Interaction and Reaction of Coadsorbed NO and CO on a Rh (100

Jun 8, 2010 - Schuit Institute of Catalysis, Eindhoven University of Technology, Postbox 513,. 5600 MB Eindhoven, The Netherlands. Received April 6, 2...
0 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Interaction and Reaction of Coadsorbed NO and CO on a Rh(100) Single Crystal Surface† Maarten M. M. Jansen,* Oguz Caniaz, Ben E. Nieuwenhuys, and J. W. (Hans) Niemantsverdriet Schuit Institute of Catalysis, Eindhoven University of Technology, Postbox 513, 5600 MB Eindhoven, The Netherlands Received April 6, 2010. Revised Manuscript Received May 20, 2010 In order to assess the possibility to follow surface reactions in a quantitative way by vibrational spectroscopy, a combination of temperature programmed reaction spectroscopy (TPRS) and reflection absorption infrared spectroscopy (RAIRS) has been used to study the decomposition of NO and the reaction between NO and CO on Rh(100). NO adsorbs in two configurations: in an almost parallel position at coverages below 0.18 ML and, in addition, in an upright position, probably on a bridge site, at all coverages. Coadsorbing NO and CO has only a minor influence on NO binding, whereas CO shifts gradually from top toward the bridge site under the influence of NO. Combining TP-RAIRS with TPRS during the reaction between CO and NO enabled us to simultaneously study site occupation and obtain qualitative surface coverages and desorption rates. At low surface coverages, NO dissociation is observed at lower temperatures than CO2 formation. Near saturation, NO dissociation becomes blocked and shifts up in temperature. NO dissociation occurs simultaneously with CO2 formation. To decompose NO, free surface sites have to be generated through surface diffusion or desorption of some CO. During NO decomposition, the formed oxygen atoms react with CO to form CO2, creating more empty sites. This may lead to an explosive surface reaction.

Introduction In heterogeneous catalysis, molecules adsorb, react on, and subsequently desorb from surfaces. The catalytic cycle consists of a series of elementary reactions steps of which many occur on the surface. The pioneering work of Gabor Somorjai and co-workers provides many excellent examples of how catalytic reaction mechanisms can be unraveled with the help of surface science.1 Only a few techniques are capable of following elementary reactions on the surface in real-time. In the past, our group has used static secondary ion mass spectroscopy (SSIMS) for this purpose.2-4 However, this technique may interfere with the reaction because it could damage the surface by creating defects. Here, we explore to what extent reflection absorption infrared spectroscopy (RAIRS) is suitable to this end. † Part of the Molecular Surface Chemistry and Its Applications special issue. *To whom correspondence should be addressed. Room number: STW3.59. Telephone: þ31 40 247 4658. Fax: þ31 40 247 3481. E-mail: m.m.m.jansen@ tue.nl.

(1) Somorjai, G. A. Introduction to Surface Chemistry and Catalysis; Wiley: New York, 1994. (2) Borg, H. J.; Reijerse, J. F. C. J. M.; Van Santen, R. A.; Niemantsverdriet, J. W. The dissociation kinetics of NO on Rh(111) as studied by temperature programmed static secondary ion mass spectrometry and desorption. J. Chem. Phys. 1994, 101 (11), 10052-10063. (3) Hopstaken, M. J. P.; Niemantsverdriet, J. W. Lateral Interactions in the Dissociation Kinetics of NO on Rh(100). J. Phys. Chem. B 2000, 104 (14), 3058-3066. (4) Hopstaken, M. J. P.; Niemantsverdriet, J. W. Reaction between NO and CO on rhodium (100). How lateral interactions lead to auto-accelerating kinetics. J. Vac. Sci. Technol. 2000, 18, 1503-1508. (5) Nieuwenhuys, B. E. The surface science approach toward understanding automotive exhaust conversion catalysis at the atomic level. Adv. Catal. 1999, 44, 259-328. (6) Taylor, K. C. Nitric Oxide Catalysis in Automotive Exhaust Systems. Catal. Rev. Sci. Eng. 1993, 35 (4), 457-481. (7) Requejo, F. G.; Hebenstreit, E. L. D.; Ogletree, D. F.; Salmeron, M. An in situ XPS study of site competition between CO and NO on Rh(111) in equilibrium with the gas phase. J. Catal. 2004, 226 (1), 83-87.

Langmuir 2010, 26(21), 16239–16245

The reaction and interaction between CO and NO on rhodium surfaces is important for automotive exhaust gas catalysis.5,6 Hence, the reaction between NO and CO has extensively been (8) Gopinath, C. S.; Zaera, F. A Molecular Beam Study of the Kinetics of the Catalytic Reduction of NO by CO on Rh(111) Single-Crystal Surfaces. J. Catal. 1999, 186 (2), 387-404. (9) Permana, H.; Ng, K. Y. S.; Peden, C. H. F.; Schmeig, S. J.; Lambert, D. K.; Belton, D. N. Adsorbed species and reaction rates for NO-CO over Rh(111). J. Catal. 1996, 164 (1), 194-206. (10) Root, T. W.; Schmidt, L. D.; Fisher, G. B. Nitric oxide reduction by carbon monoxide on rhodium(111): temperature programmed reaction. Surf. Sci. 1985, 150 (1), 173-192. (11) Root, T. W.; Fisher, G. B.; Schmidt, L. D. Electron energy loss characterization of nitric oxide on rhodium(111). II. Coadsorption with oxygen and carbon monoxide. J. Chem. Phys. 1986, 85 (8), 4687-4695. (12) Hopstaken, M. J. P.; Van Gennip, W. J. H.; Niemantsverdriet, J. W. Reactions between NO and CO on rhodium (111): an elementary step approach. Surf. Sci. 1999, 433-435, 69-73. (13) Oh, S. H.; Fisher, G. B.; Carpenter, J. E.; Goodman, D. W. Comparative kinetic studies of carbon monoxide-oxygen and carbon monoxide-nitric oxide reactions over single crystal and supported rhodium catalysts. J. Catal. 1986, 100 (2), 360-376. (14) Peden, C. H. F.; Goodman, D. W.; Blair, D. S.; Berlowitz, P. J.; Fisher, G. B.; Oh, S. H. Kinetics of carbon monoxide oxidation by oxygen or nitric oxide on rhodium(111) and rhodium(100) single crystals. J. Phys. Chem. 1988, 92 (6), 1563-1567. (15) Schwartz, S. B.; Schmidt, L. D.; Fisher, G. B. Carbon monoxide þ oxygen reaction on rhodium(III): steady-state rates and adsorbate coverages. J. Phys. Chem. 1986, 90 (23), 6194-6200. (16) Belton, D. N.; Schmieg, S. J. Oxidation of carbon monoxide by nitric oxide over rhodium(111). J. Catal. 1993, 144 (1), 9-15. (17) Peden, C. H. F.; Belton, D. N.; Schmieg, S. J. Structure sensitive selectivity of the NO-CO reaction over Rh(110) and Rh(111). J. Catal. 1995, 155 (2), 204-218. (18) Permana, H.; Ng, K. Y. S.; Peden, C. H. F.; Schmieg, S. J.; Belton, D. N. Effect of NO Pressure on the Reaction of NO and CO over Rh(111). J. Phys. Chem. 1995, 99 (44), 16344-16350. (19) Herman, G. S.; Peden, C. H. F.; Schmieg, S. J.; Belton, D. N. A comparison of the NO-CO reaction over Rh(100), Rh(110) and Rh(111). Catal. Lett. 1999, 62 (2-4), 131-138. (20) Zhdanov, V. P.; Kasemo, B. Mechanism and kinetics of the NO-CO reaction on Rh. Surf. Sci. Rep. 1997, 29 (2), 31-90. (21) Zaera, F.; Gopinath, C. S. Role of adsorbed nitrogen in the catalytic reduction of NO on rhodium surfaces. J. Chem. Phys. 1999, 111 (17), 80888097.

Published on Web 06/08/2010

DOI: 10.1021/la1013544

16239

Article

studied on various single crystal surfaces of rhodium: (111),7-24 (100),4,14,19,25,26 (110),17,19,27-32 and (331).33 Under steady-state conditions, the reaction between NO and CO proceeds via a generally accepted reaction mechanism, which includes the reversible adsorption of CO and NO, subsequent irreversible decomposition of NO forming nitrogen and oxygen atoms, and the reaction of oxygen with CO to form CO2.19,20 Atomic nitrogen recombines to form N2 at mild pressure conditions, while elevated pressure conditions also lead to recombination with NO to form N2O.19,20 NO decomposition and CO2 formation are surface structure sensitive, with the activity of both reactions on the Rh(100) surface lying in between that of the highly active Rh(110) and least reactive Rh(111) surface.19 Hopstaken and Niemantsverdriet4 studied the reaction between CO and NO on Rh(100) under ultrahigh vacuum (UHV) conditions by a combination of temperature programmed reaction spectroscopy (TPRS) and TP-SSIMS. The most crucial step is the decomposition of NO to form atomic N and O. At surface concentrations below 0.20 ML, this happens readily at 200 K. When the NO coverage increases close to saturation values, NO dissociation gradually shifts up in temperature to 450 K. Due to repulsive lateral interactions and the absence of free surface sites, NO decomposition becomes blocked. This in turn leads to an explosive surface reaction, triggered by desorption of a small fraction of CO. The free sites generated in this way are used by NO to dissociate. The formed oxygen atoms can in turn react with CO to CO2, which desorbs, and hence, more empty sites are created. In this way, the reaction becomes autoaccelerating. Although in general the kinetics of the reaction between NO and CO on Rh(100) is well documented, studies on how interactions between adsorbed reactants affect the kinetics are scarce. Several studies have been devoted to site occupation of CO and NO on Rh(111). Root et al.11 studied coadsorption of CO and NO at low temperatures and low pressures using electron energy loss spectroscopy (EELS). CO and NO both adsorb on top and (22) Nakamura, I.; Kobayashi, Y.; Hamada, H.; Fujitani, T. Adsorption behavior and reaction properties of NO and CO on Rh(111). Surf. Sci. 2006, 600 (16), 3235-3242. (23) Rider, K. B.; Hwang, K. S.; Salmeron, M.; Somorjai, G. A. High-Pressure (1 Torr) Scanning Tunneling Microscopy (STM) Study of the Coadsorption and Exchange of CO and NO on the Rh(111) Crystal Face. J. Am. Chem. Soc. 2002, 124 (19), 5588-5593. (24) Zaera, F.; Gopinath, C. S. Kinetics of NO reduction by CO on Rh(111): A molecular beam study. Stud. Surf. Sci. Catal. 2000, 130B (International Congress on Catalysis, 2000, Pt. B), 1295-1300. (25) Brandt, M.; M€uller, H.; Zagatta, G.; B€owering, N.; Heinzmann, U. Reaction of NO and CO on a Rh(100) surface studied with gas-phase oriented NO. Surf. Sci. 1996, 352-354, 290-294. (26) Hendershot, R. E.; Hansen, R. S. Reduction of nitric oxide with carbon monoxide on the rhodium(100) single-crystal surface. J. Catal. 1986, 98 (1), 150-165. (27) Rzeznicka, I. I.; Ma, Y.; Cao, G.; Matsushima, T. Removal pathways of surface nitrogen in a steady-state NO þ CO reaction on Pd(110) and Rh(110): Angular and velocity distribution studies. J. Phys. Chem. B 2004, 108 (38), 14232-14243. (28) Schmatloch, V.; Jirka, I.; Heinze, S.; Kruse, N. Towards an oscillation mechanism for the NO-CO reaction on Rh{110}: NO dissociation kinetics and oxygen subsurface diffusion. Surf. Sci. 1995, 331-333 (Pt. A), 23-29. (29) Baird, R. J.; Ku, R. C.; Wynblatt, P. The chemisorption of carbon monoxide and nitric oxide on rhodium (110). Surf. Sci. 1980, 97 (2-3), 346-362. (30) Schmatloch, V.; Kruse, N. Adsorption and reaction of carbon monoxide, nitric oxide and mixtures of CO/NO on rhodium{110}. Surf. Sci. 1992, 269-270 (Pt. A), 488-494. (31) Baraldi, A.; Dhanak, V. R.; Comelli, G.; Kiskinova, M.; Rosei, R. Nitric oxide dissociation and NO þ carbon monoxide reaction on rhodium(110): influence of the surface structure and composition on the reaction rates. Appl. Surf. Sci. 1993, 68 (3), 395-405. (32) Matsushima, T.; Ma, Y.; Nakagoe, O. Spatial and velocity distributions of desorbing products in steady-state NO þ CO and N2O þ CO reactions on Pd(110) and Rh(110). e-J. Surf. Sci. Nanotechnol. 2006, 4, 593-601. (33) Dubois, L. H.; Hansma, P. K.; Somorjai, G. A., Evidence for an oxygen intermediate in the catalytic reduction of nitric oxide by carbon monoxide on rhodium surfaces. J. Catal. 1980, 65 (2), 318-327.

16240 DOI: 10.1021/la1013544

Jansen et al.

bridge sites. Adding CO to a NO precovered surface reduced the elastic peak intensity, which indicates a well-mixed but disordered adlayer. Addition of NO to a CO precovered surface resulted in an increase of the CO stretching frequency. The latter was explained by surface segregation, which increases the local CO coverage. At higher local CO coverage, dipole-dipole coupling within the CO islands becomes stronger which results in an upward shift in frequency. In later studies, it was shown that, beside the top sites, CO and NO occupy 3-fold hollow sites and not bridge sites on Rh(111).7,22,23 Nakamura et al.22 have reported that under similar conditions NO first occupies the fcc site, then the top site, and last the hcp site, while CO adsorbs initially onto top and then onto 3-fold hollow sites. Preadsorbed top CO blocks the hcp site for NO adsorption, and near CO saturation NO adsorption is completely blocked. NO is also able to block CO: under equilibrium conditions, NO first displaces CO on hollow sites and completely blocks CO adsorption at increasing NO pressures.7,23 On Rh(100), site occupation of NO and CO has been studied separately. Up to 0.50 ML, CO occupies mainly the top site.34-38 At higher coverages, CO is observed on both top and bridge sites due to the formation of dense overlayer structures.34-36,38-40 NO occupies two kinds of sites. At low coverages, NO is adsorbed almost parallel to the surface with the bond axis over a hollow site.41-44 At intermediate and high coverages, NO is present exclusively in an upright position, most probably on a bridge site.42-44 To the best of our knowledge, there is no study on Rh(100) that describes the interaction between CO and NO and how this influences site occupation. The purpose of this paper is to discuss the interaction between NO and CO on Rh(100) and to unravel how reaction kinetics is influenced by changes in adsorption sites. First, RAIRS and TPRS are used to study the adsorption and decomposition of NO. Second, changes in bonding site upon coadsorbing NO and CO are investigated with RAIRS. Next, the influence of NO decomposition on CO2 formation is investigated in more detail. Finally, (34) Baraldi, A.; Gregoratti, L.; Comelli, G.; Dhanak, V. R.; Kiskinova, M.; Rosei, R. CO adsorption and CO oxidation on Rh(100). Appl. Surf. Sci. 1996, 99 (1), 1-8. (35) Gurney, B. A.; Richter, L. J.; Villarrubia, J. S.; Ho, W. The populations of bridge and top site carbon monoxide on rhodium(100) vs. coverage, temperature, and during reaction with atomic oxygen. J. Chem. Phys. 1987, 87 (11), 6710-6721. (36) De Jong, A. M..; Niemantsverdriet, J. W. The adsorption of CO on Rh(100): reflection absorption infrared spectroscopy, low energy electron diffraction, and thermal desorption spectroscopy. J. Chem. Phys. 1994, 101 (11), 10126-10133. (37) Richter, L. J.; Gurney, B. A.; Ho, W. The influence of adsorbate-adsorbate interactions on surface structure: the coadsorption of carbon monoxide and molecular hydrogen on rhodium(100). J. Chem. Phys. 1987, 86 (1), 477-490. (38) Strisland, F.; Ramsted, A.; Ramsvik, T.; Borg, A. CO adsorption on the Rh(100) surface studied by high resolution photoelectron spectroscopy. Surf. Sci. 1998, 415 (3), L1020-L1026. (39) Castner, D. G.; Sexton, B. A.; Somorjai, G. A. LEED and thermal desorption studies of small molecules (hydrogen, oxygen, carbon monoxide, carbon dioxide, nitric oxide, ethene, ethyne and carbon) chemisorbed on the rhodium (111) and (100) surfaces. Surf. Sci. 1978, 71 (3), 519-540. (40) Jansen, M. M. M.; Nieuwenhuys, B. E.; Curulla-Ferre, D.; Niemantsverdriet, J. W. Influence of Nitrogen Atoms on the Adsorption of CO on a Rh(100) Single Crystal Surface. J. Phys. Chem. C 2009, 113 (28), 12277-12285. (41) Bondino, F.; Comelli, G.; Baraldi, A.; Vesselli, E.; Rosei, R.; Goldoni, A.; Lizzit, S. NO adsorption on Rh(100). Part 2. Stability of the adlayers. J. Chem. Phys. 2003, 119 (23), 12534-12539. (42) Bondino, F.; Comelli, G.; Baraldi, A.; Vesselli, E.; Rosei, R.; Goldoni, A.; Lizzit, S.; Bungaro, C.; Gironcoli, S. d.; Baroni, S. NO adsorption on Rh(100). Part 1. Structural characterization of the adlayers. J. Chem. Phys. 2003, 119 (23), 12525-12533. (43) Loffreda, D.; Delbecq, F.; Simon, D.; Sautet, P. Breaking the NO bond on Rh, Pd, and Pd3Mn alloy (100) surfaces: A quantum chemical comparison of reaction paths. J. Chem. Phys. 2001, 115 (17), 8101-8111. (44) Popa, C.; Van Bavel, A. P.; Van Santen, R. A..; Flipse, C. F. J.; Jansen, A. P. J. Density functional theory study of NO on the Rh(100) surface. Surf. Sci. 2008, 602 (13), 2189-2196.

Langmuir 2010, 26(21), 16239–16245

Jansen et al.

Article

Figure 1. RAIRS spectrum of 0.15 ML NO on Rh(100) at 120 K.

we use a combination of TP-RAIRS and TPRS, under lowpressure reaction conditions, to monitor the formation of gas phase products and changes in CO and NO binding sites.

Figure 2. RAIRS and TPRS spectra of various amounts of NO on Rh(100). NO was adsorbed at 120 K. For TPRS, the surface was heated with 5 K/s. The course of the NO stretching frequency with NO coverage is plotted as an inset in (A).

Experimental Details

Results

The experiments were carried out in a home-built, two-stage stainless steel ultrahigh vacuum (UHV) system with a base pressure of 1.510-10 mbar. The system is equipped with a quadrupole mass spectrometer (Prisma QME200, Balzers) for TPRS and a Fourier transform infrared spectrometer (Galaxy 4020, Mattson) with a mercury cadmium telluride (MCT) detector (spectral range of 4000-800 cm-1) flushed with dry nitrogen for RAIRS. All RAIRS spectra consist of 512 scans taken at 4 cm-1 spectral resolution divided by a stored background spectrum of a clean surface. For TP-RAIRS, 32 scans are collected to decrease measurement time. The rhodium single crystal of (100) orientation was mounted by two tantalum wires allowing for resistive heating to 1400 K and cooling to 88 K with a continuous flow of liquid nitrogen. Temperatures were measured using a chromel-alumel thermocouple spot-welded to the back of the crystal. Crystal cleaning was done by cycles of argon ion sputtering (6 μA/cm2) at 920 K and annealing in 10-7 mbar of 20% oxygen in argon at temperatures of 900 to 1100 K and a final flash to 1400 K. Carbon monoxide (Linde Gas, 99.997% pure) and nitrogen monoxide (Linde Gas, 99.5% pure) were used without further purification. TPRS experiments were done to determine the initial NO and CO coverages. The NO coverage was obtained from the areas under the N2 and NO desorption traces. The N2 area was compared with the area of a saturated nitrogen adlayer: 0.35 ML atomic N.40 Here, the nitrogen layer was made by decomposing NO at 400 K and simultaneous removal of atomic O with CO. The saturation coverage of NO is 0.65 ML.45,46 At NO saturation, the area under the NO desorption trace is 0.42 ML: 0.65 ML (the saturation coverage of NO) minus 0.23 ML (the amount of NO that can decompose in a TPRS experiment). The initial CO coverage was obtained from the CO2 and CO desorption traces. For CO desorption, the CO saturation coverage of 0.83 ML34,36,39 was used as reference. For the area under the CO2 traces, the amount of oxygen from decomposed NO was used as reference, where we assume that atomic oxygen and CO react with 100% efficiency.47

NO on Rh(100). In this section, we discuss adsorption and dissociation of NO on Rh(100) without any coadsorbates. RAIRS is used to investigate how NO binds to the surface, while TPRS is used to study its decomposition. Figures 1 and 2A show RAIRS spectra of NO adsorbed on Rh(100). The surface was exposed to 10-9 mbar NO at 120 K followed by RAIRS and subsequently TPRS for quantification. IR absorption by molecular NO is observed in two frequency regions: a weak absorption band at 933 cm-1 and an intense band around 1600-1700 cm-1. The first feature only appears at coverages below 0.18 ML, but it is not always detected. The noise level in this frequency range is near the detection limit of the IR-MCT detector and sometimes larger than the NO feature. This feature has also been observed previously in EELS experiments.46,48,49 On the basis of a combination of density functional theory (DFT) and angle dependent near edge X-ray absorption fine structure (NEXAFS),41-44 this band was assigned to an almost parallel NO species situated above the hollow site as indicated schematically in Figure 1. DFT calculations suggest that NO in this position acts as an intermediate in NO dissociation.43 The most direct experimental evidence for this suggestion has come from synchrotron X-ray photoelectron spectroscopy (XPS).41 In the high intensity X-ray beam, the inclined species showed decomposition within minutes while upright NO stayed completely intact. Because this species is oriented almost parallel to the surface, it has a small dipole moment perpendicular to the surface, making this species difficult to detect with RAIRS. DFT calculations have shown a decrease in absorption intensity of the inclined species with respect to an upright bridge species of about 8 times due to the decrease in dipole moment.44 The second band around 1600-1700 cm-1 shows a strong frequency shift with increasing coverage attributed to dipoledipole coupling between NO molecules and chemical effects near saturation. Around 0.23-0.27 ML, there are actually two absorption bands. IR absorption in the region between 1600-1700 cm-1

(45) Ho, P.; White, J. M. Adsorption of nitric oxide on rhodium(100). Surf. Sci. 1984, 137 (1), 103-116. (46) Villarrubia, J. S.; Ho, W. Nitric oxide adsorption, decomposition, and desorption on rhodium(100). J. Chem. Phys. 1987, 87 (1), 750-764. (47) Hopstaken, M. J. P.; Niemantsverdriet, J. W. Structure sensitivity in the CO oxidation on rhodium: Effect of adsorbate coverages on oxidation kinetics on Rh(100) and Rh(111). J. Chem. Phys. 2000, 113 (13), 5457-5465.

Langmuir 2010, 26(21), 16239–16245

(48) Van Bavel, A. P. Understanding and quantifying interactions between adsorbates carbon monoxide, nitric oxide and nitrogen- and oxygen-atoms on rhodium. Ph.D. Thesis, 2005. (49) Villarrubia, J. S.; Richter, L. J.; Gurney, B. A.; Ho, W. Observation of significant nitrogen-oxygen bond weakening in nitric oxide on rhodium(100). J. Vac. Sci. Technol. A 1986, 4, 1487-1490.

DOI: 10.1021/la1013544

16241

Article

Jansen et al.

Figure 3. RAIRS spectra of various amounts of CO on NO-precovered Rh(100) with (a) θNO = 0 ML, (b) θNO = 0.16 ML, (c) θNO = 0.23 ML, and (d) θNO = 0.36 ML. Both NO and CO were adsorbed at 150 K.

is assigned to NO in an upright position.42,43 DFT calculations suggest that NO is adsorbed on a bridge site.42-44 Splitting of this band was not detected previously,46,48,49 probably due to the inherent lower resolution of EELS. We tentatively assign this phenomenon to adlayer ordering: the NO molecules are almost randomly ordered at√low coverage, while at higher coverage NO √ can order into a p(4 2  2)R45° structure.42 Figure 2B and C show TPRS spectra of NO desorption and N2 recombination with subsequent desorption. Two regimes are present with the first occurring up to about 0.23 ML NO. Here, molecular NO desorption is absent and the amount of desorbing N2 increases proportional to the NO coverage. In the second regime above 0.23 ML, the amount of desorbing N2 stays constant, whereas the amount of molecular NO that desorbs around 430 K increases. In the first regime, all NO decomposes, forming atomic N and O. As (100) surfaces generally can contain a maximum of 0.50 ML atoms, which would correspond to 0.25 ML of NO, we tentatively assume that the first regime of complete dissociation ends at 0.25 ML. In the second regime, NO decomposition becomes hindered by the presence of the N and O atoms, resulting in desorption of molecular NO. Its decomposition is probably impeded due to the lack of free sites. The 0.25 ML of NO that can decompose is lower than the values reported in several TPRS studies: up to 0.34 ML NO is suggested to decompose.3,45,46,50 The difference is probably caused by the different quantification methods used. In earlier literature,45,46,50 a desorption feature coinciding with NO desorption at mass 28 was mistakenly interpreted as a low temperature N2 desorption feature, while actually it is caused by cracking of NO in the mass spectrometer, forming N2.3 A second factor introducing inaccuracy is the correction for mass spectrometer sensitivity. Removal rates differ for different gases in a UHV system: when a calibration gas mixture of NO and N2 is introduced, the partial pressure of N2 stabilizes very fast while this is not so for NO. To mimic a TPRS experiment, the mass spectrometer is best calibrated immediately after introducing the gas mixture, while normally stabilized values are used. This underestimates the amount of NO desorbing during TPD. To (50) Siokou, A.; Van Hardeveld, R. M.; Niemantsverdriet, J. W. Surface reactions of nitrogen oxide on rhodium(100), adsorption, dissociation and desorption. Surf. Sci. 1998, 402-404, 110-114.

16242 DOI: 10.1021/la1013544

exclude these effects, we have used a different reference for the nitrogen coverage as is explained in the experimental part. This reference was obtained by using XPS on a saturated nitrogen adlayer to determine the nitrogen coverage.37 To conclude, NO binds at low coverage in an almost parallel position and in an upright bridge position at all coverages. Below 0.25 ML, all NO decomposes during heating, while above 0.25 ML NO, NO decomposition becomes blocked in the presence of atomic N and O. CO on Rh(100). Between θCO = 0.35 and 0.50 ML, CO orders into a c(2  2) structure with CO occupying top sites.34-38 Above CO = 0.50 ML, the c(2  2) layer transforms into a √ θ√ p(4 2 2)R45° structure at θCO =0.75 ML with CO occupying both top and bridge sites.35,36,38 At CO saturation, CO orders into a c(6  2) pattern with CO on both top and bridge sites.34,36,39,40 Figure 3a shows how the different adlayer structures change the RAIRS spectra of CO on Rh(100). Two absorption bands are present in the regions 1870-1940 cm-1 and 2010-2080 cm-1. These bands are assigned to bridge and top CO, respectively.35,36 At coverages up to 0.50 ML, absorption mainly by CO on top sites is observed. The CO stretching frequency gradually shifts to higher values, due to an increase in dipole-dipole interactions between CO molecules. Between θCO = 0.50 and 0.75 ML, the absorption by CO on top sites decreases first while simultaneously an increase in absorption by bridged CO is observed. At saturation, CO occupies mainly top sites again. Changes in site occupation are mainly due to the presence of different adlayer ordering. At θCO = 0.75 ML, the intensity of all absorption bands has decreased with respect to θCO = 0.58 ML and saturation. Site occupation changes from mainly top at θCO = 0.50 ML to about 0.25 ML top and 0.50 ML bridge CO at θCO = 0.75 ML to predominantly top sites again at θCO = 0.83 ML. Due to dilution of CO on top sites in an √ adlayer √ with CO on bridge sites at θCO = 0.75 ML in the p(4 2  2)R45° structure, dipole-dipole coupling decreases significantly, causing absorption intensities of both top and bridge sites to decrease. CO and NO Bonding Sites: RAIRS. We first discuss the interaction between CO and NO on Rh(100) in the absence of reactions. Figure 3b, c, and d show the RAIRS spectra with, respectively, 0.16, 0.23, and 0.36 ML of NO and various amounts of CO on Rh(100). The crystal surface is first exposed to NO then to CO, followed by RAIRS and TPRS for quantification. Langmuir 2010, 26(21), 16239–16245

Jansen et al.

Figure 4. RAIRS spectra selected from Figure 3 at a constant CO coverage with (a) θCO ≈ 0.10 ML and (b) θCO ≈ 0.27 ML. The NO coverage varies within each panel.

As indicated, roughly three IR absorption bands are present originating from bridged NO, bridged CO and top CO. In general, the absorption intensity of bridged NO decreases in the presence of CO at all three NO coverages, while the NO stretching frequency increases by about 20-40 cm-1. However, at θNO = 0.16 ML, an increase in intensity of the bridged NO band is observed at increasing the CO coverage as shown in Figure 3 from 0 to 0.18 ML. When the NO adlayer becomes mixed with CO, dipole-dipole interactions between NO molecules decrease. This, in general, decreases the NO absorption intensity. We tentatively assign the increase in intensity observed at θNO = 0.16 ML to the initially tilted NO species switching to an upright position in the presence of CO. The increase in the NO stretching frequency is most pronounced when the adlayer becomes saturated. This shift in frequency is probably due to static or chemical effects which can be understood with the back-bonding model of Blyholder.51 High coverages may cause the adsorbates to compete for metal electrons, lowering back-donation into the antibonding orbitals of the adsorbate, which in turn increase molecular vibrational frequencies. Linke et al.52 reported that frequency shifts for CO on Rh(111) are mainly due to dipole-dipole interactions at low coverages, while chemical effects predominantly cause shifts when an adlayer becomes saturated. Figure 4 shows the RAIRS spectra of Figure 3 at a constant amount of CO with varying NO coverage. At increasing the NO coverage from θNO = 0 to 0.36 ML, the CO binding site changes gradually from top to bridge. This switch from top to bridge sites is due to lateral interactions. On the ground of these experiments, we cannot attribute this to chemical or geometrical factors, because the local ordering of the adlayer is not understood. To conclude, CO probably influences the way NO binds to the surface by switching the almost parallel species to upright NO. In the presence of NO, CO changes from top to bridge sites. CO2 Formation: TPRS. In TPRS, desorption of N2, CO, NO, and CO2 are observed. Due to the low pressure conditions used, no N2O is formed.19,20 Contribution of N2O to CO2 formation, both have mass 44 amu, can also be excluded by the absence of evolution of the cracking product N (mass 14 amu) in all (51) Blyholder, G. Molecular Orbital View of Chemisorbed Carbon Monoxide. J. Phys. Chem. 1964, 68 (10), 2772-2777. (52) Linke, R.; Curulla-Ferre, D.; Hopstaken, M. J. P.; Niemantsverdriet, J. W. CO/Rh(111): Vibrational frequency shifts and lateral interactions in adsorbate layers. J. Chem. Phys. 2001, 115 (17), 8209-8216.

Langmuir 2010, 26(21), 16239–16245

Article

spectra when CO2 is formed. Beside intensity changes, no significant influence of interactions between coadsorbates is visible in the N2, CO, and NO desorption products. Changing the NO and CO coverages does have an impact on the CO2 formation behavior as shown in Figure 5, making this reaction product more interesting for discussion. The intensity of some spectra has been adapted as indicated. Figure 5a, b, and c show the CO2 formation traces at, respectively, θNO =0.16, 0.23, and 0.36 ML and various amounts of CO. We note the following CO2 formation characteristics. At low coverage as in Figure 5a, CO2 forms between 400 and 460 K. In the zero coverage limit of both oxygen and CO, CO2 forms around 450 K.47 Hence, NO has probably decomposed at lower temperatures than 400 K, leaving the surface with a mixture of CO, O, and N when CO2 formation starts. In the absence of CO, NO already decomposes around 200 K.3 When the total coverage of NO and CO exceeds 0.50 ML, CO2 is formed in a very sharp peak around 370 K. CO2 formation is limited by the dissociation of NO, which becomes hindered on an occupied surface. First, CO has to desorb to create free sites for NO to dissociate.4 When NO dissociates, atomic oxygen is formed, which subsequentially reacts with CO to form CO2 and free sites. The free sites are again used for NO dissociation. This makes the reaction autoaccelerating and in fact explosive. At intermediate coverage, two CO2 formation channels are present around 350 and 420 K. The temperature of the latter feature is very similar to CO2 formation at low coverage and is probably only mildly influenced by lateral interactions. Lateral interactions between the adsorbates can for example destabilize CO and thereby accelerate the rate of CO2 formation. This has probably caused the shift to lower temperatures for the first feature. The strongest interactions can be expected from atomic species such as N and O which are formed during the dissociation of NO.

Discussion CO and NO Bonding Sites and Reactivity. Figure 3 gives information about the way in which CO and NO bind to the Rh(100) surface at 150 K, and Figure 5 shows at what temperatures CO2 is formed. To understand how CO and NO behave under conditions where reactions occur, a TP-RAIRS study has been done. Figure 6 shows the TP-RAIRS spectra collected while heating the surface with 1 K/s. The TPRS spectra were collected in a separate experiment with a heating rate of 5 K/s. To correct for the difference in heating rate between experiments, the temperature scale in the TPRS spectra has been recalculated. Changing heating rate from 5 to 1 K/s, one can show using the Redhead equation53 that desorption features shift down in temperature by about 5%. In TPRS, the desorption products are N2 around 750 K (as in Figure 2), CO, NO, and CO2. In TP-RAIRS, the absorption spectra and integrated intensities of bridged NO, bridged CO, and top CO are presented as a function of temperature. No additional features due to N2O, NCO, or CN are present in the TP-RAIRS spectra. If these species are formed, they are very short-lived. At low coverage as in Figure 6a, NO dissociates between 210 and 270 K as can be seen in the decline of the RAIRS intensity of bridged NO. The temperature at which NO dissociates increases with increasing the CO coverage to around 300-340 K (Figure 6b) and with increasing the NO coverage to about 310-360 K (Figure 6d). Our interpretation is that this hindering of NO dissociation is due to repulsive lateral interactions in Figure 6b and due to a lack of (53) Redhead, P. A. Thermal desorption of gases. Vacuum 1962, 12, 203-211.

DOI: 10.1021/la1013544

16243

Article

Jansen et al.

Figure 5. TPRS spectra of CO2 formation by reaction between various amounts of NO and CO on Rh(100). Both CO and NO were adsorbed at 150 K. The heating rate used was 5 K/s.

Figure 6. TP-RAIRS and TPRS spectra of various amounts of CO on NO precovered Rh(100). RAIRS and TPRS spectra were collected while heating the surface with, respectively, 1 and 5 K/s. The TPRS spectra with the integrated IR absorption intensities are shown as insets.

free sites in Figure 6d. The latter becomes clear from the desorption of CO and CO2 between 290 and 390 K. CO desorbs to create free sites which are immediately used for NO decomposition. Subsequently, the formed oxygen atoms react with CO to form CO2, creating more free sites. Hence, the NO coverage decreases rapidly as is shown by the steep decrease in RAIRS intensity of NO. In Figure 6d, an additional feature in the RAIRS intensity of NO occurs between 350 and 430 K due to molecular desorption of NO. Hopstaken and Niemantsverdriet4 found similar trends in the NO dissociation temperature using SSIMS. The Rh2NOþ/ Rh2þ signal was used as a qualitative measure for the NO coverage as the integrated absorption intensity is used in Figure 6. They reported, at low coverages, NO decomposition at 200 K 16244 DOI: 10.1021/la1013544

which shifts to 350 K at high surface concentrations. In principle, static SIMS is a destructive technique which can influence the surface chemistry. Due to the interaction of the high energy primary ion beam with the surface, secondary ion clusters evolve from the surface, leaving behind surface defects. These defects could have a different (probably higher) surface reactivity. That RAIRS and SIMS give similar results indicates that the influence of the beam of ions on surface chemistry is limited and both RAIRS and SIMS are suitable tools for studying these systems. The way CO is bonded to the surface during CO2 formation at low (Figure 6a), medium (Figure 6b), and high total coverages (Figure 6c and d) will be discussed now. At low coverage, all NO Langmuir 2010, 26(21), 16239–16245

Jansen et al.

Article Table 1. Reaction Scheme of the Reaction between CO and NO on Rh(100)

NO(a) þ * f N(a) þ O(a) O(a) þ CO(a) f CO2(g) þ 2* CO(a) f CO(g) þ * NO(a) f NO(g) þ * 2N(a) f N2(g) þ 2*

low θNO,CO

medium θNO,CO

high θNO,CO

200-270 K 400-460 K excess: 430-530 K

300-340 K 330-380 þ 380-450 K excess: 430-530 K

700-800 K

700-800 K

330-360 K 350-400 K 350-400 K þ excess: 430-530 K 380-440 K 700-800 K

has dissociated, forming N and O atoms when CO reacts with O atoms to CO2. CO is mainly present on the top site above 300 K. It appears that this is the reactive species during CO2 formation around 400 K. This can be seen by the strong decrease in the absorption intensity of CO on top sites. At medium and high coverage (Figure 6b and c), the first CO2 formation peak around 300-350 K is accompanied by a decrease in CO on top sites and an increase in CO on bridge sites. We cannot conclude from these observations that bridged CO is the reacting species here, because several processes occur simultaneously on the surface. Besides the CO2 formation reaction, NO dissociates which changes the coverages of NO, N, and O. Each adsorbate may exert an influence on the CO binding site. Atomic species such as N and O are more strongly bound to the surface and will dictate how a weakly bound molecule such as CO is adsorbed. This makes it very difficult to discern between the different influences. Therefore, more structural information on how the adsorbates are distributed is necessary. Between 400 and 500 K, the excess CO that has not reacted with atomic oxygen desorbs. As indicated by the strong decrease in intensity of top CO at 470 K, it appears that mainly CO on top sites is desorbing. Nitrogen desorbs around 750 K and has very similar desorption characteristics as nitrogen desorption in Figure 2. Starting from adsorbed CO and NO, the reaction between NO and CO on Rh(100) follows the scheme in Table 1. The effect of lateral interactions plays an important role in this reaction scheme, in particular in NO decomposition and CO2 formation. At increasing adlayer concentration, NO decomposition shifts to higher temperatures due to the lack of free sites. At low coverages, NO decomposition and CO oxidation occur at different temperatures. At medium to high adlayer concentrations, NO decomposition becomes blocked and the atomic oxygen coverage is low. CO oxidation becomes dependent on NO decomposition and occurs at similar temperatures.

Conclusions A combination of TPRS and RAIRS has been used to study the decomposition of NO and the reaction between NO and CO

Langmuir 2010, 26(21), 16239–16245

on Rh(100). The goal was to unravel how reaction kinetics is influenced by changes in adsorption sites. NO adsorbs in two configurations: in a tilted or almost parallel position at coverages below 0.18 ML and in an upright position on probably a bridge site at all coverages. NO decomposition occurs in two reactivity regimes with all NO decomposing up to 0.25 ML NO, while also molecular NO desorption is observed above 0.25 ML NO. Coadsorbing NO and CO has only minor influence on the NO binding site, while the CO bonding site shifts gradually from top toward the bridge site upon increasing the NO coverage. TPRAIRS with TPRS is a very powerful combination to study the reaction between CO and NO. Site occupation, qualitative surface coverages, and desorption rates are simultaneously obtained. Varying the NO and CO coverage influences NO decomposition and CO2 formation. At low coverages, NO already decomposes around 230 K. CO2 is formed at 440 K which is close to the CO oxidation temperature in the absence of NO and atomic N. At high coverages, NO decomposition becomes blocked until a temperature of 350-370 K is reached. Free surface sites are generated through desorption of a small amount of CO. The free sites are consumed by NO to dissociate, and the formed oxygen atoms react with CO to form CO2 and more empty sites. This makes the reaction autocatalytic and in fact explosive in the formation of CO2. Changes in absorption intensities in TP-RAIRS experiments cannot always be related to a surface reaction. It works well when elementary reaction steps occur in sequence at different temperatures. However, when elementary reaction steps occur parallel to each other, it becomes very difficult to discern the causes of change in site occupation. This was the case when NO decomposition becomes blocked and NO decomposition occurs simultaneously with CO oxidation. It was not possible to conclude if bridge or top CO is the reacting species. More information on the distribution of the adsorbates from, for example, DFT or scanning tunneling microscopy (STM) is necessary to unravel this complex reaction.

DOI: 10.1021/la1013544

16245