Article pubs.acs.org/JPCA
Reactions of Ti- and V‑Doped Cu Cluster Cations with Nitric Oxide and Oxygen: Size Dependence and Preferential NO Adsorption Shinichi Hirabayashi† and Masahiko Ichihashi*,‡ †
East Tokyo Laboratory, Genesis Research Institute, Inc., 717-86 Futamata, Ichikawa, Chiba 272-0001, Japan Cluster Research Laboratory, Toyota Technological Institute: in East Tokyo Laboratory, Genesis Research Institute, Inc., 717-86 Futamata, Ichikawa, Chiba 272-0001, Japan
‡
ABSTRACT: Reactions of copper cluster cations doped with an early transition metal atom, CunTi+ (n = 4−15) and CunV+ (n = 5− 14, 16), with NO and O2 were investigated at a near-thermal collision energy using a guided ion beam tandem mass spectrometer. Most of the clusters adsorb NO and O2 under single collision conditions, and this reaction is often followed by the release of Cu atoms. For both Ti- and V-doped Cu clusters, the total cross sections for the reaction with NO increase gradually with the cluster size up to n ≈ 11 and then decrease rapidly, whereas those with O2 are almost constant in n ≤ 12 and then decrease. The size dependence of the reactivity toward NO is found to correlate with that of the adsorption energy calculated by the density functional theory method; CunTi+ clusters exhibit the larger reaction cross sections when they have the larger adsorption energies. The calculations of CunTi+ also show that a structural transition from a Tiexposed structure to Ti-encapsulated one occurs around n = 12. This indicates that a geometric property of the clusters, i.e., the position of the dopant atom, is a determining factor of reactivity. In addition, the Ti- and V-doping dramatically improves the reactivity of Cu cluster cations toward NO but it does not affect that toward O2 significantly. As a result, most of the Ti- and Vdoped Cu clusters are more reactive toward NO than toward O2. We also studied the multiple-collision reaction of Cu7Ti+ with NO and obtained the cluster dioxide, Cu3TiO2+, as a product ion, which suggests that the dissociation of NO and the subsequent formation/release of N2 take place.
1. INTRODUCTION Air pollution from vehicles is one of the most serious environmental problems,1 and the use of three-way catalysts (TWC) is currently the best way to reduce the air pollutants in exhaust gases from automotive gasoline engines.2 The TWC generally contain small metal particles of Rh, Pd, and Pt, and clean up carbon monoxide (CO), hydrocarbons (HC), and nitrogen oxides (NOx) at almost stoichiometric air-to-fuel ratios. However, it is required that the usage of these metals is reduced or minimized in the catalysts because of their high cost and scarcity. Common and abundant metals should be substituted for the platinum group metals. Additionally, the TWC cannot reduce NOx efficiently in lean-burn gasoline and diesel engines operating under the presence of excess O2 because the reductants (CO and HC) are oxidized by O2 rather than NOx.1,2 The actual conversion efficiency of NOx is thus affected by the competition between NOx and O2 to adsorb dissociatively. Copper-based catalysts are promising alternatives to the platinum group metal catalysts for NOx reduction.3,4 For instance, copper ion-exchanged zeolites have attracted considerable interest since Iwamoto et al. found that these catalysts are active for direct decomposition of NO.5 It has been shown that these are also attractive for selective catalytic reduction of © XXXX American Chemical Society
NO with ammonia or hydrocarbons under oxidizing atmospheres.4,6,7 In addition to these zeolite-supported catalysts, Al2O3-supported copper catalysts were reported to be effective for the reduction of NO by CO in the presence of O2.8 Although monomeric and dimeric Cu species have been proposed to be active sites in these catalysts,4,8 the reactivity of small Cu clusters has not been revealed yet. Therefore, it is essential to understand how the reactivity of Cu clusters changes with the cluster size. Studies of metal clusters isolated in the gas phase can provide basic information on the reactivity of the clusters themselves without the influence of the surrounding environments. The reactions of isolated neutral and ionic Cu clusters with NO have been studied by Holmgren et al.9 and by our group,10 respectively. These studies indicate that the adsorption probabilities of NO on Cun (n = 15−80) and Cun± (n = 3− 19) are very low overall at room temperature (or at a nearthermal energy), though large neutral clusters (n ≈ 40−60) and some anions (n = 8, 10, 12, 14−16) exhibit slightly high probabilities. Recently, we investigated the reactions of CunO2− Received: January 8, 2016 Revised: February 23, 2016
A
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A (n = 4−17)10,11 and CunAl+ (n = 4−6, 8−16)12 with NO at a near-thermal energy and revealed that the introduction of O and Al into specific-sized Cu clusters improves the adsorption probability and further promotes the decomposition of NO. Moreover, it was found that CunAl+ (n = 9, 11, 13, 15) show higher reactivity toward NO than toward O2.12 These studies demonstrate that the doping of foreign atoms in Cu clusters changes their properties drastically and thereby can enhance the specific reactivity of the clusters. In this article, we report on the reactivity of copper cluster cations doped with a titanium and vanadium atom, CunTi+ (n = 4−15) and CunV+ (n = 5−14, 16), toward NO and O2 at a near-thermal energy. The reactivity of these clusters is found to change with the cluster size similar to each other, and the size dependence is discussed in terms of the geometric and electronic characters of the clusters. The influence of the Ti and V dopant atom on the reactivity is elucidated by comparing with the results obtained for pure Cu cluster cations in our previous studies.10,13
the product ions, respectively. A partial reaction cross section, σp, for the formation of a given product ion was obtained using Ip σp = σr ∑ Ip (2) where Ip/ΣIp represents the branching fraction for the product ion of interest. In an electrostatic interaction model,15 the collision cross sections are calculated to be 49.2 Å2 for NO and 47.4 Å2 for O2 at Ecol = 0.2 eV.
3. RESULTS 3.1. Size Distributions of CunTi+ and CunV+. Figure 1 shows typical mass spectra of cluster ions produced by
2. EXPERIMENTAL SECTION Metal cluster ions were produced in an ion sputtering source, and their reactivity was investigated by a tandem-type mass spectrometer, as described in our previous publications.10−14 In the cluster ion source, four separate metal targets were sputtered with 8.5 keV accelerated xenon ion beams from an ion gun (CORDIS Ar25/35c, Rokion Ionenstrahl-Technologie). In this experiment, three copper plates (99.96% purity) and one titanium (99.9% purity) or vanadium (99.9% purity) plate were used as targets. The cluster cations produced by the ion sputtering were extracted by a series of ion lenses and admitted into an octopole ion beam guide (OPIG) passing through a 290 mm long cooling cell, which was filled with He gas (>10−2 Torr) at room temperature. The thermalized cluster ions were guided by the second OPIG into the first quadrupole mass filter (QMF) where clusters of interest were selected. The mass-selected ions (parent cluster ions) were further guided by the third OPIG and transferred into a 100 mm long reaction cell, where they were allowed to react with NO or O2. The pressure of the reactant gas was measured with a spinning rotor gauge (SRG-2, MKS) connected to the reaction cell and kept below 2 × 10−4 Torr in the single-collision reactions. The translational energy of the parent cluster ions in the reaction cell was measured by the retarding potential method using the third OPIG without reactants and converted to the collision energy, Ecol, in the center-of-mass frame. A typical spread of the collision energy was obtained as 0.4 eV in the full width at halfmaximum (fwhm). Unreacted parent cluster ions and product ions were mass-analyzed with the second QMF and then collected by a series of ion lenses to a secondary electron multiplier equipped with an ion conversion dynode. Signals from the secondary electron multiplier were processed in a pulse counting mode. The total reaction cross section, σr, was derived from σr =
I + ∑ Ip kBT ln Pl I
Figure 1. Typical mass spectra obtained by cosputtering of three Cu targets and one target of (a) Ti and (b) V. The peaks of CunTi+ and CunV+ are shaded and connected by thick lines.
cosputtering of three Cu targets and one Ti (or V) target. A series of copper cluster cations doped with a Ti or V atom appear in the mass spectra, together with the intense peaks of pure Cu clusters. There is no even−odd alternation in the size distributions of CunTi+ and CunV+ in contrast to that of Cun+, but a stepwise drop of the intensities of CunTi+ and CunV+ is observed beyond n = 15 and 14, respectively. Similar intensity drops were seen in the size distributions of AgnM+ (M = Ti, V)16 and AunTi+ (ref 17) at n = 14 or 15, and the enhanced stability of the clusters was attributed to the 18-valence electron system in the spherical two-step jellium model.18 In addition, Cu16Sc+ with 18 valence electrons was observed as a prominent peak in a mass spectrum, which was also interpreted in terms of an extremely stable dopant encapsulated structure featuring a closed electron shell.19,20 3.2. Single-Collision Reactions of CunTi+ with NO and O2. In the reactions of CunTi+ (n = 4−15) with NO, five different product ions, Cun−mTiNO+ (m = 0−4), are observed at Ecol = 0.2 eV: Cu nTi+ + NO → Cu n − mTiNO+ + [mCu]
(1)
(m = 0−4) (3)
where kB is the Boltzmann constant, P and T are the pressure and temperature of the reactant gas, respectively, l (= 120 mm) is the effective path length of the reaction region, and I and ΣIp represent the intensity of the unreacted parent ion passing through the reaction region and the sum of the intensities of
where the square bracket is used to indicate the ambiguity of the neutral products. As a representative example, a mass spectrum for the reaction of Cu7Ti+ with NO obtained at the pressure of ∼2 × 10−4 Torr is shown in Figure 2a, where three reactions among the five occur to produce Cu6TiNO+, B
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
the reactions. It is also possible that the dissociation of the larger clusters is suppressed efficiently because they can store the excess energy resulting from the adsorption for a longer time due to more vibrational degrees of freedom. In the reactions of CunTi+ (n = 4−15) with O2, several species, Cun−mTiO2+ (m = 1−4), Cun−mTiO+ (m = 0, 2, 3), and Cun−1Ti+ (only for n = 4), are observed as product ions at Ecol = 0.2 eV. Among them, Cun−mTiO2+ dominates in all of the clusters studied and should be obtained from the following reactions: Cu nTi+ + O2 → Cu n − mTiO2+ + [mCu]
(m = 1−4) (4)
CunTiO2+,
The transient intermediate, is not observed in any clusters studied. As shown in Figure 4a, the total cross section
Figure 2. Mass spectra of parent and product ions obtained in the reactions of Cu7Ti+ with NO under (a) single collision (∼2 × 10−4 Torr) and (b) multiple collision (∼2 × 10−3 Torr) conditions at the initial collision energy of 0.2 eV. The inset in (b) indicates the peak due to Cu3TiO2+. Here, two Gaussian functions centered at 268.5 and 270.5 amu with the fwhm of 4 amu, which is estimated from the peak widths of Cu7Ti+ and Cu4−6TiNO+, are fitted to this peak to separate the contribution of Cu3TiNO+ (see text).
Cu5TiNO+, and Cu4TiNO+. Figure 3 shows the total and partial reaction cross sections for CunTi+ + NO as a function of
Figure 4. (a) Total cross sections for the reactions of CunTi+ (n = 4− 15) with O2 as a function of the number of copper atoms in the cluster at the collision energy of 0.2 eV. (b) Partial cross sections for the formation of Cun−1TiO2+, Cun−2TiO2+, Cun−3TiO2+, and Cun−4TiO2+. Note that the measurements in the reactions of CunTi+ (n = 13−15) were performed in the narrower mass ranges including only the first three ions.
for the reaction of CunTi+ with O2 is slightly larger (∼11 Å2) at n = 4, but does not change significantly (6−9 Å2) in the size range of n = 5−12, and then decreases rapidly to zero at n = 15. Notably, Cu15Ti+ is unreactive against O2. We recognize that there are even−odd alternations in the partial reaction cross sections; the cluster dioxide ions with an odd number of copper atoms possess higher stability (Figure 4b). The dominant product ion switches roughly in the order of Cun−3TiO2+, Cun−2TiO2+, and Cun−1TiO2+ with the increase of n. 3.3. Single-Collision Reactions of CunV+ with NO and O2. The reactions of CunV+ (n = 5−14, 16) with NO at Ecol = 0.2 eV result in the formation of three kinds of NO adsorption species, CunVNO+, Cun−1VNO+, and Cun−2VNO+:
Figure 3. (a) Total cross sections for the reactions of CunTi+ (n = 4− 15) with NO as a function of the number of copper atoms in the cluster at the collision energy of 0.2 eV. (b) Partial cross sections for the formation of Cu n TiNO + , Cu n−1 TiNO + , Cu n−2 TiNO + , Cun−3TiNO+, and Cun−4TiNO+. Note that the measurements in the reactions of CunTi+ (n = 13−15) were performed in the narrower mass ranges including only the first three ions.
the cluster size, n. The total reaction cross section increases gradually with the number of Cu atoms in the cluster from 15 Å2 at n = 4 to 32 Å2 at n = 11 and declines rapidly to zero at n = 15. As shown in Figure 3b, even−odd alternations appear in the cross sections of the Cun−mTiNO+ (m = 1−3) formation; the cross sections of the products with an even number of Cu atoms are larger than those of their neighbors with an odd number of Cu atoms. Cun−2TiNO+ or Cun−3TiNO+ dominates the products in n = 4−6, and Cun−1TiNO+ is dominant in n = 7−11. The transient intermediate, CunTiNO+, is observed as a major product only at n = 12 and 13. These changes in the dominant products can simply reflect the thermochemistry for
Cu nV + + NO → Cu n − mVNO+ + [mCu]
(m = 0−2) (5)
As shown in Figure 5a, the total reaction cross section increases with the number of Cu atoms up to n = 11 and then decreases rapidly. Small clusters with n = 5−12 show relatively large cross sections (>10 Å2), whereas the larger clusters are quite unreactive. Figure 5b indicates that Cun−2VNO+ and C
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Figure 5. (a) Total cross sections for the reactions of CunV+ (n = 5− 14, 16) with NO as a function of the number of copper atoms in the cluster at the collision energy of 0.2 eV. (b) Partial cross sections for the formation of CunVNO+, Cun−1VNO+, and Cun−2VNO+. Note that the measurements in the reactions of CunV+ (n = 13, 14, 16) were performed in the narrower mass ranges including only the first two ions.
Figure 6. (a) Total cross sections for the reactions of CunV+ (n = 5− 14, 16) with O2 as a function of the number of copper atoms in the cluster at the collision energy of 0.2 eV. (b) Partial cross sections for the formation of CunVO2+, Cun−1VO2+, Cun−2VO2+, and Cun−3VO2+. Note that the measurements in the reactions of CunV+ (n = 13, 14, 16) were performed in the narrower mass ranges including only the first three ions.
Cun−1VNO+ dominate at n = 5 and 6−11, respectively. The CunVNO+ intermediate is observed particularly at n = 11 and 12. In the reactions of CunV+ (n = 5−14, 16) with O2, the following O2 adsorption with and without Cu release occurs at Ecol = 0.2 eV: Cu nV + + O2 → Cu n − mVO2+ + [mCu]
(m = 0−3) (6)
Other product ions such as Cu5O2+, Cu5O+, Cu4VO+, and Cu3+ are observed only in the reaction of Cu5V+, which is the smallest cluster studied here. The total reaction cross sections are almost flat within the narrow range of 4−8 Å2, as shown in Figure 6a. The exception is Cu14V+, which exhibits a relatively small O2 adsorption cross section. As the cluster size increases, the number of Cu atoms released in the reaction tends to decrease (Figure 6b). The O2 adsorption without Cu release is dominantly observed at the large cluster sizes of n = 14 and 16. 3.4. Multiple-Collision Reactions of Cu7Ti+ with NO. As mentioned above, both CunTi+ and CunV+ (n ≤ 12) have large adsorption cross sections for the first NO molecule. To investigate the subsequent reactions of adsorbed NO molecules, NO-pressure-dependent experiments were carried out under multiple collision conditions for the reaction of Cu7Ti+ as a typical example. As shown in Figure 2b, a peak (∼270 amu) assignable to Cu3TiNO+ and/or Cu3TiO2+ appears in the mass spectra measured at the NO pressure higher than 6 × 10−4 Torr, together with the single-collision product ions (Cu6TiNO+, Cu5TiNO+, and Cu4TiNO+). Because we did not have enough intensity of this peak (∼270 amu) to improve the mass resolution for the isotope splitting against the transmittance of the second QMF, the abundances of these products are estimated by fitting two Gaussian functions to the mass peak. As a result, we found that the contribution of Cu3TiNO+ is negligible and this peak is almost exclusively due to Cu3TiO2+ (the inset of Figure 2b). Figure 7 shows the NO-pressure dependence of the relative intensities of
Figure 7. Relative intensities of the parent ion and the major product ions in the reactions of Cu7Ti+ with NO as a function of the NO pressure at the initial collision energy of 0.2 eV. The dashed lines represent the first- and second-order dependences on the NO pressure for eye guides.
the parent ion and main product ions. The relative intensity of Cu3TiO2+ exhibits roughly the second-order dependence on the NO pressure. Then, it is concluded that NO decomposition proceeds as follows: Cu 7Ti+ + 2NO → Cu3TiO2+ + [4Cu, N2]
(7)
Similar pressure dependence was observed previously for Cun−2O4− and Cu6AlO2+ in the reactions of CunO2− (n = 8, 10, 12) + NO11 and Cu9Al+ + NO,12 respectively.
4. DISCUSSION In the mass spectra of CunTi+ and CunV+, we observe stepwise intensity drops from Cu15Ti+ to Cu16Ti+ and from Cu14V+ to Cu15V+ (Figure 1). This indicates that Cu15Ti+ and Cu14V+ do not dissociate easily because of the large binding energy of Cu atoms in comparison with Cu16Ti+ and Cu15V+. It is expected D
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A that the Ti (or V) atom is located at the center of Cu15Ti+ (or Cu14V+) to produce as many Cu−Ti (or V) bonds as possible. Thus, the first geometric shell of Cu atoms around the Ti (or V) atom is probably completed at Cu15Ti+ (or Cu14V+) and an outer Cu shell starts from Cu16Ti+ (or Cu15V+). Similar stable clusters, Cu16Sc+,19,20 Ag15Ti+, Ag14V+,16 and Au15Ti+,17 were previously reported by Lievens and co-workers. Actually, we carried out density functional theory (DFT) calculations of CunTi+ (n = 4−16) using the BPW91 functional21,22 with the LanL2DZ basis set in the GAUSSIAN 09 program package.23 First, the structures of Cu12Ti+ were explored in the spin multiplicities (2S + 1) of 2, 4, and 6, and three motifs of geometries (icosahedron, hexagonal antiprism, and bicapped pentagonal antiprism) were considered. Two Ti-centered structures of Cu12Ti+, 12-(a) and 12-(b) in Figure 8, with
adsorption energy by the DFT method. Initial structures were generated by attaching an NO molecule to possible sites of the most stable CunTi+ clusters. As shown in Figure 9, an NO
Figure 9. Structures of CunTi(NO)+ (n = 10−15) optimized by DFT(BPW91/LanL2DZ) calculations. The spin multiplicities and the adsorption energies of NO are also given.
molecule binds to a Ti−Cu−Cu hollow at n = 10 and the ontop site of the Ti atom for 11−13, whereas it attaches onto a hollow or bridge site of the Cu cage at n = 14 and 15. The NO adsorption energy of CunTi+ does not change significantly in n = 10−12 but decreases rapidly in n = 13−15 (Figure 9). This tendency is comparable to the size dependence of the measured total reaction cross section for CunTi+ + NO, which indicates that the higher reactivity toward NO can be attributed to the larger adsorption energy related to the geometric configuration of the Ti atom in the cluster. Even−odd alternations with the cluster size are clearly observed in the partial reaction cross sections for CunTi+ (Figures 3b and 4b). In the reaction with O2, the formation cross section of CuxTiO2+ is larger at x = odd than those of its neighboring sizes with x = even. This product ion may correspond to a combined species of (Cux+)(TiO2), and the positive charge is located in Cux which has smaller ionization energy than TiO2.24,25 The stability of (Cux+)(TiO2) can be estimated from that of Cux+ in the electronic shell model,26 and actually this species with x = odd is more abundant than that with x = even as shown in Figure 4b. Similarly, CuxTiNO+ should be expressed as (Cux−1+)(CuO)(TiN). Because the unit of Cux−1+ has an even number of valence electrons at x = even, (Cux−1+)(CuO)(TiN) with x = even are more stable than their neighbors with x = odd. Thus, the stability of CuxTiO2+ and CuxTiNO+ can be determined by the number of valence electrons in Cux+ and Cux−1+, respectively. In contrast, there is no even−odd alternation in the partial cross sections for the reactions of CunV+ with NO and O2 (Figures 5b and 6b). Probably, the consideration of the valence electrons in Cux+ of CuxVO2+ and in Cux−1+ of CuxVNO+ is invalid. Neither VO2 nor VN may be isolated in the clusters electronically because of the ambiguous valence states of a vanadium atom, as is common in its compounds. The total cross sections for the reactions of CunTi+ and CunV+ with NO are compared with those of Cun+1+ and CunAl+ obtained in our previous studies10,12,13 in Figure 10a. It is obvious that the substitution of a Ti or V atom drastically improves the reactivity toward NO and that the total cross sections of CunTi+ and CunV+ are larger than those of CunAl+ in the size range of n ≤ 12. On the contrary, CunTi+ and CunV+
Figure 8. Structures of CunTi+ (n = 4−16) optimized by DFT(BPW91/LanL2DZ) calculations. The spin multiplicities are also given.
comparable energies were obtained as the lowest-energy isomers among them. Next, we surveyed the lowest-energy structures of the larger clusters, CunTi+ (n ≥ 13), by using similar motifs. For n ≤ 11, copper atoms were removed from the Cu12Ti+ obtained, and the geometry optimization was performed. As shown in Figure 8, the Ti atom is located inside the Cu cage of the large clusters whereas it is exposed on the surface of the small clusters. In particular, the Ti atom almost occupies the center of the clusters at n = 12 and 14−16. Note that Cu12Ti+-(a) has an icosahedral geometry whereas the structure of Cu12Ti+-(b) is similar to those of Cu11Ti+ and Cu13Ti+. The first Cu shell around the Ti atom is completely filled in Cu15Ti+, and the next Cu atom begins to form an outer shell. The structural transition from a Ti-exposed geometry to Tiencapsulated one occurs at n = 12, though the Ti atom is incompletely encapsulated at n = 13 again. If the reactant molecule prefers to bind to the Ti atom, it is very likely that the size dependence of the reaction cross section is related to the position of the Ti atom in the cluster. Then, the geometric structures of CunTi+(NO) (n = 10−15) were calculated to obtain the most stable adsorption site for NO and the E
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Figure 11. Ratios of the total cross sections, σr(NO)/σr(O2), for the reactions of CunTi+ and CunV+ with NO to those with O2 as a function of n. The ratios obtained from our previous data for Cun+1+ (refs 10 and 13) and CunAl+ (ref 12) are also plotted for comparison. The collision energy is 0.2 eV. Figure 10. Comparison of the total cross sections for the reactions of CunTi+ and CunV+ with (a) NO and (b) O2 as a function of n. The results for Cun+1+ (refs 10 and 13) and CunAl+ (ref 12) are also plotted. The collision energy is 0.2 eV.
5. CONCLUSIONS We have studied the gas-phase reactions of CunTi+ and CunV+ with NO and O2 at a near-thermal energy. The comparison with the results of pure Cu clusters10,13 shows that the doping of a Ti or V atom improves the adsorption probability of NO molecules whereas it does not affect the O2 adsorption probability significantly. The preference for NO over O2 prevails in a wider size range, compared with the effect of the Al doping.12 The total cross sections for the reactions of CunTi+ and CunV+ are characterized by a gradual increase with the cluster size for NO (no significant change for O2) and the subsequent rapid drop at n ≈ 12. The size-dependent reactivity is mainly interpreted in terms of the geometric properties, specifically, the location of the dopant atom that acts as an effective adsorption site. Furthermore, it was found that the cluster dioxide is produced through the double-collision reactions with NO for an early transition-metal-doped copper cluster, Cu7Ti+, indicating that the decomposition of NO proceeds to release N2. The information obtained here would be useful for the rational design of Cu-based catalysts for the NOx reduction under lean burn conditions.
with n ≥ 13 have relatively small cross sections compared with the corresponding CunAl+. This difference can be explained by the position of the dopant atom; for large clusters, the Ti and V atoms are encapsulated while the Al atom stays on the surface.12 In addition, several Cu atoms often evaporate in the reactions of CunTi+ and CunV+ with NO, which gives contrast to those of Cun+1+ and CunAl+. The efficient Cu release implies that the NO adsorption energy increases considerably by the substitution of Ti or V. This increase is supported by our DFT calculations; the NO adsorption energies are calculated to be ∼3.1 eV for CunTi+ (n = 10−12), which are much larger than that for Cu10+ (1.39 eV).12 The reactivity of Cu clusters toward O2 is less affected than that toward NO by the substitution of Ti or V, except for the disappearance of the even−odd alternation (Figure 10b). However, there is a notable difference at n = 15; Cu15Ti+ is like noble metal and unreactive toward O2 at all whereas Cu16+ has some reactivity. This result can be explained in terms of the stable structure of Cu15Ti+ as mentioned above. A similar drastic decrease in reactivity by substitution and its interpretation were reported for the reactions of cationic and neutral Co12V clusters with H2.27,28 Figure 11 compares the ratio of the total cross sections for the reactions of CunX+ (X = Al, Ti, V, Cu) with NO to those with O2. The preference for NO over O2 is commonly observed by the substitution of Ti and V, whereas the pure Cu clusters prefer O2 over NO. In contrast, the preference of CunAl+ changes with the cluster size.12 Finally, it is worth noting that the production of cluster dioxide (i.e., the decomposition of NO) is observed in the multiple-collision reactions of Cu7Ti+ with NO. The NO decomposition efficiency can be roughly evaluated by comparing the relative intensities of the cluster dioxide ions at the same NO pressure (e.g., 1 × 10−3 Torr). It is found that the efficiency of Cu7Ti+ is higher than that of Cu9Al+ (ref 12) by about 1 order of magnitude. This result indicates that the activity of Cu clusters toward NO decomposition can be optimized by the selection of the dopant atom.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (M. Ichihashi). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Calculations were performed using the Fujitsu PRIMERGY RX300 S7 of the Research Center for Computational Science, Okazaki Research Facilities, National Institutes of Natural Sciences. This work was supported by the Special Cluster Research Project of Genesis Research Institute, Inc., and by JSPS KAKENHI Grant Number 25390004.
■
REFERENCES
(1) Heck, R. M.; Farrauto, R. J.; Gulati, S. T. Catalytic Air Pollution Control: Commercial Technology, 3rd ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2009. (2) Kašpar, J.; Fornasiero, P.; Hickey, N. Automotive Catalytic Converters: Current Status and Some Perspectives. Catal. Today 2003, 77, 419−449.
F
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A (3) Centi, G.; Perathoner, S. Nature of Active Species in CopperBased Catalysts and Their Chemistry of Transformation of Nitrogen Oxides. Appl. Catal., A 1995, 132, 179−259. (4) Deka, U.; Lezcano-Gonzalez, I.; Weckhuysen, B. M.; Beale, A. M. Local Environment and Nature of Cu Active Sites in Zeolite-Based Catalysts for the Selective Catalytic Reduction of NOx. ACS Catal. 2013, 3, 413−427. (5) Iwamoto, M.; Yokoo, S.; Sakai, K.; Kagawa, S. Catalytic Decomposition of Nitric Oxide over Copper(II)-Exchanged Y-type Zeolites. J. Chem. Soc., Faraday Trans. 1 1981, 77, 1629−1638. (6) Iwamoto, M.; Hamada, H. Removal of Nitrogen Monoxide from Exhaust Gases through Novel Catalytic Processes. Catal. Today 1991, 10, 57−71. (7) Iwamoto, M. Heterogeneous Catalysis for Removal of NO in Excess Oxygen - Progress in 1994. Catal. Today 1996, 29, 29−35. (8) Yamamoto, T.; Tanaka, T.; Kuma, R.; Suzuki, S.; Amano, F.; Shimooka, Y.; Kohno, Y.; Funabiki, T.; Yoshida, S. NO Reduction with CO in the Presence of O2 over Al2O3-Supported and Cu-Based Catalysts. Phys. Chem. Chem. Phys. 2002, 4, 2449−2458. (9) Holmgren, L.; Andersson, M.; Rosén, A. NO on Copper Clusters. Chem. Phys. Lett. 1998, 296, 167−172. (10) Hirabayashi, S.; Ichihashi, M. Reactions of Size-Selected Copper Cluster Cations and Anions with Nitric Oxide: Enhancement of Adsorption in Coadsorption with Oxygen. J. Phys. Chem. A 2014, 118, 1761−1768. (11) Hirabayashi, S.; Ichihashi, M. NO Decomposition Activated by Preadsorption of O2 onto Copper Cluster Anions. J. Phys. Chem. C 2015, 119, 10850−10855. (12) Hirabayashi, S.; Ichihashi, M. Stability of Aluminum-Doped Copper Cluster Cations and Their Reactivity toward NO and O2. J. Phys. Chem. A 2015, 119, 8557−8564. (13) Hirabayashi, S.; Ichihashi, M.; Kawazoe, Y.; Kondow, T. Comparison of Adsorption Probabilities of O2 and CO on Copper Cluster Cations and Anions. J. Phys. Chem. A 2012, 116, 8799−8806. (14) Ichihashi, M.; Hanmura, T.; Yadav, R. T.; Kondow, T. Adsorption and Reaction of Methanol Molecule on Nickel Cluster Ions, Nin+ (n = 3−11). J. Phys. Chem. A 2000, 104, 11885−11890. (15) Gioumousis, G.; Stevenson, D. P. Reactions of Gaseous Molecule Ions with Gaseous Molecules. V. Theory. J. Chem. Phys. 1958, 29, 294−299. (16) Janssens, E.; Neukermans, S.; Nguyen, H. M. T.; Nguyen, M. T.; Lievens, P. Quenching of the Magnetic Moment of a Transition Metal Dopant in Silver Clusters. Phys. Rev. Lett. 2005, 94, 113401. (17) Neukermans, S.; Janssens, E.; Tanaka, H.; Silverans, R. E.; Lievens, P. Element- and Size-Dependent Electron Delocalization in AuNX+ Clusters (X = Sc, Ti, V, Cr, Mn, Fe, Co, Ni). Phys. Rev. Lett. 2003, 90, 033401. (18) Janssens, E.; Neukermans, S.; Lievens, P. Sells of Electrons in Metal Doped Simple Metal Clusters. Curr. Opin. Solid State Mater. Sci. 2004, 8, 185−193. (19) Veldeman, N.; Höltzl, T.; Neukermans, S.; Veszprémi, T.; Nguyen, M. T.; Lievens, P. Experimental Observation and Computational Identification of Sc@Cu16+, a Stable Dopant-Encapsulated Copper Cage. Phys. Rev. A: At., Mol., Opt. Phys. 2007, 76, 011201. (20) Höltzl, T.; Veldeman, N.; De Haeck, J.; Veszprémi, T.; Lievens, P.; Nguyen, M. T. Growth Mechanism and Chemical Bonding in Scandium-Doped Copper Clusters: Experimental and Theoretical Study in Concert. Chem. - Eur. J. 2009, 15, 3970−3982. (21) Becke, A. D. Density-Functional Exchange-Energy Approximation with Correct Asymptotic Behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (22) Perdew, J. P.; Wang, Y. Accurate and Simple Analytic Representation of the Electron-Gas Correlation Energy. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 45, 13244−13249. (23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. GAUSSIAN 09, Revision D.01; Gaussian Inc.: Wallingford, CT, 2013.
(24) Knickelbein, M. B. Electronic Shell Structure in the Ionization Potentials of Copper Clusters. Chem. Phys. Lett. 1992, 192, 129−134. (25) Hildenbrand, D. L. Mass Spectrometric Studies of the Thermochemistry of Gaseous TiO and TiO2. Chem. Phys. Lett. 1976, 44, 281−284. (26) de Heer, W. A. The Physics of Simple Metal Clusters: Experimental Aspects and Simple Models. Rev. Mod. Phys. 1993, 65, 611−676. (27) Nonose, S.; Sone, Y.; Onodera, K.; Sudo, S.; Kaya, K. Structure and Reactivity of Bimetallic ConVm Clusters. J. Phys. Chem. 1990, 94, 2744−2746. (28) Nakajima, A.; Kishi, T.; Sugioka, T.; Sone, Y.; Kaya, K. Structure and Reactivity of Bimetallic ConVm+ Cluster Ions. J. Phys. Chem. 1991, 95, 6833−6835.
G
DOI: 10.1021/acs.jpca.6b00206 J. Phys. Chem. A XXXX, XXX, XXX−XXX