Interaction of small molecules with Au(3 1 0): Decomposition of NO

Interaction of small molecules with Au(3 1 0): Decomposition of NO

Applied Catalysis A: General 291 (2005) 93–97 www.elsevier.com/locate/apcata Interaction of small molecules with Au(3 1 0): Decomposition of NO C.P. ...

260KB Sizes 0 Downloads 12 Views

Applied Catalysis A: General 291 (2005) 93–97 www.elsevier.com/locate/apcata

Interaction of small molecules with Au(3 1 0): Decomposition of NO C.P. Vinod a,b, J.W. Niemantsverdriet Hansa, B.E. Nieuwenhuys a,b,* a

Schuit Institute of Catalysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands b Leiden Institute of Chemistry, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands Received 20 September 2004; received in revised form 12 January 2005; accepted 13 January 2005 Available online 11 May 2005

Abstract The interaction of NO, N2O, CO, O2, H2 and N2 with a stepped Au(3 1 0) surface has been studied using X-ray photoelectron spectroscopy (XPS) and temperature programmed desorption (TPD). In the temperature (80–500 K) and pressure (up to 5  106 mbar) range used, no adsorption of O2, H2, CO and N2 was observed. However, NO and N2O adsorb on this stepped surface. Decomposition of NO to N2O was observed at temperatures as low as 80 K. A possible mechanism for the formation of N2O from NO has been discussed on the basis of our TPXPS results. # 2005 Elsevier B.V. All rights reserved. Keywords: Nitric oxide (NO); Au(3 1 0); Steps and kinks; Defects; X-ray photoelectron spectroscopy (XPS); Temperature programmed desorption (TPD); Gold catalysis; Oxygen; Nitrous oxide (N2O)

1. Introduction One of the exciting discoveries in the field of heterogeneous catalysis in the last decade is the surprisingly high activity of gold-based catalysts towards several reactions of industrial and environmental importance [1–3]. Gold, considered as an uninteresting material for catalysis because of its nobleness [4], was found in the form of nanoparticles, to be an extremely efficient catalyst for oxidation and reduction of several small molecules including CO and NO. However, the origin of the catalytic activity of gold-based catalyst is still a debate among the researchers. A number of different models have been proposed to explain this reactivity in terms of either special reaction sites [5], special electronic structure of gold nanoparticle [6,7], the nature of the support [8–11], the particle–support interface [12–16] or a combination of these factors. An interesting and unambiguous observation though is that a decrease in particle size of gold is often * Corresponding author. E-mail address: [email protected] (B.E. Nieuwenhuys). 0926-860X/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2005.01.046

accompanied by an increase in catalytic activity, with possibly, a maximum around 3–5 nm [17]. Both theoretical and experimental results, therefore, may suggest that steps and kinks on a gold surface might play a very crucial role in determining the activity of gold clusters as the defect density on a metal surface increases with decrease in particle size [18–21]. The concept of the active site in catalysis was proposed as early as 1920s [22]. Interaction of small molecules like dioxygen, CO and CO2 towards gold clusters and well-defined Au surfaces has been an active area of theoretical research for the last few years. Recently, Xu and Mavrikakis [23] have shown by density functional calculations that the stepped Au(2 1 1) surface can dissociate oxygen molecules. The dissociation barrier for dioxygen was decreased by 0.6 eV on a stepped surface. Hu and co-workers [24] have also carried out DFT calculations on CO oxidation on gold. They conclude that the oxygen atom chemisorption energy is highest for step atoms and least for a flat Au(1 1 1) surface. They also find that CO can adsorb rather well on defects and poorly on flat Au(1 1 1). The huge effect of steps, kinks, dislocations or lattice strain on the chemical properties of transition metal

94

C.P. Vinod et al. / Applied Catalysis A: General 291 (2005) 93–97

surfaces has been demonstrated experimentally by many studies. Recently, using STM/RAIRS and temperature programmed desorption (TPD), Freund and co-workers [25] studied the effect of particle size on the strength of CO adsorption and concluded that a high heat of adsorption is associated with highly uncoordinated gold atoms. The experimental results are partly motivated to understand the crucial role of stepped Au(3 1 0) (formed by (1 1 0) steps and (1 0 0) terraces) towards small molecules especially oxides of nitrogen. We were also encouraged by the recent studies using VT-STM from our group which showed nitric oxide induced lifting of the surface reconstruction of Au(1 0 0) [26]. The binding of NO and N2O to gold and to transition metal surfaces in general is of interest from two points of view: one related to the catalytic reduction of NO for pollution control and a more fundamental perspective related to the formation of the chemisorption bond. NO has a single electron in the antibonding 2p* orbital causing complex adsorption behavior on metal surfaces. In addition, the adsorption of NO on metal surfaces at low temperatures is found to produce N2O [27–30]. There is a debate regarding the mechanism of formation of such species as through dissociation or through NO–NO pairing, either by formation of dinitrosyl species or via dimers of NO. In addition, the oxides of nitrogen are of technological importance in air pollution and also as strong oxidizing agents. Most of the surface science studies concerning NOx adsorption are dealing with Pt group metals because of the application of Pt-based catalysts for automotive pollution control. Most of the studies show that Pt(1 1 1) [31] and Pt(1 1 0) [32] are inactive for NO dissociation while NO dissociates on Pt(1 0 0) at room temperature [33]. Stepped and kinked surfaces such as Pt(2 1 0) [34], Pt(4 1 0) [35] and Pt(3 1 0) [36] have also been shown to dissociate NO at room temperature. Literature on the interaction of oxides of nitrogen with gold surfaces is very scarce. Bartram and Koel [37] studied the interaction of NO2 on Au(1 1 1) using TPD and HREELS. They found molecularly adsorbed NO2 at 100 K with saturation coverage of 0.4 ML. The NO2 surface chelate in their case is found to react with NO to produce N2O3. They could not find NO and N2O adsorption on the same surface at 95 K. Since there is lot of interest in goldbased catalysts for several reactions of industrial importance we thought it appropriate to study the interaction of several small molecules (NO, N2O, CO, H2, O2 and N2) on a stepped gold surface.

could also be cooled to as low as 80 K via a liquid nitrogen reservoir in contact with the sample mount. The Au(3 1 0) surface was cleaned by repeated cycles of Ar ion (2 keV) bombardment and annealing at 950 K. The cleaning procedure mentioned above was sufficient to get a clean surface devoid of any carbon, which was the only contaminant on the surface. The sample temperature was measured using a chromel/alumel thermocouple spot welded at the back of the crystal. The gas exposures were carried out directly onto the crystal through a fine leak valve in the main chamber of the spectrometer. All the filaments were switched off prior to dosing in order to prevent any possible effects of electron beam damage on NO. All the XPS spectra were corrected to Au(4f7/2) binding energy of 84 eV. The clean Au(4f7/2) had a natural line width of 1.5 eV.

3. Results In Fig. 1, we show TPD spectra obtained by exposing Au(3 1 0) at 80 K to NO, N2O, O2, CO, N2 and H2. High exposures of the order of 106 mbar for several minutes were used for all the gases except N2O. For N2O, dosing was done at 107 mbar at 80 K for 15 min. The TPD was carried out from 80 to 500 K with a heating rate of 5 K s1. The desorption of NO was detected by monitoring mass 30 and N2O by mass 44. From TPD profile, it is clear that dosing NO at 80 K on Au(3 1 0) also produces N2O desorption from the surface. The desorption maximum for NO from our TDS results corresponds to 120 K and that of N2O (produced by dosing NO) at around 150 K. The activation energies for desorption estimated using Redhead’s method and first order desorption kinetics (n = 10131 s1; b = 5 K s1) are 30 and 35 kJ mol1, respectively. The high background tail is due to

2. Experimental The experiments were carried out in a custom built ultrahigh vacuum system, which has facilities for X-ray photoelectron spectroscopy (XPS) and TPD. The Au(3 1 0) surface was mounted on tungsten wires by which the sample could be resistively heated to above 1000 K. It

Fig. 1. TPD spectra obtained by dosing (a) H2, (b) N2, (c) O2, (d) CO and (e) NO at 80 K on Au(3 1 0) (5  106 mbar; 30 min). Spectra (f) correspond to N2O desorption following decomposition of NO and (g) that of dosing pure N2O at 107 mbar for 15 min. Spectra were shifted and presented for better clarity.

C.P. Vinod et al. / Applied Catalysis A: General 291 (2005) 93–97

95

Fig. 2. N 1s and O 1s core level spectra obtained after dosing NO at 80 K (5  106 mbar; 30 min) and subsequent warming. Top spectra correspond to N 1s and O 1s core level after dosing NO at 160 K maintaining the same exposure conditions.

the high exposures and slow pumping speed of NO and N2O. The small NO desorption peak found at 320 K can be attributed to that emanating from NO adsorbed on the sample manipulator. This suggestion is supported by XPS observations that show absence of N and O containing species above 300 K. The TPD plots obtained from the surface by dosing the surface to O2, CO, N2 and H2 are also shown in Fig. 1. It should be noted that employing the same exposure conditions we could not find any adsorption of these species on the surface. We have also carried out TPD by dosing pure N2O on Au(3 1 0) at 80 K. It should be noted that pure N2O desorbs from the surface at a much lower temperature (120 K) compared to N2O produced from NO on the surface. In Fig. 2, we show the N 1s and O 1s XP spectra obtained by dosing NO (5  106 mbar; 30 min) at 80 K and subsequent warming. The N 1s core level shows that two types of nitrogen species are produced on the surface characterized by binding energies of 401.5 and 403.5 eV. The latter species is assigned to molecularly adsorbed NO. This assignment matches well with values reported for NO on several transition metal surfaces like Ni [38]. Since the reported binding energy of NO on surface like Ag(1 1 1) comes 402 eV, a possibility that NO (403.5 eV) forming a surfaces complex with Oads similar to one reported on Pt(2 1 1) cannot be ruled out in our case [39]. The species with binding energy of 401.5 eV can be assigned to N2O (as will be shown later). On subsequent warming, first the NOads desorbs from the surface at around 120 K and then N2O at around 220 K. The corresponding O 1s spectra at 80 K show a broad peak centered around 532 eV with contributions

from NOads, N2Oads and Oads. We have not attempted to fit the O 1s spectra to different nitrogen oxides and Oads as the spectrum is not as well defined as N 1s. But it is interesting to note from the O 1s spectra that the reminiscent of oxygen in the form of Oads (531 eV) is still there on the surface even at 280 K. The formation of Oads with binding energy 531 eV by adsorption of molecules like NO, O2 on surfaces like Cu and Ag is well manifested in literature [40]. Our results are interesting in the light of previous NO adsorption studies on Au(1 1 1) surface which suggest practically no adsorption at temperatures as low as 90 K probably due to low exposures employed in their experiment or the absence of step sites [37]. The adsorption of NO at elevated temperature is also presented in Fig. 1. The top spectra correspond to the adsorption of the same dosage of NO at 160 K. It is clear from the N 1s core level that the only nitrogen species present on the surface is N2O (401.5 eV). The corresponding O 1s spectra show two well-defined peaks at 531 eV (Oads) and 532.5 eV (N2Oads). It is worth mentioning here that we could not find any N2O adsorption on this surface by dosing pure N2O (5  106 mbar; 30 min) at temperatures above 100 K. In Fig. 3, we show the adsorption of N2O on Au(3 1 0) at 80 K and subsequent warming. It is clear that the there are two well-defined peaks at 401.5 and 405.5 eV. These two peaks can be associated to NNO and NNO’s of nitrogen [49]. On warming the surface to 120 K the N2O completely desorbs from the surface. It should be noted that we could only find one peak (at 401.5 eV, NNO) for N2O formed by decomposition of NO.

96

C.P. Vinod et al. / Applied Catalysis A: General 291 (2005) 93–97

The formation of N2O via dissociative adsorption of NO has been reported for many metal surfaces [29,45–47]. The evidence for the dissociative adsorption of NO on Cu [29,45] comes from the fact that XPS and LEED could identify the presence of Oad from the surface at 80 K and at subsequent higher temperatures. Support for the decomposition mechanism also comes from a recent NO adsorption study of Mukerji et al. where RAIRS could not detect the presence of (NO)2 on Pt(2 1 1). Their DFT calculations also showed that (NO)2 species was unstable on Pt(2 1 1) [39].

Fig. 3. XPS spectra obtained by dosing Au(3 1 0) at 80 K to N2O at 1  107 mbar for 15 min and subsequent warming.

4. Discussion Exposure of Au(3 1 0) to O2, H2, CO and N2 did not result in adsorption in agreement with the earlier reports for gold films [41], Au(1 1 0) [42] and Au(1 1 1) [50]. Apparently, the presence of steps is not sufficient to enable adsorption under our experimental conditions. On the other hand, NO and N2O were found to adsorb fairly well on Au(3 1 0) as indicated by our XPS and TPD. Interestingly, NO is found to decompose on Au(3 1 0) at temperatures as low as 80 K. Even though, the formation of N2O following adsorption of NO is well known on many metal surfaces it is striking to note such an event happening on a relatively ‘‘inert’’ massive gold surface. In principle, two mechanisms may lead to N2O formation: (i) pairing of two NO molecules or dimer mechanism: NOg ! NOads (1) (2) 2NOads ! N2 Oads þ Oads According to this mechanism, N2O is formed by Npairing of two NOads molecules before the NO bond is broken. The presence of dinitrosyl species or NO dimers may result in the pairing of the N-atoms [43,44]. On Ag(1 1 1) surface, there is evidence for the formation of NO dimers in NO multilayers at temperatures as low as 40 K based on RAIRS observations [30]. On warming the multilayer to 90 K (NO)2 transforms to N2O. (ii) via N–O bond breaking and a reaction of NOads with Nads [36,38]. NOads ! Nads þ Oads

(3)

2Nads ! N2g

(4)

NOads þ Nads ! N2 Oads

(5)

Recently, the role of reactive Nads in catalysis has been discussed in detail [48]. The Nads formed by the decomposition of NO can either react with NO to form N2Oads (equation (3)) or combine with Nads to N2 (equation (4)). Equation (3) seems to be the preferred path way for N2O formation on surfaces like W(1 1 0) [46], Cu(1 0 0) and Cu(1 1 1) [45]. The fate of Nads formed by equation (3) is metal specific. For example, on Cu(1 0 0) and Cu(1 1 1) there is no evidence for chemisorbed Nads using XPS. All the Nads being accounted for the formation of N2O. This also seems to be the case in our experiments where we could not detect any Nads (binding energy 397 eV) by XPS. The interesting point though is that at 160 K, the O 1s spectra clearly demonstrated the signatures of decomposition. At 160 K, the NOads decomposes to Nads, which immediately reacts with NO molecules to leave N2O and Oads on the surface. Another observation is the stability of the N2O formed by the decomposition of NO. N2O thus formed being stable up to 220 K whereas N2Oads formed by dosing pure N2O desorbs from the surface at around 120 K as shown by XPS and TPD. Though more experiments need to be done to prove this, it indirectlypointstothestabilizingeffectofOads towardsN2Oon the surface.Theresultspresented here are alsointerestinginthe light of recent NO experiments carried out on Pt(2 1 1) surface using TPD where NO adsorption at 120 K and warming produces N2O by decomposition [34]. Since there are no reports of NO interaction with gold surfaces, a direct comparison on the effect of steps on Au(3 1 0) could not be made. The presence of NO-dimers are generally observed at very low temperatures (below 70 K). Hence, it is very unlikely thatinourcaseN2O formationproceedsviadimerformationon Au(3 1 0). In the light of ourexperiments,we are encouraged to suggest that NO decomposes on Au(3 1 0) following equations (3) and (5) and produces N2O on the surface, similar to Pt(2 1 1). The earlier workfrom ourgroupdescribed Au/Al2O3 (if the particle size is in nanometer range) as a very active catalyst in NO and N2O reduction [51,52]. The present results suggest that the NOx reduction proceeds via decomposition on special Au sites present on the Au nanoparticle.

5. Conclusions In conclusion, our TP-XPS and MS-TPD results clearly support some of the theoretical calculations predicting

C.P. Vinod et al. / Applied Catalysis A: General 291 (2005) 93–97

enhanced reactivity of stepped surfaces of gold towards small molecules, in our case nitric oxide. The NO molecule decomposes at temperatures as low as 80 K and forms N2Oads and Oads on the surface. These results thus may contribute to a better understanding of gold catalysis, especially with respect to the role of steps and their reactivity towards nitric oxide.

Acknowledgement This work was supported by the Netherlands Organization for Scientific Research (NWO #700.99.037 and 047.015.003).

References [1] G.C. Bond, D.T. Thompson, Cat. Rev. Sci. Eng. 41 (1999) 319. [2] M. Haruta, Chem. Records 3 (2003) 75. [3] R. Grisel, K.J. Weststrate, A. Gluhoi, B.E. Nieuwenhuys, Gold Bull. 35 (2002) 39. [4] B. Hammer, J.K. Norskov, Nature 376 (1995) 238. [5] M. Ruff, S. Frey, B. Gleich, R.J. Behm, Appl. Phys. A 66 (1998) 513. [6] M. Valden, X. Lai, D.W. Goodman, Science 281 (1998) 1647. [7] L. Guczi, G. Peto, A. Beck, K. Frey, O. Geszti, G. Molnar, C. Daroczi, J. Am. Chem. Soc. 125 (2003) 4332. [8] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. Genet, B. Delmon, J. Catal. 144 (1993) 175. [9] M.A.P. Dekkers, M.J. Lippits, B.E. Nieuwenhuys, Catal. Lett. 56 (1998) 195. [10] M. Haruta, N. Yamada, T. Kobayashi, S. Ijima, J. Catal. 115 (1989) 301. [11] S. Lin, M. Bollinger, M. Vannice, Catal. Lett. 17 (1993) 245. [12] G.R. Bamwenda, S. Tsubota, T. Nakamura, M. Haruta, Catal. Lett. 44 (1997) 83. [13] J.D. Grundwaldt, A. Baiker, J. Phys. Chem. B 103 (1999) 1002. [14] B.E. Nieuwenhuys, in: W. Joyner, R.A. Van Santen (Eds.), Elementary Reaction Steps in Heterogeneous Catalysis, Kluwer Academic Publishers, 1993. [15] K. Hayek, R. Kramer, Z. Paal, Appl. Catal. A 162 (1997) 1. [16] L.M. Molina, B. Hammer, Phys. Rev. Lett. 90 (2003) 20. [17] M. Haruta, Catal. Today 36 (1997) 153. [18] S. Dahl, E. Tornqvist, I. Chorkendorff, J. Catal. 192 (2000) 381.

97

[19] T. Zambelli, J. Wintterlin, J. Trost, J. Ertl, Science 273 (1996) 1688. [20] M. Mavrikakis, P. Stolze, J.K. Norskov, Catal. Lett. 64 (2000) 101. [21] J. Greenley, J.K. Norskov, M. Mavrikakis, Ann. Rev. Phys. Chem. 53 (2002) 319. [22] H.S. Taylor, Proc. R. Soc. Lond. A 108 (1925) 105. [23] Y. Xu, M.J. Mavrikakis, Phys. Chem. B 107 (2003) 9298. [24] Z.P. Liu, P. Hu, A. Alavi, J. Am. Chem. Soc. 124 (2002) 14770. [25] C. Lemire, R. Meyer, S. Shaikhutdinov, H.J. Freund, Angew. Chem. Int. Ed. 43 (2004) 118. [26] E.D.L. Rienks, G.P. van Berkel, J.W. Bakker, B.E. Nieuwenhuys, Surf. Sci. 571 (2004) 187. [27] C. Nyberg, P. Uvdal, Surf. Sci. 204 (1988) 517. [28] H. Ibach, S. Lehwald, Surf. Sci. 76 (1978) 1. [29] J.F. Wendelken, Appl. Surf. Sci. 11–12 (1982) 172. [30] W.A. Brown, P. Gardner, D.A. King, J. Phys. Chem. 99 (1995) 7065. [31] R.J. Gorte, J.L. Gland, Surf. Sci. 102 (1981) 348. [32] M.E. Lesley, L.D. Schmidt, Surf. Sci. 155 (1985) 215. [33] E.D.L. Rienks, J.W. Bakker, A. Baraldi, S.A.C. Carabineiro, S. Lizzit, C.J. Weststrate, B.E. Nieuwenhuys, Surf. Sci. 516 (2002) 109. [34] R.J. Mukerji, A.S. Bolina, W.A. Brown, J. Chem. Phys. 119 (20) (2003) 10844. [35] Y.O. Park, W.F. Banholzer, R.I. Masel, Surf. Sci. 155 (1985) 57. [36] S. Sugai, K. Takeuchi, T. Ban, H. Miki, K. Kawasaki, Surf. Sci. 282 (1993) 67. [37] M.E. Bartram, B.E. Koel, Surf. Sci. 213 (1989) 137. [38] C.R. Brundle, M.W. Roberts, Proc. R. Soc. A 331 (1972) 383. [39] R.J. Mukerji, A.S. Bolina, W.A. Brown, Z.P. Liu, P. Hu, J. Phys. Chem. B 108 (2004) 289. [40] M.W. Roberts, Chem. Soc. Rev. 25 (1996) 437, and references therein. [41] B.M.W. Trapnell, Proc. R. Soc. A 218 (1952) 566. [42] D.A. Outka, R.J. Madix, Surf. Sci. 179 (1987) 351. [43] M. Shelef, G.W. Graham, Catal. Rev. Sci. Eng. 36 (1994) 433. [44] T.R. Ward, P. Alemany, R. Hoffmann, J. Phys. Chem. 97 (1993) 7691. [45] D.W. Johnson, M.H. Matloob, M.W. Roberts, J. Chem. Soc., Faraday Trans. 75 (1979) 2143. [46] R.I. Masel, E. Umbach, J.C. Fuggle, D. Menzel, Surf. Sci. 79 (1979) 26. [47] B.E. Nieuwenhuys, Adv. Catal. 44 (1999) 259. [48] M.W. Roberts, Catal. Lett. 93 (2004) 29. [49] J. Grimblot, P. Alnot, R.J. Behm, J. Electron Spectrosc. Relat. Phenom. 52 (1990) 175. [50] N. Saliba, D.H. Parker, B.E. Koel, Surf. Sci. 410 (1998) 270. [51] A.C. Gluhoi, S.D. Lin, B.E. Nieuwenhuys, Catal. Today 90 (2004) 175. [52] M.A.P. Dekkers, M.J. Lippits, B.E. Nieuwenhuys, Catal. Today 54 (1999) 381.