Possible deNOx Management under Net Oxidizing ... - ACS Publications

ARTICLE pubs.acs.org/JPCC

Possible deNOx Management under Net Oxidizing Conditions: A Molecular Beam Study of 15NO + CO + O2 Reaction on Pd(111) Surfaces Sankaranarayanan Nagarajan,† Kandasamy Thirunavukkarasu,†,‡ and Chinnakonda S. Gopinath*,†,# † #

Catalysis Division, National Chemical Laboratory, Dr. Homi Bhabha Road, Pune 411 008, India Center of Excellence on Surface Science, National Chemical Laboratory, Dr. Homi Bhabha Road, Pune 411 008, India

bS Supporting Information ABSTRACT: Isothermal kinetic measurements of 15NO reduction with CO on Pd(111) surfaces were carried out under net-oxidizing conditions with 15NO + CO + O2 , using a molecular beam instrument (MBI). Transient state (TS) and steady state (SS) kinetic details of the above reaction were obtained for a wide range of temperature and beam compositions, especially with O2-rich compositions. Increasing O2 content, generally, suppresses 15NO reduction in the SS; nonetheless, irrespective of O2 content, 15N2 was produced in TS, and to a significant extent under SS conditions too. Sustainable N2 production between 450 and 600 K and with low to moderate amount of oxygen was observed, and the extent of NO decomposition was also quantified. The ratio of 15N2:15N2O was generally found to be around 8:1 under most of the reaction conditions. Maxima in the SS reaction rates of all products were observed between 500 and 600 K. Compared to other elementary reaction steps, a slow decay observed with N + N f N2 step under SS beam oscillation conditions demonstrates its contribution to the rate limiting nature of the overall reaction. Fast beam switching experiments have been performed alternately between O2-lean and -rich conditions, thus highlighting the effectiveness of 15NO reduction in TS, irrespective of the beam composition. Possibly in a future technology initiative, this aspect could be exploited to manage more 15NO reduction on Pd-based catalysts.

1. INTRODUCTION The nitric oxide emission from automobile exhaust is one of the key environmental issues of the 21st century and it is directly affecting human health in many ways. Regulations for the control of NOx emissions from automobile exhausts have dramatically increased all over the world.1
4 The basic aim of the three way catalytic converter (TWC) catalyst is to convert the pollutants (NOx, CO, CxHy) to harmless gases (N2, H2O, CO2).5 Improvements in automobile internal combustion (IC) engine design and the developments in fuel quality led to more fuelefficient vehicles. Nonetheless, problems exist, particularly in the reduction of NO to N2, especially under net oxidizing conditions with fuel-efficient lean-burn engines.6 In the past decade Pd-based TWC converters (Pd supported on Al 2O3 , CeO2 , ZrO2 , SiO2 , La2 O3 , or a combination of supports) were increasingly introduced, and it replaced the traditional Rh
Pt TWCs, due to its significant performance in reducing NO under net oxidizing conditions, relatively low cost, and the higher natural abundance compared to Rh. Metals like Pt and Rh are found to be passive in oxygen-rich conditions due to irreversible oxide formation.7
10 However, there is no such poisoning effect observed on the Pd surface, even under oxygenrich conditions for CO oxidation.11
14 Hence an understanding of NOx reduction reactions, under net oxidizing conditions on palladium catalysts, is attracting more attention since the last decade. Due to the complex nature of catalyst materials, the molecular level understanding and the microkinetic details remain to be explored. Hence fundamental studies on TWC reactions, especially on single crystal surfaces with Pd, are a r 2011 American Chemical Society

necessity. Modern IC engines are operating at high air to fuel ratio, and consequently more oxygen is found in the exhaust gases too. To achieve NO reduction is highly challenging in such O2-rich conditions. It is important to note that Pd remains active even in such high O2 content, because of the reversible oxidation capacity of Pd. This makes Pd a better catalyst. The O2 interaction with Pd has been pursued, and its influence on oxidation catalysis has been explored by a few groups.11
19 NO:CO and NO:CO:O2 reactions were also carried out on noble metal single crystal surfaces by a few groups.20
24 Zaera et al. studied NO:CO reaction extensively on Rh(111) surfaces and it was found that maximum N2 production was observed between 500 and 600 K.25,26 The addition of O2 inhibits the rate of production of N2 and CO2 under the actual reaction conditions. This behavior was explained by postulating the poisoning of surface active sites by the dissociated O atoms. The addition of excess CO also inhibits the rate of production of both N2 and CO2, below 500 K due to the CO poisoning. The addition of O2 to CO-rich beams enhances the conversion of NO to N2. The added O2 obviously helps to remove the excess CO from the surface. Zaera et al26,27 observed that the maximum product formation (N2 and CO2) occurs at stoichiometric ratio of adsorbates (NO and CO) on Rh(111) surfaces. In our earlier studies of the same NO + CO reaction it has been found that

Received: July 25, 2011 Revised: September 13, 2011 Published: September 18, 2011 21299

dx.doi.org/10.1021/jp207092s | J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C

ARTICLE

significant N2 production occurs around 500 K, following catalytic decomposition of NO on Pd(111).28 Goodman et al.29
31 carried out the NO + CO reaction on Pd(111) and Pd(100) surfaces and showed higher activity of Pd(111) compared to Pd(100). Matsushima et al32,33 identified the following three elementary steps (1
3) for nitrogen removal, which constitute the main events in the NO:CO:O2 reaction on Pd(110): Although there have been a few reports on the SS and the TS behavior of the NO:CO:O2 reaction on Rh surfaces2 and also a few on Pd-based powder catalysts,34
37 to the best of our knowledge there is no report on Pd(111), especially under conditions such as high temperatures and a wide range of NO: CO:O2 ratios. In the present paper, we address the issue of NO dissociation under net-oxidizing conditions and also with beam switching experiments, between O2-lean and -rich compositions, to effect NO reduction. This work is a part of our ongoing study on the detailed kinetics and mechanism of catalytic converter reactions and associated aspects.38
40 N2 OðadsÞ f N2ðgÞ þ OðadsÞ

ð1Þ

N2 OðadsÞ f N2 OðgÞ

ð2Þ

2NðadsÞ f N2ðgÞ

ð3Þ

2. EXPERIMENTAL SECTION All experiments reported here were performed in a home-built 12 L capacity stainless-steel ultrahigh vacuum (UHV) chamber38,39 evacuated with a 210 L/s turbo molecular pump (Pfeiffer, TMU261) to a base pressure of about 2  10
10 Torr. This system is equipped with quadrupole mass spectrometer (Pfeiffer, Prisma QMS 200M3), which can detect gas phase species up to 300 amu. A detailed description of the doser setup, its calibration, and other details of MBI have been given elsewhere.11
14,38 A total flux (F) of reactants of 0.32 monolayer per second (ML/s) was used in all the experiments reported here, unless otherwise specified. Isothermal experiments with the 15NO:CO:O2 reaction on Pd(111) surfaces were conducted in a MBI between 400 and 800 K and with x15NO:yCO:zO2; hereafter x:y:z represents the composition of the respective individual components varied from x = 1 to 6, y = 1 to 14, and z = 0 to 7. The contribution from the background to the measurements of the reaction rates was deemed to be less than 5% of that from the direct beam by independent calibration experiments11,13,28,38 and was not considered for calculations of the SS rates and coverages. A five-way valve was connected to the gas manifold system to have fast beam switching experiments, which is reported elsewhere.39a The Pd(111) single crystal (Metal Oxide Ceramics, U.K.) used in all the experiments is a circular disk 8 mm in diameter and 1 mm thick. The crystal was cleaned by the standard procedure31 of Ar+ sputtering in the presence of oxygen and argon together (total pressure of 1.5  10
6 Torr) at 950 K, and subsequent flashing to 1100 K. A detailed sample cleaning procedure is available in our publications.11
14,28,38 TPD spectra were recorded at a constant heating rate of 10 K/s. Isotopically labeled 15NO (SG Spectra Gases 99%), CO, and oxygen (Inox Air Products Ltd., 99.9%) were used without any further purification.

Figure 1. 15NO:CO:O2 molecular beam (of 2:2:1) is directed onto a clean Pd(111) surface as the temperature kept at 550 K and the partial pressures of all the reactants (15NO, CO, and O2) and products (15N2 and CO2) are followed as a function of time. The beam was blocked at 210 s and unblocked at 230 s in order to measure the SS rates due to the gases in the beam directly, as they are proportional to the drop in partial pressure of the products and/or the rise in partial pressure of the reactants from their SS values. Note the instantaneous changes in CO2 signal with blocking and unblocking of the beam, in contrast with the slower response of the 15N2 and 15N2O trace.

The mass spectrometer signals of reactants and products are calibrated independently, by leaking the corresponding pure component. All the rates obtained are accurate within 5% error, the main source of errors originating, likely, from the mass spectrometer calibration and the contribution of the background gases to the overall measured reaction rates. In relative terms, the reported values were reproducible to within 5% for a given beam composition, beam flux, and surface temperature.

3. RESULTS 3.1. General Considerations. All kinetic experiments reported here for 15NO:CO:O2 reactions on Pd(111) surfaces were carried out in a manner identical to that of reported in our earlier publications.11,28 The clean Pd(111) surface is first heated to a predecided temperature and then exposed to an effusive molecular beam with required 15NO:CO:O2 ratio, while the partial pressure of the different gases of interest is followed by QMS as a function of time. Figure 1 shows typical raw kinetic data obtained in this manner. Figure 1 shows the evolution of partial pressure traces of 15NO, CO, O2, CO2, 15N2, and 15N2O for a beam of 15NO:CO:O2 (2:2:1) composition directed on to Pd(111) surfaces at 550 K. Although 15NO2 (amu = 47) was also recorded, no measurable intensity was observed suggesting that there is no 15NO2 production under the present experimental conditions. Indeed no NO2 was observed in a similar work on Rh(111) surfaces.7 Various steps involved in most of the experiments are explained below with reference to Figure 1: (1) At t = 10 s, a molecular beam of a mixture of the reactants was turned on; an immediate rise in reactants partial pressures could be seen. Some adsorption of the reactants from the background cannot be avoided at this stage. 21300

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C It is also to be mentioned here that a shutter is in place and blocking the beam reaching Pd(111) surface. (2) At around t = 13 s, the shutter was opened (solid arrow) in Figure 1 to allow the beam to interact directly with the Pd(111) surface kept at 550 K. An immediate decrease in the partial pressure of the reactants was observed indicating the adsorption of reactants on the Pd(111) surface. These data are helpful to explore the competitive adsorption characteristic of reactants from a mixture of reactants. An increase in the partial pressure of the products 15N2 (30 amu), CO2 (44 amu), and 15N2O (46 amu) was also observed. The changes observed from the time the shutter opened (t = 13 s) until the SS reached is termed as the TS; SS normally reaches within 1 min under the present experimental conditions. (3) In the SS, the reaction rate was measured by blocking the beam deliberately for about 20 s (between t = 210 and 230 s) with the shutter. An increase (decrease) in the partial pressure of all reactants (products) was observed. A slow increase in 15NO partial pressure (up to 10%) was due to 15NO adsorption on the UHV chamber walls, which decreases with time, and does not indicate a change in the 15NO flux (FNO) on Pd(111).28 An increase in the overall partial pressure of 15NO and O2 was observed, while blocking the beam in the SS; a sharp change in the CO partial pressure was observed only at the point of blocking and unblocking of the beam. The above observation and other results presented in this article highlight the net adsorption is significantly influenced by the reaction conditions. The measured changes in the partial pressure of products allow direct determination of the SS reaction rates, indeed after calibration with pure components. A sharp decrease in the CO2 partial pressure was observed when the beam was blocked, whereas a slow decay in partial pressure was observed for the 15N2 and 15N2O. (4) At t = 230 s the shutter was opened to resume the SS. Immediately on shutter opening CO, O2, and CO2 intensities revert back instantly; however 15N2 and 15N2O gradually increase and the time required to reach the earlier SS depends on reaction conditions. In the present case shown in Figure.1, it takes about 10 s to reach the earlier SS value again. On shutter opening, large 15NO adsorption, resembling that of TS adsorption occurs, which is in contrast to CO and O2. The reaction was followed to continue for some more time. (5) At about t = 320 s, the beam was turned off to stop the reaction. After the pressure of the UHV chamber reached the initial background level, the sample temperature was ramped to 1000 K at a heating rate of 10 K/s to record the TPD of all the relevant species. A systematic study of the 15NO + CO + O2 reaction kinetics on Pd(111) surfaces was carried out by following the above procedure as a function of temperature and 15NO:CO:O2 composition. During the reaction measurements, the total pressure in the chamber increases to about (1
2)  10
8 Torr and that induces some additional adsorption/reaction apart from the direct beam onto Pd(111). As shown in Figure 1, the reactant pressure increases in the same manner, whether the beam was blocked or not, except while opening/closing the shutter. CO2 shows an increase in pressure and displays that there is reaction to some extent due to background adsorption of reactants. In the

ARTICLE

Figure 2. Time evolution of the partial pressure of the reactant uptake ((a) 15NO), and product desorption ((b) CO2, (c) 15N2, and (d) 15N2O) in kinetic experiments such as that described in Figure 1 for the compositions of (1:1:2) 15NO:CO:O2 at different temperatures. CO2 production occurs at all temperatures, but 15N2 and 15N2O formation is observed only between 500 and 600 K. Shutter open and shutter close are indicated by dotted and solid arrows, respectively, in all the figures.

case of 15N2, a significant increase in the background pressure is mostly due to 1% 14NO (amu 30). Indeed, at low temperature (600 K. Interestingly, the kinetics observed in the SS for beam oscillation (immediately after shutter closing or opening) is quite different for N-containing products and CO2 . Changes in CO2 pressure are instantaneous upon blocking and unblocking the beam, whereas the 15N2 and 15N2O pressure changes slowly. Instantaneous changes observed with CO2 partial pressure for beam oscillation measurements demonstrate the associated steps, such as adsorption of CO, 15NO, and O2, and dissociation of 15NO and O2 are also fast. In contrast to the above, a slow decay/recovery in 15N2 and 15N2O partial pressure, up to 10 s, was observed for beam oscillation under SS conditions (Figures 1, 2c, and 2d). The main conclusion from this observation is that the diffusion controlled nature of recombination of N atoms to molecular nitrogen (15N + 15N f 15N2) should be the ratedetermining step (RDS) for the whole process under the conditions of those experiments. 15N2O being a minor product in the parallel pathway for both N-containing products, it is ruled out that 15N2O product formation can be the RDS. More detailed arguments and mechanistic aspects are available for 15NO + CO reaction on Pd(111) surfaces.28 As in Figure 1, extent of 15NO adsorption observed in the TS is more than the SS adsorption, despite 50% oxygen content in the

beam. An important fact is a strikingly different TS kinetics observed for CO2 and 15N2 production. For instance, between 500 and 650 K, the 15N2 production rate increases abruptly in TS and then slowly stabilizes to the SS. However, CO2 production increases gradually to the SS, without any large CO2 production as seen at 500 and 450 K in the TS. These are due to the change in competitive adsorption of reactants as well as desorption of products due to change in surface composition of various species.41 Fast increase in 15N2 production highlights a favorable dissociation of 15NO followed by 15N + 15N recombination. Slow increases in CO2 production underscore the relatively slow O2 uptake under competitive adsorption conditions and, hence, slow CO2 generation in TS. 3.3. Beam Composition Dependence. Figure 3 displays the SS rate of products formation for (a) CO2, (b) 15N2, and (c) 15 N2O on Pd(111) during the exposure of 15NO:CO:O2 = 2:1:z composition, with z varying between 0 and 7, and at temperatures between 450 and 800 K. CO2 production was observed from 450 to 800 K with the rate maxima observed at 550 K for all beam compositions. In the case of 2:1:0 beam composition, the CO oxidation to CO2 occurs exclusively due to the atomic oxygen provided by 15NO dissociation, whereas in the case of 2:1:z (z g 1), O2 is an additional and main reactant source for supply of atomic oxygen. SS rate of CO2 production with 2:1:0 composition is lesser than that of 2:1:1 and 2:1:2 hints the limited O supply by 15NO dissociation. A sudden jump in the rate of CO2 production at g550 K with 2:1:1, 2:1:2, and 2:1:4 compared to 21302

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C

ARTICLE

Figure 4. Transient state temporal evolution of partial pressures of (a) 15NO, (b) O2, (c) 15N2, and (d) CO2 from 2:1:z, where z = 1, 2, 4, and 7, of 15 NO:CO:O2 composition at 500 K on Pd(111).

2:1:0 highlights the competitive O2 adsorption. A decrease in 15 N2 production >550 K demonstrates a decrease in 15NO dissociation due to predominant θ0 on the surface. However, an overall decrease in the CO2 rate was observed for O2-rich 2:1:4 and 2:1:7 compositions due to decreasing FCO relative to F O2. Sustainable CO 2 production was observed in the SS for O2-rich beam composition demonstrating the nonpoisoning of Pd surface with adsorbed O atoms, in contrast to the loss of activity on Rh(111) surfaces25 under comparable conditions. Indeed O-predosed Rh(111) surfaces show hardly any CO uptake and CO2 production above 450 K,25 in contrast to the complete removal of adsorbed O atoms through CO titration on Pd(111) surfaces.11,13 There is a shift in the rate maxima observed with respect to CO2 production from 500 to 550 K with 2:1:0 to 550
600 K with increase in O2 content in the 2:1:z beam. 15N2 production at 550 K for different oxygen content (z) was also observed in the range of 450 and 600 K; however, the maximum 15N2 production was observed with 2:1:0 beam composition. The overall production of N-containing products decreases with increase in O2 content; however, albeit lower rate values, a sustainable 15N2 production was observed with 2:1:z (z = 1
4) beam compositions. It is also to be noted that the 15NO content in 2:1:0 composition is 66%, whereas it decreases to 50, 40, and 28.6% with z = 1, 2, and 4, respectively, and a reason for an overall decrease in the rate of N-containing products. A quantitative calculation of balancing the rate of N-containing products to that of CO2 highlights the extent of 15NO decomposition and O2 utilization for CO oxidation with eqs 4 and 5. RCO2
rCO2 ¼ 2R15 N 2 þ R15 N 2 O ¼ RNODiss

ð4Þ

RCO2
2R15 N 2
R15 N 2 O ¼ rCO2

ð5Þ

RCO2 and rCO2 correspond to the rate of total CO2 and CO2 production exclusively by utilizing molecular O2, respectively. A simple substitution of rate values from Figure 3 for 550 K data indicating rCO2 is 0.027 for z = 1 and 2, 0.02 for z = 4, and 0.014

ML/s for z = 7 (see inset in Figure 3a). Nonetheless, 25% atomic O originates from 15NO dissociation for z = 1, 2, and 4 compositions to be highlighted. This also suggests an increasing extent of 15NO dissociation occurs with increasing O2 content (z e 4) underscoring a significant ability of Pd(111) surface to decompose 15NO under the net oxidizing conditions. However, with 2:1:7 composition, hardly there is any 15NO dissociation, and too high O2 content in the beam as well as on the surface apparently retards 15NO adsorption and dissociation. The rate maxima observed at 550 K with 2:1:1 and 2:1:2 compositions shift to 500 K with increase in the O2 content for 2:1:4. A similar trend is observed in 15N2O also. The reason for rate maxima shift to lower temperature for N-containing products is likely due to the repulsive interaction exerted by more chemisorbed oxygen under O2-rich conditions, which decreases NO
Pd bond strength. Generally an increase in FO2 decreases the rate maximum temperature for CO:O2 compositions11
14,41,42 However, a shift in rate maximum to higher temperature (550
600 K) with increase in O2 content in a 15NO:CO:O2 mixture indicates the dominating role of 15NO at e600 K. This is a reason for sustainable 15NO reduction 600 K can alter the adsorption properties and the Pd is found to be mildly oxidized.11,18,19 The behavior of Pd-based systems supported on alumina and ceria
zirconia was investigated with stoichiometric C3H6: NO:CO:O2 composition by Garcia et al34 and underscores the importance of the oxidation state of Pd on supported catalysts. It has been found that PdO-like species get reduced to metallic Pd on Al2O3, but not on ceria
zirconia supported catalyst. Changes in adsorption dynamics were observed due to competition among NO, CO, and O2 molecules for adsorption sites, and it varies with reactants compositions, surface composition of adsorbed species, and the reaction temperature.7,41 Both CO and NO exhibit a large extent of adsorption with high sticking coefficient below 550 K; however, the extent of O2 adsorption is predominant between 550 and 900 K on Pd(111) surfaces with high sticking coefficient.11
14 O2 may inhibit the adsorption of other reactants but it also removes adsorbed CO as CO2 and hence helps the overall 15NO reduction reaction. In the case of CO-rich conditions, it is involved in the removal of excess oxygen chemisorbed on the surfaces, especially below 500 K. Other effects, like the influence of 15NO on CO:O2 reaction, need to be considered to understand the overall reactivity of the 15NO:CO:O2 mixture. Results obtained in the present kinetic studies of the ternary 15 NO:CO:O2 mixture on the Pd(111) system are analyzed in light of the above points. 4.1. Influence of O2 on 15NO + CO Reaction. From the results derived from Figures 3, 4, 6, 8, 9, and 10, increase in the oxygen content always results in lesser 15N2 production under SS conditions. However, CO2 production increases as oxygen content increases. Excess oxygen does not seem to poison the Pd surface, irrespective of CO flux. Indeed this conclusion is in good agreement with our earlier results11 on CO:O2 reaction as well as with that of Schalow et al. on Pd/Fe2O3.44 Though the initial sticking coefficient of CO (SO CO on clean Pd(111) is negligible above 550 K,7,41 it increases with increasing subsurface O coverage on Pd(111) from 15NO:CO:O2 mixture g550 K. Indeed this is the reason why the Pd(111) surface is active for oxidation at 750 K with O2-rich compositions (see Figure 6) On the other hand, O atoms occupy the adsorption sites and 15NO adsorption capacity decreases with increasing temperature as well as with increasing θ0 on the surface. Hence a retardation for 15 NO decomposition on the Pd(111) surface was observed due to oxygen at high temperatures (>550 K). NO decomposition also increasingly contributes to the RDS as observed for NO: CO/Pd(111) systems. The conclusion with respect to the O2 on NO:CO is that the presence of oxygen significantly decreases the adsorption as well as dissociation of NO over CO adsorption on Pd(111) resulting a decrease in the N-containing products (15N2 and 15N2O). The suggestion is to maintain a relatively oxygenfree surface to achieve NO reduction on Pd(111) surfaces under SS conditions. 21307

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C 4.2. Effect of 15NO on CO + O2 Reaction. The role of 15NO

on the overall 15NO:CO:O2 reaction could be better understood from the experiments carried out below 600 K. Figures 5 and 7 provide some information about the role of 15NO in 15NO:CO: O2 reactions. It has to be remembered here that 15NO dissociation starts around 400 K but is still not complete at 450 K. Further, under the reaction conditions the equilibrium coverage of CO and oxygen should be high, and hence finding two adjacent vacant sites for 15NO dissociation is difficult. Rate of CO desorption increases above 450 K, and hence sizable 15NO dissociation occurs just above 450 K and starts contributing to the overall reaction. Alternatively, a significant amount of 15NO adsorbs and dissociates (irrespective of O2 content) at the time of initial shutter opening and makes the O and 15N atoms available for CO oxidation and 15N + 15N recombination, respectively. It is well-known that 15N + 15N recombination is a diffusion controlled reaction and a slow process. It is further demonstrated in Figures 1, 2, 5, 9, and 10 that for beam oscillation under SS conditions CO2 partial pressure reaches the new SS instantaneously, whereas N-containing products take at least few seconds (up to 10 s) delay. The presence of adsorbed oxygen retards the 15NO dissociation and hence a sizable molecular 15NO available on the surface makes the 15N2O formation possible. Increasing oxygen content increases CO2 production rate at all the temperatures investigated. However, when we compare these rates of CO2 formation between 15NO + CO + O2 and CO:O2 reactions,7 lower rates were observed e475 K with the former. Relatively large oxygen adsorption >600 K, leading an increasingly O-covered Pd(111) decreases the 15NO adsorption, changes the nature of the overall reaction increasingly toward CO:O2 reaction at high temperatures. Graham et al.35 compared the rate of CO2 production from three different reactions, NO + CO + O2, NO + CO, and CO + O2 on Pd(100) at a partial pressure of 1 Torr and between 430 and 580 K. The rates for CO + O2 and NO + CO reactions follow Arrhenius behavior; however, the rate of NO + CO reaction was 2 orders of magnitude lower than that of CO + O2, and hence the rate of NO + CO + O2 reaction is not different from the latter. The situation regarding the NO + CO + O2 reaction provides a greater contrast between Pd and Rh. In the case of Rh(111), the addition of NO to CO + O2 has been found to greatly reduce the rate of CO2 formation,7,35 whereas Pd(111) demonstrates the reverse trend. Lambert et al45 observed 100% NOx conversion in the presence of CO and H2 as reductants on Pd nanoparticles supported on TiO2/Al2O3. The NCO formation on Pd is critical on titania, whereas alumina promotes the hydrolysis of NCO to NH3, which in turn reduces NOx. Indeed, there is not much work reported on fundamental studies of NO + H2 + O2 on Pd surfaces, and it is important to explore in detail due to the possibility of maximum NO conversion to N2 under net oxidizing conditions.46 4.3. NO Reduction Management under Net-Oxidizing Conditions. NO reduction activity of the catalysts is normally poisoned by oxygen under oxygen-rich conditions, which can be recovered by the exposure of fuel-rich mixtures. It appears that surface-reducing agents (CO in the present case, and VOC and H2 in automotive engines) can effectively remove the surface oxygen as CO2. Hence, the duration of the fuel-rich pulses used to reduce the catalyst becomes critical. Beam switching experiments have been performed between O2-lean and O2-rich conditions to understand the effectiveness of NO reduction under net oxidizing environments. A few points are worth highlighting

ARTICLE

in this section: (a) The surprising thing to be noticed is an increase in the SS CO2 production, even though the FCO decreased to 40% in O2-rich conditions (compared to 60% CO in O2-lean conditions) in Figure 9. (b) More interesting is the observation of large 15N2 production in the TS with O2-rich compositions. As mentioned in the Results section, relatively clean or less-oxygen-covered Pd surfaces help in 15NO dissociation. However, a subsequent O2-rich beam builds up significant oxygen coverage on Pd and hence SS 15NO dissociation decreases. Nonetheless, the above significant θ0 helps for fast CO oxidation in the TS with O2-lean beam and simultaneously significant 15NO reduction and 15N2 production could be observed in the TS. By alternating the composition between O2-rich and O2-lean compositions, oxygen build-up on Pd surface could be minimized, which helps toward NO dissociation. Although the above method would not reach 100% efficiency toward NO reduction, by fast and appropriate cycling between the O2-rich and O2-lean exhaust compositions, a better NO reduction management is possible. Making use of a significantly large NO dissociation on relatively oxygen-free Pd surface is the key finding from this experiment. 4.4. Comparison of 15NO + CO + O2 on Pd(111) and Rh(111) Surfaces. Some new information and the reaction mechanism have been derived from 15NO:CO:O2 on Pd(111) systems. A comparison and contrast could be made between the present results on Pd(111) to that on Rh(111) surfaces. Oxygen was the dominant surface species above 600 K and in O2-rich conditions, which directly affects 15NO dissociation, on both surfaces under SS conditions. In the case of Rh(111), no 15N2O was produced in the reactions reported,2,7 whereas 15N2O production with all 15NO + CO + O2 composition has been observed on Pd(111) surfaces, but the intensity of 15N2O is roughly 8 times less than 15N2 production. On the basis of SS kinetic measurements observed by Zaera et al47,48 on Rh(111) with isotopic nitrogen label switching(14N and 15N), that N2 production under the conditions of the catalytic reaction is likely to involve the formation of an N
NO intermediate. This intermediate is presumed to be common for the production of both N2 and N2O, and the selectivity between the two pathways must be dominated by the relative probabilities for N
NO decomposition (to N2 + adsorbed O) versus molecular N2O desorption. Exclusive N
NO decomposition to N2 seems to be the case on Rh(111),47,48 and both pathways are possible on Pd(111). The main supporting evidence for the claim comes from the slow response of the 15N2 partial pressure trace while blocking and unblocking of the beam seen below 600 K (Figures 1, 2, 3, and 5), which is applicable for Rh(111) also. Nonetheless, the above observations do not rule out the direct N + N recombination to N2 on both surfaces. One of the most important observations is that there is a build-up of a critical coverage of nitrogen atoms on the Rh surface before N2 desorption can be seen.7 However such an observation is not made with 15NO + CO + O2 beams on Pd(111) surfaces. However, 15 NO dissociation step contributes in a major way toward the RDS on both Pd and Rh.30,47,48

5. CONCLUSION A detailed kinetics of the 15NO:CO:O2 reaction was carried out on Pd(111) using a mixed molecular beam in a wide range of temperatures between 400 and 800 K and with different ratios of 21308

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C reactant compositions. The beam compositions varied drastically, between O2-rich and O2-lean compositions. Overall, the maximum reactivity was found between 500 and 600 K for all the beam composition ranges. The kinetic data were presented to demonstrate the role of oxygen in the overall reaction; the main effect of oxygen addition to the 15NO:CO mixture is that it inhibits the NO dissociation rates in the SS and clearly thus enhances the rate of CO2 production under most commonly encountered circumstances. High surface oxygen coverage under the SS conditions decreases the net 15NO adsorption and dissociation, especially above 600 K. However, it helps also to maintain the CO oxidation reaction without any poisoning effect. The overall reaction resulting in N2 production was largely controlled by the 15N + 15N recombination between 450 and 550 K. The 15NO dissociation increasingly contributes to the RDS with increasing temperature as well as the oxygen content in the beam. Pd(111) surface shows reversible oxygen adsorption and without any poisoning effect, irrespective of the oxygen content in the reaction mixture. Large 15NO dissociation occurs on (a) relatively oxygen-free Pd surfaces,and in (b) the transient state and independent of oxygen concentration, which is observed through fast beam switching experiments. Indeed this particular aspect could be utilized for better deNOx management with the present TWC setup in automobiles by fast cycling.

’ ASSOCIATED CONTENT

bS

Supporting Information. Complementary experimental data and discussion to section 3.5.3. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Website: www.ncl.org.in/csgopinath. Present Addresses ‡

National Center for Catalysis Research, Indian Institute of Technology, Madras, Chennai 600 036, India.

’ ACKNOWLEDGMENT S.N. and K.T. thank CSIR for research fellowships. ’ REFERENCES (1) Taylor, K. C. Catal. Rev. Sci. Eng. 1993, 3, 457. (2) Shelef, M.; Graham, G. W. Catal. Rev. Sci. Eng. 1994, 36, 433. (3) Shelef, M. Chem. Rev. 1995, 95, 209. (4) Belton, D. N.; Taylor, K. C. Curr. Opin. Solid State Mater. Sci. 1999, 4, 97. (5) Twigg, M. V. Phil. Trans. R. Soc. A 2005, 363, 1013. (6) Jobson, E. Top. Catal. 2004, 28, 191. (7) Gopinath, C. S.; Zaera, F. J. Catal. 2001, 200, 270. (8) Suhonen, S.; Valden, M.; Hietikko, M.; Laitinen, R.; Savimaki, A.; Harkonen, M. Appl. Catal., A 2001, 218, 151. (9) Weaver, J. F.; Chen, J.-J.; Gerrard, A. L. Surf. Sci. 2005, 592, 83. (10) Shumbera, R. B.; Kan, H. H.; Weaver, J. F. Surf. Sci. 2007, 601, 4809. (11) Nagarajan, S.; Thirunavukkarasu, K.; Gopinath, C. S. J. Phys. Chem. C. 2009, 113, 7385. (12) Thirunavukkarasu, K.; Nagarajan, S.; Gopinath, C. S. Appl. Surf. Sci. 2009, 256, 443. (13) Gopinath, C. S.; Thirunavukkarasu, K.; Nagarajan, S. Chem. Asia. J. 2009, 4, 74.

ARTICLE

(14) Nagarajan, S.; Gopinath, C. S. J. Indian. Inst. Sci. 2010, 90, 245. (15) Zheng, G.; Altman, E. I. Surf. Sci. 2000, 462, 151. (16) Zheng, G.; Altman, E. I. Surf. Sci. 2002, 504, 253. (17) Zheng, G.; Altman, E. I. J. Phys. Chem. B 2002, 106, 1048. (18) Teschner, D.; Pestryakov, A.; Kleimenov, E.; Haevecker, M.; Bluhm, H.; Sauer, H.; Knop-Gericke, A.; Schloegl, R. J. Catal. 2005, 230, 186. (19) Ketteler, G.; Ogletree, D. F.; Bluhm, H.; Liu, H.; Hebenstreit, E. L. D.; Salmeron, M. J. Am. Chem. Soc. 2005, 127, 18269. (20) Ramsier, R. D.; Gao, Q.; Waltenburg, H. N.; Yates, J. T., Jr. J. Chem. Phys. 1994, 100, 6837. (21) Bertolo, M.; Jacobi, K.; Nettesheim, S.; Wolf, M.; Hasselbrink, E. Vacuum 1990, 41, 76. (22) Bertolo, M.; Jacobi, K. Surf. Sci. 1990, 226, 207. (23) Vesecky, S. M.; Koranne, M.; Oh, W. S.; Goodman, D. W. J. Catal. 1997, 167, 234. (24) Ozensoy, E.; Goodman, D. W. Phys. Chem. Chem. Phys. 2004, 6, 3765. (25) Gopinath, C. S.; Zaera, F. J. Catal. 1999, 186, 387. (26) Zaera, F.; Gopinath, C. S. J. Mol. Catal. A 2001, 167, 23. (27) Zaera, F.; Gopinath, C. S. J. Chem. Phys. 1999, 111, 8088. (28) Thirunavukkarasu, K.; Thirumoorthy, K.; Libuda, J.; Gopinath, C. S. J. Phys. Chem. B 2005, 109, 13272. (29) Vesecky, S. M.; Chen, P.; Xu, X.; Goodman, D. W. J. Vac. Sci. Technol. A 1995, 1, 1539. (30) Szanyi, J.; Goodman, D. W. J. Phys. Chem. 1994, 98, 2972. (31) Szanyi, J.; Kuhn, W. K.; Goodman, D. W. J. Phys. Chem. 1994, 98, 2978. (32) Ma, Y. S.; Rzeznicka, I. I.; Matsushima, T. Appl. Surf. Sci. 2005, 244, 558. (33) Ma, Y. S.; Rzeznicka, I. I.; Matsushima, T Cat. Lett. 2005, 101, 109. (34) García, M. F.; Juez, A. I.; Arias, A. M.; Hungría, A. B.; Anderson, J. A.; Conesa, J. C.; Soria, J. J. Catal. 2004, 221, 594. (35) (a) Graham, G. W.; Logan, A. D.; Shelef, M. J. Phys. Chem. 1993, 97, 5445. (b) Graham, G. W.; Logan, A. D.; Shelef, M. J. Phys. Chem. 1994, 98, 1746. (36) Joh
anek, V.; Schauermann, S.; Laurin, M.; Gopinath, C. S.; Libuda, J.; Freund, H.-J. J. Phys. Chem. B. 2004, 108, 14244. (37) Fisher, G. B.; Oh, S. H.; Dimaggio, C. L.; Schmieg, S. J.; Goodman, D. W.; Peden, C. H. F. In Proceedings 9th International Congress on Catalysis; Phillips, M. J.; Ternan, M., Eds, The Chemical Institute of Canada: Ottawa, 1988, p 1355. (38) (a) Thirunavukkarasu, K.; Gopinath, C. S. Catal. Lett. 2007, 119, 50. (b) Nagarajan, S.; Thirunavukkarasu, K.; Gopinath, C. S.; Prasad, S. D. J. Phys. Chem. C. 2011, 115, 15487. (c) Dongare, M. K.; Ramaswamy, V.; Gopinath, C. S.; Ramaswamy, A. V.; Sch€urell, S.; Br€uckner, M.; Kemnitz, E. J. Catal. 2001, 199, 209. (39) (a) Nagarajan, S.; Thirunavukkarasu, K.; Gopinath, C. S.; Counsell, J.; Gilbert, L.; Bowker, M. J. Phys. Chem. C. 2009, 113, 9814. (b) Bowker, M.; Counsell, J.; El-Abiary, K.; Gilbert, L.; Morgan, C.; Nagarajan, S.; Gopinath, C. S. J. Phys. Chem. C. 2010, 114, 5060. (c) Shiju, N. R.; Anilkumar, M.; Mirajkar, S. P.; Gopinath, C. S.; Rao, B. S.; Satyanarayana, C. V. J. Catal. 2005, 230, 484. (40) Thirunavukkarasu, K.; Thirumoorthy, K.; Libuda, J.; Gopinath, C. S. J. Phys. Chem. B. 2005, 109, 13283. (41) Gopinath, C. S.; Zaera, F. J. Phys. Chem. B. 2000, 104, 3194. (42) Engel, T.; Ertl, G. J. Chem. Phys. 1978, 69, 1267. (43) Leisenberger, F. P.; Koller, G.; Sock, M.; Surnev, S.; Ramsey, M. G.; Netzer, F. P.; Kl€otzer, B.; Hayek, K. Surf. Sci. 2000, 445, 380. (44) Schalow, T.; Brandt, B.; Starr, D. E.; Laurin, M.; Shaikhutdinov, S. K.; Schauermann, S.; Libuda, J.; Freund, H.-J. Angew. Chem., Int. Ed. 2006, 45, 3693. (45) Macleod, N.; Cropley, R.; Keel, J. M.; Lambert, R. M. J. Catal. 2004, 221, 20. (46) (a) Shibata, J.; Hashmoto, M.; Shimizu, K.; Yoshida, H.; Hattori, T.; Satsuma, A. J. Phys. Chem. B 2004, 108, 18327. (b) Machida, 21309

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310

The Journal of Physical Chemistry C

ARTICLE

M.; Watanabe, T. Appl. Catal., B 2004, 52, 281. (c) Savva, P. G.; Efstathiou, A. M. J. Catal. 2008, 257, 324. (47) (a) Zaera, F.; Gopinath, C. S. J. Chem. Phys. 2002, 116, 1128. (b) Zaera, F.; Gopinath, C. S. Phys. Chem. Chem. Phys. 2003, 5, 646. (48) (a) Zaera, F.; Wehner, S.; Gopinath, C. S.; Sales, J. L.; Gargiulo, V.; Zgrablich, G. J. Phys. Chem. B. 2001, 105, 7771. (b) Bustos, V.; Gopinath, C. S.; U~nac, R.; Zaera, F.; Zgrablich, G. J. Chem. Phys. 2001, 114, 10927. (c) Zaera, F.; Gopinath, C. S. Chem. Phys. Lett. 2000, 332, 209.

21310

dx.doi.org/10.1021/jp207092s |J. Phys. Chem. C 2011, 115, 21299–21310