25578
2006, 110, 25578-25581 Published on Web 11/30/2006
N+NO Reaction on Rh(111) Surfaces Studied with Fast Near-Edge X-ray Absorption Fine Structure Spectroscopy: Role of NO Dimer as an Extrinsic Precursor Ikuyo Nakai,†,§ Hiroshi Kondoh,*,† Toru Shimada,†,| Masanari Nagasaka,† Reona Yokota,†,⊥ Kenta Amemiya,†,# Hideo Orita,‡ and Toshiaki Ohta†,) Department of Chemistry, School of Science, The UniVersity of Tokyo, Tokyo 113-0033, Japan, National Institute of AdVanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8565, Japan, Institute for Molecular Science, Myodaiji, Okazaki 444-8585, Japan, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan, Sekisui Chemical CO., Ltd., Tsukuba, Ibaraki 300-4292 Japan, High Energy Accelerator Research Organization, and SR Center, Ritsumeikan UniVersity, 1-1-1 Noji-Higashi, Kusatsu, Shiga 525-8577, Japan ReceiVed: October 7, 2006; In Final Form: NoVember 13, 2006
We studied the mechanism of the N+NO reaction on Rh(111) surfaces by means of fast near-edge X-ray absorption fine structure spectroscopy. This reaction is important as a basis of NOx reduction reactions on platinum-group metal surfaces. Atomic nitrogen layers on Rh(111) were titrated with NO at various temperatures. N2O is exclusively formed and desorbs into the gas phase below 350 K. The consumption rate of atomic nitrogen exhibits strange temperature dependence between 100 and 350 K; the reaction proceeds slower with increasing temperature. Reaction kinetics analyses and isotope-controlled experiments have revealed that the surface N atoms do not react with chemisorbed NO molecules but with NO dimers weakly bound on top of the chemisorbed layer, which play a role as an extrinsic precursor. The present results may support the possibility that NO dimers participate in various NO-related synthetic, biochemical, and surface reactions as an intermediate.
The adsorption behavior of NO and the NO+CO reaction on Rh(111) surfaces have been extensively studied as model processes of the NOx reduction reaction.1,2 NO molecules dissociatively adsorb on Rh(111) above 350 K. The NO+CO reaction proceeds in a steady state only above 350 K since the NO dissociation is necessary for the progress of the reaction.1,2 One of the major products in the whole catalytic NO+CO reaction is N2. Zaera et al. proved that the N2 molecules are not formed by association of atomic N but via formation and decomposition of N2O on Rh(111).2 Though the reaction mechanism has been investigated at high temperatures where NO dissociation proceeds in a steady state, this reaction is expected to occur even at lower temperatures. In order to understand the N+NO reaction comprehensively, it is necessary to investigate it over a wide temperature range including low temperatures where no dissociation of NO takes place. For this purpose, the mechanistic study of N+NO reaction on artificially prepared N-covered Rh surfaces is a promising approach. So far the number of surface science studies on the N+NO reaction using N-precovered Rh surfaces has been quite limited.3 * Corresponding author. Tel&Fax: 81-3-5841-4418; E-mail: kondo@ chem.s.u-tokyo.ac.jp † Department of Chemistry, School of Science, The University of Tokyo. ‡ National Institute of Advanced Industrial Science and Technology (AIST). § Institute for Molecular Science. | National Institute for Materials Science. ⊥ Sekisui Chemical Co., Ltd. # High Energy Accelerator Research Organization. ) Ritsumeikan University.
10.1021/jp0665933 CCC: $33.50
Here we studied the N+NO reaction on N-precovered Rh(111) surfaces by monitoring with fast near-edge X-ray absorption fine structure (NEXAFS) spectroscopy. It was revealed that transiently trapped NO dimer (NO)2 in an extrinsic precursor state plays an important role in the reaction below 350 K. The NO dimer has been assumed to be an essential intermediate in various NO-related synthetic, biochemical, and surface reactions.4-7 Since the NO monomer is relatively inactive, it is important to understand the role of the dimer in the complicated reaction mechanisms. Recently, the contribution of the NO dimer to NO-induced oxidation and nitrosation reactions was proposed by molecular orbital calculations.4 However, NO dimers have not been directly detected under reaction conditions. Spectroscopic observation of the NO dimer as a reaction intermediate will help the understanding of the mechanisms of NO-related reactions. The experiments were carried out at BL-7A of the Photon Factory (Tsukuba, Japan). The real-time monitoring of the reaction was conducted with the fast NEXAFS on the basis of the energy dispersive technique.8 The atomic nitrogen covered surface was prepared on a clean Rh(111) surface following the previously reported procedure: first molecularly adsorbed NO is thermally dissociated into atomic N and O, and then the atomic O was removed by dosing H2.9 The partially N-covered surfaces were exposed to NO gas at a constant pressure of 5.0 × 10-9 Torr at constant surface temperatures ranging from 70 to 480 K. During the NO exposure, N K-edge NEXAFS spectra were obtained every 13 s. © 2006 American Chemical Society
Letters
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25579 Arrhenius plot (inset of Figure 1) to be a negative value of -0.04 eV. Figures 2a and 2b show time evolution of adsorbate coverages and an example of NEXAFS spectra, respectively, taken at 120 K. Atomic N and molecular NO were observed on the surface during the reaction in this temperature range. When the gas-phase species were monitored with a quadrupole mass spectrometer (QMS) during the NEXAFS measurement, only N2O was observed accompanying consumption of the surface atomic N. Decomposition of N2O into N2 and O, which was observed in the reaction above 350 K, was not observed at all. These results indicate that the total reaction in the temperature range of 100-350 K can be regarded as
N(a) + NO(a) f N2O(g) Figure 1. Temperature dependence of the initial reaction rate (V0) at the NO pressure of 5 × 10-9 Torr. The reaction rates are normalized with the initial N coverage (θ0) to exclude the effect of variety in the initial coverage of N (in the range of 0.18-0.22 ML). The solid line is guide for eyes. The inset is the Arrhenius plot for the reaction rate in the temperature range from 100 to 200 K.
Figure 2. Time evolution of the coverages of surface adsorbates during reactions at 120 K (a) and 70 K (c) and example of corresponding N-K NEXAFS spectra and curve-fitting analysis results at 120 K (b) and 70 K (d). The red, green and blue lines in the spectra show the fitting component for atomic N, chemisorbed NO and physisorbed NO dimer (NO)2, respectively, and the gray lines show the sums of them. For the experimental run at 70 K(c), the NO exposure was stopped at 3170 s. The small structure around 400 eV in (b) is due to inhomogeneous sensitivity of the imaging detector, which can be neglected in the curve-fitting analyses.
Consumption of atomic N by the reaction was observed during the NO exposure. Time evolution of coverages of surface species was obtained by the curve fitting analysis for each spectrum (examples of curve fitting are presented in Figures 2 b and d). The errors in the estimated coverages were below 5%. The reaction rate was estimated from the slope of the consumption curve of atomic N. The reaction rate (N consumption rate, namely) at the initial stage of the NO exposure was measured as a function of temperature as shown in Figure 1. The reaction-rate change is classified into three temperature regions: (1) below 100 K, (2) 100-350 K, and (3) above 350 K. The reaction above 350 K, which proceeds faster with increasing temperature, was revealed to be identical process with the one in the steady-state NO+CO reaction2 where N2 is formed via dissociation of transiently formed N2O. The detail of the reaction above 350 K will be presented elsewhere. In the temperature range from 100 to 350 K, the reaction proceeds slower with increasing temperature, which is opposite to the cases of usual elementary reaction steps. The activation energy was estimated from the slope of the
(1)
Negative activation energies have been so far observed for surface reactions where weakly adsorbed species, whose desorption rate is faster than its reaction rate, are involved.10,11 In Figure 2a, the consumption of N exhibits a delayed start at around 300 s where NO is accumulated up to almost the completion of the N+NO monolayer on the surface (induction period). This result suggests that the reactive NO species is an extrinsic-precursor, which is weakly bound on top of the alreadyformed chemisorbed monolayer.12 Since the desorption and reaction rate of the NO molecules trapped in the extrinsicprecursor state is faster than the adsorption rate, its residence time may be too short to be observed with NEXAFS. At sufficiently low temperatures, however, the desorption and reaction rate of the NO precursor should become slower than the adsorption rate. Furthermore, if the desorption rate is slower than the reaction rate, the temperature dependence of the total reaction rate (the consumption rate of atomic N) is expected to turn positive and the trapped precursors are expected to be observed on the surface during the reaction. The temperature dependence of the reaction rate is actually positive below 100 K (Figure 1). As shown in Figures 2c and 2d, a new species besides atomic N and molecular NO is observed on the surface below 100 K. The new species is attributed to NO dimer (NO)2, since the NEXAFS spectrum of the new species is almost the same as that of NO dimer.13 The formation of the NO dimer and the consumption of atomic N start simultaneously at around 400 s, though the start of N consumption is not clear due to the low consumption rate. These results suggest a possibility that the NO dimer formed on top of the chemisorbed layer acts as an extrinsic precursor. When the NO exposure was stopped during the reaction at 3170 s in Figure 2c, a part of the NO dimers decomposed and desorbed into the gas phase. However, the consumption of atomic N still continued by the reaction of remaining NO dimer and atomic N even after the stop of NO exposure. To confirm the validity of the extrinsic-precursor-mediated reaction model, we checked contribution of the precursor species to the reaction with an isotope-controlled experiment. First, the surface was completely precovered with 14N and chemisorbed 14NO. Then it was exposed to 15NO gas while monitoring the mass signals of the desorbing N2O with the QMS. The result is shown in Figure 3. 14N2O does not appear but 14N15NO is exclusively observed, which evidences that the extrinsicprecursor species in the second layer directly reacts with atomic N in the first layer. Note that no intermixing between the firstand second-layer NO species occurs under this condition. Chemisorbed NO does not react with N probably due to a strong N-NO repulsion in the first layer.14 From these experimental results, the reaction formula (1) is rewritten as follows:
25580 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Letters
Figure 3. Time evolution of mass signal intensity of m/e ) 45 (15N14NO) and m/e ) 44 (14N2O) when a Rh(111) surface completely covered with chemisorbed 14N and 14NO was exposed to 15NO gas (T ) 120 K, PNO ) 1 × 10-6 Torr). The small structure in the m/e ) 45 curve at around 5 s is ascribed to a drop of NO pressure.
2NO(g) a (NO)2 (precursor)
(2-1)
(NO)2 (precursor) + N(a) f N2O(g) + NO(a) (2-2) Negative activation energies have been sometimes interpreted in terms of the difference between the real activation energy of the reaction (Er) and that of the desorption of the weakly bound reactant (Ed): Er - Ed.10,11 If Er is smaller than Ed, a negative apparent activation energy is observed. Figure 4a shows the energy diagram obtained from the NEXAFS experiments. Ed was estimated from the rate of decomposition-desorption process of the NO dimer on top of a pure NO monolayer, and Er was obtained from the consumption rate of atomic N in the reaction below 100 K. The estimated energies are consistent with one another within their errors, supporting the proposed reaction mechanism (2-1 and 2-2). Surface reactions where extrinsic precursors directly participate are very rare.11 Although the microscopic reaction mechanism including NO dimers cannot be deduced from the present experimental results, we tentatively propose a reaction model based on the density functional theory calculations15 as illustrated in Figure 4b. The reaction proceeds via formation of a new N-N bond between an N atom and an NO dimer (ii). According to a recent theoretical study, NO dimer is electrophilic and easily forms a chemical bond with electron-rich species.4 Then, one of the NO units of the complex (ii) replaces the N atom bound to the surface concomitantly with elongation of the N-N bond of the dimer (iii). Finally, the N-N bond dissociates to release an N2O molecule with an NO molecule remaining on the surface (iv). Physisorbed NO monomers were not detected in the present NEXAFS observations.16 It should be noted that we cannot completely exclude the possibility that a small amount of NO monomers physisorbed on top of the chemisorbed monolayer and/or supplied by dissociation of dimers react with the atomic N. However, we assume that the NO monomers do not contribute primarily to the reaction; if they can react with N atoms, the reaction should immediately start to occur without any induction period after the NO exposure is started. That the reactivity of the NO dimer is higher than the monomer is presumably due to its electrophilicity4 and increased van der Waals interactions caused by the larger molecular weight. It was revealed that in the N+NO reaction on Rh(111) at 70-350 K, the NO dimer in the extrinsic precursor state directly reacts with atomic N to form N2O and NO. The practical catalytic NO reduction on Rh surfaces, however, proceeds at higher temperatures above 350 K. In the present study, the role of the extrinsic precursor was prominent at the lower temper-
Figure 4. (a) Experimentally obtained energy diagram for the reaction. The errors for Ed, Er, and Ea′ was 0.05, 0.02, and 0.01 eV, respectively. (b) Proposed reaction model based on the density functional theory (DFT) calculations.15
atures, but the extrinsic-precursor-mediated N2O formation may not be negligible, even under realistic conditions at the higher temperatures due to high NO pressures. In conclusion, we found from the real-time monitoring of the N+NO reaction on Rh(111) with the fast NEXAFS technique that an extrinsic precursor formed on the chemisorbed N+NO layer reacts with atomic N to form N2O. The precursor is identified as an NO dimer species with an N-N bond. The direct observation of the NO dimer contributing to the reaction may support the possibility that NO dimers are involved as a reaction precursor in many NO-related reaction systems. Acknowledgment. The present work is supported by the Grant-in-Aid for Scientific Research (KAKENHI) in Priority Area “Molecular Nano Dynamics” from MEXT. The present work has been performed under the approval of the Photon Factory Program Advisory Committee (PF PAC No.2001S2003 and 2004G-320). Supporting Information Available: Comparison between an N-K NEXAFS spectrum of the precursor (Figure S1) and that of NO dimer (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Zhdanov, V. P.; Kasemo, B. Surf. Sci. Rep. 1997, 29, 31 and references therein. (2) (a) Zaera, F.; Gopinath C. S. Chem. Phys. Lett. 2000, 332, 209. (b) Zaera, F.; Gopinath C. S. Phys. Chem. Chem. Phys. 2003, 5, 646. (3) Belton, D. N.; DiMaggio, C. L.; Schmieg, S. J.; Ng, K. Y. S. J. Catal. 1995, 157, 559. (4) Zhao, Y.-L.; Bartberger, M. D.; Goto, K.; Shimada, K.; Kawashima, T.; Houk, K. N. J. Am. Chem. Soc. 2005, 127, 7964. (5) (a) Lim, M. D.; Lorkovic, I. M.; Ford. P. C. Inorg. Chem. 2002, 41, 1026.(b) Goto, K.; Hino, Y.; Kawashima, T.; Kaminaga, M.; Yano, E.; Yamamoto, G.; Takagi, N.; Nagase, S. Tetrahedron Lett. 2000, 41, 8479. (6) Nitric Oxide, Biology and Pathobiology; Ignarro, L. J. Ed.; Academic Press: San Diego, 2000. (7) Brown, W. A.; Gardner, P.; King, D. A. J. Phys. Chem. 1995, 99, 7065. (8) Amemiya, K; Kondoh, H.; Yokoyama, T.; Ohta, T. J. Electron. Spectrosc. Relat. Phenom. 2002, 124, 151. (9) van Hardeveld, R. M.; van Santen, R. A.; Niemantsverdriet, J. W. J. Phys. Chem. B 1997, 101, 998. (10) (a) Ertl, G.; Lee, S. B.; Weiss, M. Surf. Sci. 1982, 114, 515. (b) Elliott, A. J.; Hadden, R. A.; Tabatabaei, J.; Waugh, K. C.; Zemicael, F. W. J. Catal. 1995, 157, 153. (11) Celio, H.; Scheer, K. C.; White, J. M. J. Am. Chem. Soc. 2001, 123, 2990. (12) Kang, H. C.; Weinberg, W. H. Surf. Sci. 1994, 299/300, 755.
Letters (13) (a) Pe´rez-Jigato, M.; Termath, V.; Gardner, P.; Handy, N. C.; King, D. A, Rassias, S.; Surman, M. Mol. Phys. 1995, 85, 619. (b) Supporting Information. (14) DFT calculations15 indicated that the chemisorbed NO molecules cannot react with the atomic N due to a high activation barrier. They react with the N atoms above 350 K as their diffusion becomes active at the high temperatures. (15) We conducted DFT calculations to check this reaction model qualitatively because NO dimer is one of the molecules that are not treated explicitly within the state-of-the-art DFT methods.17 The detailed calculation methods were described in the previous works.18,19 The DFT calculations were performed with the program package DMol3 in Materials Studio (version 2.2) of Accelrys Inc. A (2 × 2) surface unit cell of a slab of four layers’ thickness (including 16 Rh atoms) was used. The slab was repeated periodically with a 30 Å of vacuum region between the slabs. The adsorbates and the two top layers of metal surfaces were allowed to relax in all the
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25581 calculations without symmetry restriction. Transition state search was performed with synchronous transition methods to estimate activation.17,18 We have found that the reaction scheme is energetically reasonable also for the DFT calculations, although the absolute adsorption energies are overestimated within the tendency of the state-of-the-art DFT method. The negative activation energy is also confirmed. (16) The N-K NEXAFS of the physisorbed NO monomer is expected to be similar to that of isolated NO in the gas phase. It exhibits a sharp N1s f π* peak at around 399 eV (Kosugi, N.; Adachi J.; Shigemasa E.; Yagisita. A. J. Chem. Phys. 1992, 97, 8842). Such a peak is not observed in the present spectra. (17) East, A. L. L. J. Chem. Phys. 1998, 109, 2185. (18) Orita, H.; Nakamura, I.; Fujitani, T. Surf. Sci. 2004, 571, 102. (19) Orita, H.; Nakamura, I.; Fujitani, T. J. Chem. Phys. 2005, 122, 014703.
