the intensity of Ti6O13+ decreases similarly to Ti6O12+, and that no evidence of the increase of Ti6O11+ has been observed. This finding indicates that the binding energy of O2 in Ti6O13+ is higher than the binding energy of the network of the titanium oxides cluster ions, probably Ti5O11+ and TiO2. The structure of oxygen-excess Ti6O13+ was studied by the density functional calculation as shown in Figure 7: there is a common feature in titanium oxide clusters that an oxygen atom bridges two titanium atoms in the cluster.17,27,32 In addition, two dangling oxygen atoms exist in the cluster, which was manifested in the Ti6O12+ cluster. For the case of the Ti6O13+ cluster, the excess oxygen atom attaches next to the one dangling oxygen and forms TiOO bonding with the geometric structure of skeleton intact. The binding energies of the terminal O of TiOO in Ti6O13+ and the dangling O in Ti6O12+ are 3.67 and 3.78 eV, respectively. At these large binding energies, the atomic oxygen dissociation would not occur for either species at the experimental temperatures. The binding energy of molecular oxygen in Ti6O13+ is still large (2.09 eV), and the molecular oxygen dissociation from Ti6O13+ also would not occur. These results imply the contribution of another dissociation pathway and are consistent with the experimental observation. Formation of AumTinO2n+x+ Following Electron Transfer. Gold clusters were prepared by the laser ablation of a gold metal rod in a He gas, and the gas flow was merged with the gas flow of titanium oxide clusters into one downstream of the laser ablation block, where the gold clusters attach to the titanium oxide clusters. As we observed positively charged cluster ions, either of the clusters should be positively charged and the other one should be neutral, and the aggregation of both positively charged clusters is not likely. 11482
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A
ARTICLE
Figure 9. Relative reactivity of AumTin18O2n+x+ clusters for the reaction with CO at room temperature as color maps. (a) m = 0, (b) m = 1, (c) m = 2, (d) m = 3, and (e) averaged reactivity calculated from all available n. Red in the color scale corresponds to the higher reactivity, and white and blue mean inert. Figure 8. Abundance of Ti518O911+ and Au24+ clusters with changing the delay time between vaporization laser and acceleration of TOF-MS at room temperature. Origin of delay is defined at the appearance timing of the cluster signal. Filled symbols represent the intensity with single laser vaporization conditions (Ti only), and open symbols represent the intensity with double laser vaporization conditions (Ti and Au), respectively.
It is highly likely that, at first, an electron transfers from Aum to TinO2n+x+ when they encounter in the merged flow. In fact, the intensity of Au3+ increases by 10 times after the encounter, as shown in Figure 3. In order to examine the electron transfer mechanism more closely, we measured the spatial distribution of positively charged ions in the gas flow: In practice, the spatial distribution was measured by changing the time of pulse extraction for the mass analysis (see Figure 1). Note that a bunch of ions in the flow is at least 100 mm long, while the detection region by the pulse extraction is as long as ∼10 mm. We are able to detect the ions in any region by changing the delay time of pulse extraction. Figure 8a shows ion intensities of Ti5O10+x+ as a function of delay time. When the pulse laser for ablation of the gold metal rod is off, the ion intensities of Ti5O10+x+ peak at ∼45 μs. Then, the intensities of Ti5O10+ and Ti5O11+ around 45 μs are significantly reduced, and the peak position looks shifted to longer delay time, when the gold clusters are on. By contrast, Ti5O9+ scarcely disappears. These changes were observed for different sized TinO2n+x+: cluster ions with x = 0 and 1 are reduced, whereas those with x = 1 are not significantly reduced. On the other hand, Figure 8b shows ion intensities of Au2+, Au3+ and Au4+ as a function of delay time. It is evident that the intensities of Au3+ and Au4+ around 45 μs increase drastically, when gold clusters are allowed to encounter titanium oxide cluster ions. However, the increment of Au2+ intensity is limited. Ionization energies, Ei's, of Aum were known to be 9.26 ( 0.10, 9.16 ( 0.10, 7.27 ( 0.15, and 8.60 eV for m = 1, 2, 3,42 and 4,43 whereas the ionization energies of TinO2n are 6.5, 7.8, and 7.7 eV for n = 1, 2, and 3.27 As far as we know, there are no available data for the ionization energies of nonstoichiometric TinO2n+x clusters (x = 1 and 1) except for calculated ones [TiO (4.8 eV), Ti2O3 (3.8 eV), and Ti3O5 (3.9 eV)]27 and experimental ones (TiO3, TiO4, Ti2O6; 6.5 < Ei < 10.5 eV).22 Compared to the ionization energy of (TiO2)n (∼8 eV), it is conceivable that TinO2n+x with x = 1 has much lower ionization energy than TinO2n. As mentioned above, the occurrence of electron transfer
strongly relates to the ionization energy: electron transfer occurs between low Ei neutral gold clusters (Au3 and Au4) and high Ei titanium oxide cluster ions (TinO2n+x+ (x g 0)), increasing the abundance of Au3+ and Au4+. In contrast, low Ei TinO2n+x+ (x = 1) are not subjected to electron transfer. The formation mechanism of AumTinO2n+x+ can be summarized in terms of the electron transfer and the attachment reactions: electron transfer : Aum þ Tin O2nþx þ f Aum þ þ Tin O2nþx
ð1Þ attachment : Aum þ þ Tin O2nþx f Aum Tin O2nþx þ
ð2Þ
The electron transfer is significant for m = 3, 4 and larger clusters, because actually, the abundances of Aum+ after the electron transfer reaction are much higher than that of nascent Aum+. As the counter species, x = +1 is the most dominant ion as the nascent TinO2n+x+ prepared by the laser ablation. Hence, AumTinO2n+1+ is presumed to be most abundant. In fact, as shown in Figure 4, this is the case for m = 2, 4, and the larger ones (not shown). However, exceptionally, x = 0 is most abundant for m = 1 and 3. It is interesting to note that the number of the oxygen atom is extended to more than x = +1. Thus the contribution of the attachment of the O atom toward gold atoms also needs to be considered. This is an issue for further investigation. Reactivity of AumTinO2n+x+ with CO. We further examined the reactivity of AumTinO2n+x+ with CO. Relative chemical reactivities were obtained from the depletion ratios of AumTinO2n+x+ clusters as the result of the reactions with a CO gas. We estimated the relative reactivity of the CO attachment reaction of all clusters observed to within an experimental uncertainty of +/20%. Figure 9 shows the relative chemical reactivities of AumTinO2n+x+ clusters with m, n, x as a color map. It was found that the reactivity of AumTinO2n+x+ depends on m and x, and not significantly on n, which is manifested by the horizontal lines in the plot. Then, we averaged the depletion ratio over n, and plotted it against m and x. As seen in Figure 9e, there is a general trend that Aum on TinO2n-1+ are less reactive than those on the other TinO2n+x+. In addition, the reactivity is highest when Aum (m = 1 and 3) is on the stoichiometric titanium oxide (x = 0), whereas the reactivity is also high when Au2 is on the oxygen-rich titanium oxide (x = 1). 11483
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A In order to understand the reactivity depending on m and x, the structure of Au1,2Ti6O12,13+ was studied by the density functional calculation as shown in Figure 7. Gold atoms tend to attach next to the dangling oxygen atom. The propensity is consistent with the H+(TiO2)n cluster calculated by Calatayud et al., where H+ is adsorbed on the terminal O atom and the skeleton remains the same as for the neutral clusters.44 The structure of H(TiO2)6+ presented in Figure 5 in ref 44 resembles that of the Au1Ti6O12+ cluster in this study. According to the optimized geometrical structure, the second gold atom of Au 2Ti6 O12 + is likely to attach to the first gold atom, forming a gold dimer at the end of the dangling oxygen. For Au2Ti6 O13 +, the excess oxygen atom inserts into the gold dimer, and bridges two gold atoms. We are continuing the calculations for clusters with more than two gold atoms. To understand the experimentally observed reactivity, binding energies of CO to Au1,2Ti6O12,13+ clusters were calculated. As a whole, the binding energies at the skeletal Ti sites range within 0.60.7 eV. The difference of the binding energies arises when CO adsorbs on tail sites such as Au, OAu, Au2, and AuOAu. As shown in Figure 7, CO is able to bind to Au of Au1Ti6O12+ with a binding energy of 0.77 eV. By contrast, CO does not bind to Au1Ti6O13+, as the terminal O atom attaches on the end of the gold atom. Generalizing the results of the DFT calculations, one can explain the reason why Au1TinO2n+ and Au2TinO2n+1+ binds strongly and Au1TinO2n+1+ does not as follows: (1) CO does not bind strongly to a terminal oxygen of OAuTinO2n+. The terminal O is locally neutral (triplet), and the interaction with CO is weak. (2) CO binds to the Au atom behind the terminal O atom strongly. The Au atom is positively charged, and it causes the chargedipole or chargeinduced dipole interaction with CO.45 (3) CO can bind to the Au dimer with moderate strength. The positive charge is delocalized over the Au dimer, and the interaction with CO is weaker than in the positive charge localized in the single Au atom.
’ CONCLUSION Titanium oxide clusters were formed in the gas phase by the laser ablation of a Ti rod in the presence of oxygen in He gas. In addition, gold clusters, Aum (m = 14), were generated by the laser ablation, which were then deposited on TinO2n+x clusters. The enhancement of Au2,3+ clusters was observed when TinO2n+x and Au clusters are both produced. The formation of AumTinO2n+x+ follows electron transfer from Aum to TinO2n+x+. The reactivity of the AumTinO2n+x+ cluster ions with CO was examined for different m, n, and x by the mass spectrometry. It was found that the reactivity is highest when Aum (m = 1 and 3) is on the stoichiometric titanium oxide (x = 0), whereas the reactivity is also high when Au2 is on the oxygen-rich titanium oxide (x = 1). The reactivity was found to relate to geometrical structures of AumTinO2n+x+, which were studied by the density functional calculation. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Present Addresses †
Institute for Molecular Science, Okazaki, Aichi, Japan.
ARTICLE
’ ACKNOWLEDGMENT This work is supported by the Grant-in-Aids for Scientific Research (B) (No. 22350004) from the Ministry of Education, Culture, Sports, Science and Technology, Japan (MEXT) and by the Genesis Research Institute, Inc. for the cluster research. F.M. acknowledges Professor Greg Metha and Professor Mark A. Buntine for helpful discussions on the experimental setup. ’ REFERENCES (1) Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. J. Catal. 1989, 115, 301–309. (2) Haruta, M. Catal. Today 1997, 36, 153–166. (3) Chen, M. S.; Goodman, D. W. Science 2004, 306, 252–255. (4) Campbell, C. T. Science 2004, 306, 234–235. (5) Cox, D. M.; Brickman, R.; Creegan, K.; Kaldor, A. Z. Phys. 1991, 19, 353–355. (6) Hagen, J.; Socaciu, L. D.; Heiz, U.; Bernhardt, T. M.; W€oste, L. Eur. Phys. J. D 2003, 24, 327–330. (7) Veldeman, N.; Lievens, P.; Andersson, M. J. Phys. Chem. A 2005, 109, 11793–11801. (8) Bernhardt, T. M.; Socaciu, L. D.; Hagen, J.; W€oste, L. Appl. Catal., A 2005, 291, 170–178. (9) Johnson, G. E.; Reilly, N. M.; Tyo, E. C.; Castleman, A. W., Jr. J. Phys. Chem. C 2008, 112, 9730–9736. (10) Xing, X.; Yoon, B.; Landman, U.; Parks, J. H. Phys. Rev. B 2006, 74, 165423. (11) Gruene, P.; Rayner, D. M.; Redlich, B.; van der Meer, A. F. G.; Lyon, J. T.; Meijer, G.; Fielicke, A. Science 2008, 321, 674–676. (12) Kimble, M. L.; Moore, N. A.; Johnson, G. E.; Castleman, A. W., Jr.; B€urgel, C.; Mitric, R.; Bonacic-Koutecky , V. J. Chem. Phys. 2006, 125, 204311. (13) Sanchez, A.; Abbet, S.; Sanchez, A.; Heiz, U.; Schneider, W.-D.; H€akkinen, H.; Barnett, R. N.; Landman, U. J. Phys. Chem. A 1999, 103, 9573–9578. (14) H€akkinen, H.; Abbet, S.; Sanchez, A.; Heiz, U.; Landman, U. Angew. Chem. 2003, 42, 1297–1300. (15) Molina, L. M.; Rasmussen, M. D.; Hammer, B. J. Chem. Phys. 2004, 120, 7673–7680. (16) Arenz, M.; Landman, U.; Heiz, U. Chem. Phys. Chem. 2006, 7, 1871–1879. (17) Yu, W.; Freas, R. B. J. Am. Chem. Soc. 1990, 112, 7126–7133. (18) Guo, B. C.; Kerns, K. P.; Castleman, A. W., Jr. Int. J. Mass. Spectrom. Ion Proc. 1992, 117, 129–144. (19) Chen, Z. Y.; Walder, G. J.; Castleman, A. W., Jr. Phys. Rev. B 1994, 49, 2739–2752. (20) Foltin, M.; Stueber, G. J.; Bernstein, E. R. J. Chem. Phys. 1999, 111, 9577–9586. (21) von Helden, G.; van Heijnsbergen, D.; Meijer, G. J. Phys. Chem. A 2003, 107, 1671–1688. (22) Matsuda, Y.; Bernstein, E. R. J. Phys. Chem. A 2005, 109, 314–319. (23) Velegrakis, M.; Sfounis, A. Appl. Phys. A: Mater. Sci. Process. 2009, 97, 765–770. (24) Jadraque, M.; Sierra, B.; Sfounis, A.; Velegrakis, M. Appl. Phys. B: Laser Opt. 2010, 100, 587–590. (25) Liu, X.-H.; Zhang, X. G.; Li., Y.; Wang, X.-Y.; Lou, N. Q. Int. J. Mass Spectrom. 1998, 177, L1–L4. (26) Chaoui, N.; Millon, E.; Muller, J. F. Chem. Mater. 1998, 10, 3888–3894. (27) Hagfeldt, A.; Bergstr€om, R.; Siegbahn, H. O. G.; Lunell, S. J. Phys. Chem. 1993, 97, 12725–12730. (28) Wu, H.; Wang., L. S. J. Chem. Phys. 1997, 107, 8221–8228. (29) Zhai, H. J.; Wang, L. S. J. Am. Chem. Soc. 2007, 129, 3022–3026. (30) Albaret, T.; Finocchi, F.; Noguera, C. J. Chem. Phys. 2000, 113, 2238–2249. (31) Hamad, S.; Catlow, C. R. A.; Woodley, S. M.; Lago, S.; Mejías, J. A. J. Phys. Chem. B 2005, 109, 15741–15748. 11484
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
The Journal of Physical Chemistry A
ARTICLE
(32) Qu, Z.-W.; Kroes, G.-J. J. Phys. Chem. B 2006, 110, 8998–9907. (33) Li, S.; Dixon, D. A. J. Phys. Chem. A 2008, 112, 6646–6666. (34) Jeong, K. S.; Chang, C.; Sedlmayr, E.; S€ulzle, D. J. Phys. B: At., Mol. Opt. Phys. 2000, 33, 3417–3430. (35) Zhang, D.; Sun, H.; Liu, J.; Liu, C. J. Phys. Chem. C 2009, 113, 21–25. (36) Deng, H. T.; Kerns, K. P.; Castleman, A. W., Jr. J. Phys. Chem. 1996, 100, 13386–13392. (37) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (38) Hehre, W.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257–2261. (39) Stevens, W. J.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81, 6026–6033. (40) Stevens, W. J.; Krauss, M.; Basch, H.; Jasien, P. G. Can. J. Chem. 1992, 70, 612–630. (41) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347–1363. (42) Cheeseman, M. A.; Eyler, J. R. J. Phys. Chem. 1992, 96, 1082–1087. (43) Wang, J.; Wang, G.; Zhao, J. Phys. Rev. B 2002, 66, 035418. (44) Calatayud, M.; Maldonado, L.; Minot, C. J. Phys. Chem. C 2008, 112, 16087–16095. (45) Natural charge distribution of clusters is estimated by the natural population analysis.46 Natural charge on the gold moieties are Au(+0.75)-Ti6O12 and Au(+0.65)-O(0.59)-Au(+0.67)-Ti6O12, for Au1Ti6O12+ and Au2Ti6O13+, respectively. (46) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83, 735–746.
11485
dx.doi.org/10.1021/jp202125g |J. Phys. Chem. A 2011, 115, 11479–11485
