# Fileset

[Kurahashi_alignment-controlled-o2-chemisorption-and-catalytic-co-oxidation-on-curved-pt(111).pdf](https://mdr.nims.go.jp/filesets/63535e52-1f13-4883-a12e-f08cac96f4a4/download)

## Creator

[Mitsunori Kurahashi](https://orcid.org/0000-0002-3802-4513)

## Rights

[Creative Commons BY-NC Attribution-NonCommercial 4.0 International](https://creativecommons.org/licenses/by-nc/4.0/)

## Other metadata

[Alignment-Controlled O<sub>2</sub> Chemisorption and Catalytic CO Oxidation on Curved Pt(111)](https://mdr.nims.go.jp/datasets/b17e6445-33f6-4877-a9ec-69e6d2d51d92)

## Fulltext

Alignment-controlled O2 chemisorption andcatalytic CO oxidation on curved Pt(111)Mitsunori Kurahashi∗National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, 305-0047, JapanE-mail: kurahashi.mitsunori@nims.go.jp1AbstractSteric effects in O2 chemisorption and catalytic CO oxidation on stepped Pt(111)surfaces have been investigated by scanning a narrow alignment-controlled O2 beamacross a curved Pt(111) surface, on which the densities of {100} (A-type) and {111}(B-type) steps vary smoothly with the distance from the center of the crystal. Thecontribution of trapping-mediated and activated chemisorption to these processes onthe terraces and steps were discussed based on the dependence of these processes onthe kinetic energy (80-500 meV) and alignment of O2. The results indicate that (i)theprobability of activated O2 chemisorption on surfaces with type-B steps is slightly lowerthan that on surfaces with type-A steps, (ii) CO oxidation proceeds both at terracesand steps at the sample temperature of 693 K while it happens only at terraces at423 K. The results of CO oxidation experiments at 423 K indicate that the rate ofCO2 production at terraces of the stepped Pt(111) surface is lower than at a flat (111)surface and depends on the step structure.IntroductionO2 adsorption and CO oxidation on platinum (Pt) have been investigated intensively dueto their relevance to technologically important catalytic processes such as the O2 reductionreaction in fuel cell and car exhaust gas purification. Since steps are considered to play keyroles in these reactions on nanoparticle catalysts, much effort has been made to clarify themechanism of these processes on stepped Pt using vicinal single-crystalline surfaces as modelsystems.1 A cylindrical crystal where the step density varies smoothly on the surface, 2,3 or acurved crystal, which is a part of the cylindrical single crystal, 4–12 have also been employedsince scanning the analysis point on the surface allows us to investigate the step density andstructure dependence of the reactivity under the same experimental conditions. Studies usingthis methodology have also clarified the subtle effects of steps on the properties of adjacentterraces, which may be difficult to observe with the use of high-index single-crystal surfaces.2For instance, the combined use of high-resolution X-ray photoemission spectroscopy (XPS)and a curved Pt(111) crystal has shown that the electronic properties of microterraces instepped surfaces are not identical to those of a flat surface, which can be attributed to thestrain induced by the step.8 The effects of the step structure on the temperature dependenceof CO oxidation together with the XPS characterization of the chemical species existing onthe surface under millibar pressure have been investigated using Pt, 4 Rh5 and Pd(111)6,7,13curved crystals.On the other hand, the mechanism of gas-surface reactions depends on the kinetic energy,orientation, and internal state of gas-phase molecules that collide with the surface. Molecu-lar beams with well-defined initial states have been used to clarify this. 14,15 Regarding theO2/Pt(111) system, previous studies have shown that O2 dissociates via molecular chemisorp-tion followed by thermal dissociation, and that two basic mechanisms exist for O2 molecularchemisorption.16 One is the trapping-mediated process where O2 is initially trapped into thephysisorbed state and then thermally activated to the chemisorbed state. The other is thechemisorption activated by the kinetic energy of impinging O2. The contributions of thesetwo processes mix in reaction experiments with thermal gases. Non-state-resolved O2 molec-ular beam studies on high index surfaces 17 and curved crystal surfaces11 have shown thatsteps enhance the probability of trapping-mediated chemisorption, while the probabilities ofactivated chemisorption at the steps and terraces are not largely different. State-resolvedO2 beam studies on flat18 and high-index surfaces11,19 have shown that the probability ofactivated chemisorption depends strongly on the O2 alignment relative to the surface lo-cal structure while the trapping-mediated process does not. Considering the fact that thestrain induced by the step affects the electronic properties of terraces as demonstrated byprevious XPS/curved crystal studies,8 it might be reasonable to expect that the step alsoaffects alignment-dependent O2 chemisorption and the subsequent catalytic reactions on theadjacent terrace. However, no attempts have been made to investigate this issue.In the present study, we have studied alignment-dependent O2 chemisorption and cat-3alytic CO oxidation on a curved Pt(111) surface. A narrow state-selected O2 beam wasscanned across a curved crystal to investigate the step density and O2 alignment depen-dence of these processes. Although the overall behavior of the O2 sticking probability canbe understood based on the results of previous studies on flat and stepped Pt(111) surfaces,the results show that O2 sticking probabilities on the terrace of the stepped Pt(111) are notidentical with those on a flat (111) surface. The step density dependence of the CO oxidationrate indicates that the low-temperature CO oxidation proceeds only at the terrace of thestepped surface. Using this fact, we will discuss the difference in the reactivity between aterrace of the stepped Pt and a flat (111) surface.ExperimentsThe experiments were conducted with a single spin-rotational state-selected [(J ,M) =(2,2)]O2 beam prepared by combining a supersonic seeded O2/He beam with a hexapole mag-net.20–22 The translational energy (E0) of the state-selected O2 beam was varied within arange of 0.08-0.5 eV by adjusting the length of the hexapole and tuning the velocity of thesupersonic seeded O2 beam. The O2 internuclear axis is mainly perpendicular to the definingmagnetic field (Hdef ) while its electron spin is oriented parallel to the magnetic field in thisstate.20,21 There are three non-equivalent geometries on a non-magnetic stepped Pt(111) sur-face [Fig. 1(a)]. The O2 axis is mainly parallel to the (111) terrace [helicopter geometry (H)]when Hdef// [111], while parallel and perpendicular O2 are distributed equally (cartwheelgeometry) when Hdef is parallel to the (111) terrace. Two non-equivalent cartwheel geome-tries exist. The O2 axis is perpendicular to the step direction if Hdef//[11̄0] [CZ in Fig. 1(a)]while it is parallel to the step (CY ) if Hdef // [1̄1̄2]. A comparison of the reactivities in theCY and CZ geometries permits the discussion of the alignment effect relative to the stepdirection.11 In this study, however, since we focused our attention to the alignment effectin O2 chemisorption on the terrace of the stepped Pt surfaces, only the difference in the O24sticking probability and CO oxidation between the H and CZ geometries will be presented.The molecular beam apparatus is connected to an analysis chamber where the O2 stickingand surface reaction experiments are conducted. A slit with an opening of ∼0.2×2 mm wasused as the final aperture of the beamline to prepare a narrow aligned O2 beam. The O2sticking probability was measured at normal incidence to the (111) terrace and at a sampletemperature of 313 K using the King and Wells (KW) method 23 with an ion gauge (IG). TheKW traces at about 10 different points on the surface were taken at one time for a cleanedcurved Pt(111) surface by scanning the sample position with a stepping motor. Since theO2 pressure in the analysis chamber during the KW measurement was < 2×10−10 Torr andthe beam irradiation time for taking one KW trace was ∼20 seconds, the exposure to thebackground O2 gas while taking the ten KW traces is estimated to be < 0.04 Langmuirs.The percentage of O2 in the primary beam that passes through the final aperture was foundto be nearly 100% from the gas analysis during the beam irradiation. This comes from thefact that the magnetic hexapole focuses only O2 and the center stop located in the beamlineremoves most of the He atoms contained in the He-seeded O2 beam.21 Isotopically labelled13CO used for monitoring the catalytic CO oxidation reaction on the surface was introducedto the analysis chamber using a gas doser to obtain higher local CO pressures at the sampleposition. Hdef of about 1 Gauss was applied to the sample region using 3 pairs of Helmholtzcoils for controlling the O2 alignment. No difference was found in the IG signal between theH and CZ geometries when the O2 beam was scattered by an inert flag, which was positionedin front of the surface, indicating that the direction of the applied magnetic field does notaffect the flux of the O2 beam. An example of this measurement is shown in Fig. 10 in Ref.[21]. Here, the vacuum vessels that house the ion sources of the IG and the residual gasanalyzer (RGA) were magnetically shielded with permalloy plates to prevent the effects ofthe magnetic field change on their signals during the alignment-controlled experiments.We used a polished curved Pt(111) crystal obtained from Surface Preparation Laboratory,the Netherlands. It is a 31◦ section of a 30 mm diameter cylinder and the surface provides a5continuous range of substrates, from a flat Pt(111) surface up to vicinal angles (α) of ±15.5◦ [Fig. 1(a)]. Steps are along the close-packed [11̄0] direction and α varies along the [1̄1̄2]direction. The (111) terraces of the vicinal surfaces are separated by {100} like steps, whichare denoted as A-type steps, in the [1̄1̄2] direction and by {111}-like steps, which are denotedas B-type steps, in the [112̄] direction. The crystal surface was cleaned in a preparationchamber, which is connected to the analysis chamber mentioned above, by repeating 1 kVAr+ sputtering at 823 K, oxidation at 823 K under the O2 pressure of 1×10−8 Torr, andannealing at 1023 K. The final annealing process before reaction experiments was conductedat 823 K since a previous study showed that the low-temperature annealing prevents therestructuring and faceting of the vicinal surfaces. 9 The preparation chamber is equippedwith a low energy electron diffraction (LEED) optics and a cylindrical mirror analyzer forAuger electron spectroscopy (AES) measurements. The LEED pattern of the surface showedsharp spots from the vicinal surface [Fig. 1(b)]. The spot separation, which is known tobe nearly proportional to the terrace width, 24 varied smoothly with the distance from thecrystal center (y), indicating smooth terrace-width variation on the surface. No impuritywas found within the detection limit of the AES measurements.ResultsO2 sticking probabilityFigure 2(a) shows the initial sticking probability (S0) of O2 for the helicopter (H) andcartwheel (CZ) geometries measured at different positions (y) on the curved Pt(111) crystal.Different panels correspond to the results taken at different O2 kinetic energies. Here, theposition y=0 corresponds to the flat surface, and the A- and B-type steps are distributedat y<0 and y>0, respectively. Figure 2(b) shows the sticking probability difference betweenthe two geometries [∆S0=S0(H)-S0(CZ)]. Contributions of the trapping-mediated process,which is dominant at low E0 and shows no alignment dependence, and the activated O26Figure 1: (a) Three non-equivalent geometries for O2 molecules in the (J ,M)=(2,2) state andthe structure of the curved Pt(111) crystal used in this study. Coordinate y is taken alongthe [112̄] direction and is graduated from the center of the crystal. {100}- and {111}-likesteps exist at the y<0 and y>0 sides, respectively. (b) LEED pattern of the curved Pt(111)surface taken at y=+3 mm with beam energy of 225 eV. The position dependence of theLEED spots indicated by the red dashed rectangle in the upper panel is shown below.7chemisorption, which is dominant at high E0 and shows a large alignment dependence, needto be considered to discuss these results.The S0 and ∆S0 values show the following behaviors at the lowest energy (E0=76 meV).The first panel of Fig. 2(a) shows that S0 increases linearly with the distance from the crystalcenter (|y|) while the slope of the increase is larger for the B-type step side. This can beunderstood based on the previous studies on (553) and (335) surfaces, 11,17 which have shownthat steps enhance precursor-mediated chemisorption and the efficiency is higher for thetype-B step. The linear increase of S0 with |y| can therefore be attributed to the increaseof the probability of the precursor-mediated process with increasing the step density. Incontrast, as shown in Fig. 2(b), ∆S0 is largest at the flat surface (y=0) and decreases largelywith increasing |y|. Here, ∆S0 is associated with the activated process, which happens bothat terraces and steps. To discuss the y dependence of S0 and ∆S0, we need to consider thepercentage of terrace and step areas within the stepped surface as well as the reactivity ofthese areas. The (553) plane, which has a five atom wide terrace, is positioned at y=+3.2mm of the present curved Pt(111) crystal. Although the geometric percentage of the terracearea in the (553) plane is about 60% as will be discussed later, the ∆S0 value at y∼+3 mmis shown to be less than 30 % of the value at y=0.With increasing E0, the y dependence of S0 becomes more gradual while the alignmentdependence in S0 becomes more clear. This reflects the dominance of the activated processat higher E0. As to the alignment effect, Fig. 2(b) shows that the y dependence of ∆S0becomes weaker with increasing E0. It is noted that, at E0 > 200 meV, the S0 values onthe type-B step side are a little lower than those on the type-A step side. The lower S0 onthe type-B step side was not observed in the previous non-state-resolved O2 chemisorptionstudy on curved Pt(111).11 This might be because the measurement was conducted at 150K and with a relatively low beam energy (E0< 260 meV),11 where the contribution of thetrapping-mediated process may be larger than the present case.8Figure 2: (a)The initial sticking probability (S0) of O2 for helicopter (H) and cartwheel (CZ)geometries measured as a function of the position (y) on the Pt(111) curved crystal and (b)the difference in S0 between the two geometries. Different panels correspond to the resultsfor the O2 kinetic energies shown. Error bars were derived from the fluctuation of the iongauge signals used for deriving the S0 values.9CO oxidationTemperature dependence of CO oxidation at low O2 kinetic energyFigure 3 shows the time evolution of the steady-state CO oxidation reaction measured whilethe aligned O2 beam with E0=76 meV irradiates the curved Pt(111) surface under a back-ground CO pressure of 3×10−7 Torr. The three profiles correspond to those measured atsample temperatures (TS) of (a)593 K, (b)523 K and (c)423 K. The CO2 production ratewas monitored with the RGA at the position (y) on the sample shown in Fig. 3(a). Thebeam irradiation (about 160 sec.) and the movement to the next position with the O2 beamoff were repeated. The H and CZ geometries of O2 were alternated every 20 seconds duringirradiation. The H geometry resulted in higher CO2 production rates than the CZ geometrydue to the higher O2 sticking probability for the helicopter geometry. The behavior observedat y∼0 is similar to that for the O2 alignment-controlled CO oxidation on a flat Pt(111). 25The O2 pressure equivalent to the flux of the O2 beam used for measuring the profiles shownin Fig. 3 was roughly estimated to be 3.4×10−6 Torr from the O2 pressure in the analysischamber during the beam irradiation, the nominal pumping speed of the turbomolecularpump and the beam size on the sample. In the present experiment, CO gas was initiallyintroduced to the analysis chamber, and O2 beam irradiation followed. It is known that COoxidation does not occur if the surface is completely covered with CO, but CO oxidationhappens above 305 K on a flat Pt(111) surface since some CO molecules desorb above thistemperature.26 It would be because the sample temperature employed in this study was >305 K that the steady-state CO oxidation reaction was observed.The profile for the CO2 production rate at 593 K [Fig. 3(a)] exhibits a minimum at y∼0and increases with increasing distance from the crystal center (|y|) while the CO2 yields aty> 0 is shown to be higher than at y< 0. These results indicate that CO oxidation happensboth at the terrace and step at 593 K. The behaviors are consistent with the fact that S0increases with increasing |y| at E0=76 meV and S0 is higher on the type-B step side [Fig.10Figure 3: Time evolution of the CO2 production rate measured during O2 beam irradiation(E0=76 meV) on a curved Pt(111) crystal at TS of (a) 593 K, (b) 523 K and (c)423 Kunder the background CO pressure (PCO=3 x 10−7 torr). The reaction measurement at theposition (y) shown in the figure and the movement to the next position with the O2 beam offwas repeated. The H and CZ geometries of O2 were alternated during measurement. TheH geometry resulted in higher CO2 production rates.112(a)]. The difference in the CO2 yield between the H and CZ geometries is largest for theflat surface and decreases with increasing |y|. This is also consistent with the decrease of∆S0 with increasing |y| [Fig. 2(b)]. The similarity of the y dependence of the CO2 yield at593 K and the O2 sticking probability at the clean Pt surface suggests that the rate constantof CO oxidation at 593 K is similar between the step and the terrace at this CO pressurecondition.Figures 3(a)-(c) show that the CO2 production rate changes significantly with temper-ature at around 523 K while the threshold temperature, so-called ignition temperature,depends on the step density. At y∼0, the CO2 yield at 423 K is about 10% of that at 593K while the yields at 523 K and 593 K are similar. This is consistent with previous stud-ies showing that the CO2 production rate on a flat Pt(111) surface increases drastically at500-550 K with increasing temperature and changes gradually at higher temperatures. 2,4,27Therefore, we judged that the CO oxidation rate at the flat surface region is saturated at593 K, although experiments at higher temperatures were not conducted. In contrast, aty∼±3 where the step density is high, the CO2 yield at 523 K is 30-50% of that at 593 K.The profile at 423 K [Fig. 3(c)] shows that the CO2 yield is highest at y∼0 and decreasesmonotonically with increasing |y|. These behaviors indicate that the low-temperature COoxidation is inefficient at steps and the profile at 423 K virtually reflects the reactions atthe terrace of the stepped surface. A previous thermal desorption study 28 on Pt(335) hasshown that the binding energy of CO at the step site is higher than that at the terracesite. Oxidation of CO molecules adsorbed at steps would, therefore, require higher sampletemperatures, causing the inefficient low-temperature CO oxidation at steps. In the presentexperiment, since the surface was initially exposed to the background CO gas, step siteswould be blocked by strongly bound CO molecules and would not contribute to the COoxidation reaction at 423 K.It is possible that the CO oxidation changes the atomic arrangement at the step sincethere have been experimental evidences showing that O2 chemisorption induces the step12doubling of transition metal surfaces. 29,30 The structural analysis with LEED at the positionof the O2 beam irradiation was not conducted in this study. However, Fig. 3(a) showsthat the CO2 yield varies smoothly with the distance from the crystal center at 593 K,suggesting that the step density of the clean surface would not be changed largely in thepresent conditions.Low-temperature CO oxidation at high O2 kinetic energyFigure 4(a) shows the CO oxidation profile measured with the aligned O2 beam of E0=275meV at 423 K. Since trapping-mediated O2 chemisorption is inefficient at this energy 11,16–18and steps do not contribute to CO oxidation at 423 K as shown above, these profiles reflectCO oxidation caused by the activated O2 chemisorption at terraces of the stepped Pt. Thethree panels correspond to the profiles at different relative CO pressures.We first discuss the CO pressure dependence of the profile at y∼0. Figure 4(a) shows thatthe ratio of the CO2 intensity for the helicopter geometry [I(H)] to that for the cartwheelgeometry [I(CZ)] is about 1.4 at P(CO)/Pe(O2) =0.15 [the top panel of Fig. 4(a)] whileit increases to ∼1.8 at P(CO)/Pe(O2)> 0.3. On the other hand, Fig. 2(a) indicates thatthe sticking probability ratio between the H and CZ geometries [S0(H)/S0(CZ)] at E0=275meV and at y∼ 0 is ∼ 1.7, which is similar to the value of I(H)/I(CZ) at P(CO)/Pe(O2)>0.3. This can be understood as follows. CO oxidation on Pt proceeds via the Langmuir-Hinshelwood mechanism, in which O2 chemisorption followed by a reaction of adsorbed Oand CO forms CO2.27 The CO2 production rate would be proportional to the product ofthe concentrations of CO and atomic O on the surface. In the steady-state reaction, the COconcentration is determined by competition between CO adsorption from the gas phase andits consumption for CO2 production. Because the sticking probability of helicopter O2 [S(H)]is higher than that for cartwheel O2 [S(CZ)], the CO concentration during the steady-statereaction for helicopter O2 would be lower because more CO molecules are consumed for CO2production. This would result in I(H)/I(CZ) < S(H)/S(CZ). However, if the CO pressure is13Figure 4: (a) Time evolution of the CO2 production rate measured while the O2 beam withE0=275 meV irradiates the position (y) on the curved Pt(111) at TS=423 K. The H and CZgeometries of O2 were alternated during measurement. The three panels correspond to theresults taken under different background CO pressures. The ratio between the backgroundCO pressure and the O2 pressure equivalent with the beam flux [P e(O2)] is shown. (b)Theposition dependence of the CO2 production rate at the helicopter geometry normalized bythe value at y=0. The ratio of the terrace atoms within the surface is shown with solid bluelines. (c)The position dependence of the difference in the CO2 production rate between theH and CZ geometries normalized by the value at y=0. Error bars shown in (b) and (c) wereestimated from fluctuations in the CO2 signal.14high enough to cause the CO adsorption rate ≫ the O2 sticking probability, the CO coverageduring the reaction for the two geometries would be similar. In such conditions, the CO2production rate is expected to be proportional to the O2 sticking probability.We secondly discuss the CO oxidation at stepped surfaces. The profile at P(CO)/Pe(O2)= 0.15 shows that the ratio I(H)/I(CZ) increases with increasing |y|. For example, thisratio is ∼1.4 at y=-0.3 while it is ∼1.8 at y=2.7. If the reactivity of the microterrace ofthe stepped surface is identical with that of the flat surface, I(H)/I(CZ) for the microterraceshould be identical to that for the flat surface. In addition, the profiles on the A- and B-typestep sides are not symmetrical. It is shown that the CO2 yield is lower and I(H)/I(CZ) ishigher on the B-type step side. These behaviors may be accounted for if the O2 stickingprobability is lower on the type B step side. The fact that S0 on the B-stepped surface isslightly lower at E0=275 meV [Fig. 2(a)] might be associated with this.Figures 4(b) and (c) summarize the position dependence of the CO2 yield for the he-licopter geometry [ICO2(H)] and the difference between the H and CZ geometries [∆ICO2 ],respectively. The data taken at high CO pressures where the CO2 production rate may beproportional to the O2 chemisorption probability are plotted. These quantities are normal-ized by the values at y=0, which were estimated by the fit of the data points at -1<y< 1to parabolic functions. To discuss the contribution of the terrace area to the reactivity ofthe stepped Pt surfaces, the ratio of the number of terrace atoms (NT ) to the total num-ber of terrace, edge and corner atoms on the surface [nT=NT/(NT+2)] is shown with solidblue lines. The data points for ICO2(H) and ∆ICO2 are expected to be around these linesif the terrace of the stepped surface has the same reactivity with a flat (111) surface andthe step gives no contribution to ICO2(H) and ∆ICO2 . In previous O2 chemisorption studyon Pt(553),17 ∼50% of the surface area was assumed to act like the Pt(111) surface fordiscussing the contribution of the terrace area to the measured sticking probability. The(553) surface, which has five atom wide terraces (nT = 0.6), is at y=+3.2mm in the presentcurved Pt(111) crystal.15Figures 4(b) and (c) show that, unlike S0 and ∆S0 at E0=275 meV [Fig. 2(a)], ICO2(H)and ∆ICO2 decrease largely with increasing |y|. This reflects the following situations. Acti-vated O2 chemisorption occurs at the terraces and steps and its probability depends on theO2 alignment. The contribution of the terraces to S0 and ∆S0 decreases while that of thesteps increases with increasing |y|, resulting in the weak position dependence of S0 and ∆S0.However, since the low-temperature CO oxidation happens only at the terrace, ICO2(H) and∆ICO2 decreases with the decrease of the terrace area. ICO2(H) and ∆ICO2 show similarprofiles, implying that the O2 alignment effects in CO oxidation at the microterraces andthe flat surface are similar. We note that the normalized values for ICO2(H) and ∆ICO2 are20-50 % lower than the ratio of the terrace atoms within the surface (solid blue lines).Figure 5: (a) Time evolution of the CO2 production rate measured while the O2 beam withE0=275 meV irradiates the indicated position (y) of the curved Pt(111) crystal at 523 K.The H and CZ geometries of O2 were alternated during measurement. The three panelscorrespond to the results taken under different background CO pressures. The positiondependence of (b) the CO2 production rate at the helicopter geometry and (c) the differencein the CO2 production rate between the H and CZ geometries normalized by the value aty=0. Error bars shown in (b) and (c) were estimated from fluctuations in the CO2 signal.Figure 5 shows the CO2 production rate measured at 523 K and E0=275 meV. As shown16in Fig. 3, this temperature is around the ignition temperature of CO oxidation on theflat surface. It is shown that the CO2 production rate changes gradually with y at highlystepped regions (|y| >2). This reflects the situation that CO oxidation also happens at stepsat 523 K although its efficiency is lower than at the terrace. As to the shape of the profilefor the CO2 production rate at around y=0, it is found from the comparison of Figs. 4(b)and 5(b) that the width of the profile at 523 K is narrower than at 423 K, and that theCO2 production rate at 523 K decreases more steeply with increasing |y|. This cannot beattributed to the effect of the temperature on the O2 chemisorption step since the activatedO2 chemisorption dominant at this energy is insensitive to TS. We speculate that the stepmight affect the ignition temperature of CO oxidation on the adjacent terrace, althoughadditional information is needed for further discussions.DiscussionThe present results indicate that the reactivity of the terrace of the stepped surface is likelyto be different from that of the flat surface. The previous XPS study on curved Pt(111) 8have shown that the Pt 4f level of the terrace surface shifts to the higher binding energyside with increasing the step density. This has been associated with the lattice contractioninduced by the step. The contraction of ∼2% has been estimated for the terrace of (335) and(553) surfaces while the estimated contraction is slightly larger for the (553) surface withthe B-type step.8 On the other hand, a theoretical study by Mavrikakis at al. predicted thatthe lattice contraction reduces the adsorption energy of molecules, 31 while the O2 reductionefficiency on strained Pt films on nanoparticles has been shown to be largely different fromthat on bulk Pt.32 Considering the results of these studies, it may be reasonable to expectthat the lattice contraction increases the activation energy for molecular chemisorption,reducing the probability of the activated chemisorption of O2. Fig. 2(b) shows that ∆S0decreases largely with the step density at E0=76 meV. The values at highly stepped regions17can not be explained if we assume that the terrace of the stepped Pt has the same reactivitywith a flat Pt(111) surface. For example, the ∆S0 value at y∼2 mm, where the n(T ) valueshown in Fig. 4 is ∼ 0.7, is much less than 0.7 x ∆S0 (y=0). In addition, Figure 4 shows thatICO2(H) and ∆ICO2 for the low-temperature CO oxidation on the terrace of stepped Pt areconsiderably smaller than those expected for the reactivity of a flat Pt(111) surface. Theseresults may be consistent with the lattice contraction induced by the step, which wouldincrease the activation energy for chemisorption at the terrace of the stepped Pt, because,in such cases, the O2 sticking probability and its alignment dependence would also becomesmaller.It is known that S0 of O2 on a flat Pt(111) surface is much less than unity (< 0.3) evenwhen the O2 kinetic energy is increased to ∼ 1.0 eV.16,18 This has been associated witha situation in which, although energetic O2 molecules can surmount the activation barrierfor molecular chemisorption, the probability of being scattered off from the surface by therepulsive wall of the interaction potential increases with increasing E0.18 We speculate that,if the lattice contraction increases the activation barrier for O2 chemisorption, it wouldalso increase the probability of scattering, making the sticking probability of energetic O2molecules lower than that on a flat Pt(111) surface. The lower S0 on the B-step side at highE0 conditions, which is shown in Fig. 2(a), may be consistent with the theoretical simulationpredicting that the lattice contraction induced by the B-type step is larger than that inducedby the A-type step.8To discuss the origin of the reduced reactivity of the microterraces, we may also have toconsider the effective terrace width. The percentage of the terrace area in stepped surfacesshown in Figs. 4 and 5 was estimated by assigning the edge and corner atoms as the step area.The effective terrace width can be narrower than the width estimated with this assumptionbecause the corner and edge atoms, which have been predicted to be strongly strained, 8 canaffect the potential of neighboring area. If, for example, we assign additional 0.5 atom row tothe step area, the percentage of the terrace area is reduced to 50% for (553) surface, which is18the value employed by Jacobse et al.17 To make quantitative discussions as to how much thereactivity of the microterrace is reduced, additional information about the effective terracewidth would be necessary.Figure 3 shows that CO oxidation at low temperature is less efficient at steps than atterraces. This appears inconsistent with the results of the recent near-ambient pressure(NAP)-XPS study4 of CO oxidation on curved Pt(111), which showed that stepped andflat Pt(111) surfaces have identical ignition temperatures. The NAP-XPS study of COoxidation on curved Pt(111) was conducted with a CO and O2 gas mixture at millibarpressures, and the presence of subsurface oxygen on the (111) terrace was suggested tocause the identical ignition temperature. In the present experiment, however, the O2 beamirradiates the CO-exposed Pt surface where no subsurface oxygen exists. The difference inthe reaction conditions between the two experiments might result in the different temperaturedependences of CO oxidation.ConclusionBy scanning a narrow rotational-state-selected O2 beam on a curved Pt(111) surface wherethe step density varies smoothly across the crystal, we have investigated the effect of thestep on the alignment-dependent O2 chemisorption and catalytic CO oxidation on vicinalPt(111) surfaces. To the best of our knowledge, this is the first state-selected molecularbeam reaction experiment on a curved crystal surface. It has been shown that this approachhelps us to separate contributions of the precursor-mediated and activated processes onterraces and steps in vicinal surfaces. The increase in the probability of trapping-mediated O2chemisorption with increasing the step density, its step-type dependence and the dominanceof activated chemisorption at high energy conditions were observed and found to be consistentwith previous studies on stepped Pt. However, the results show that the probability ofactivated chemisorption on surfaces with type-B steps is slightly lower than on surfaces with19type -A steps. The step density dependence of the low-temperature (423 K) CO oxidation,which was found to occur only at terraces of stepped Pt surfaces, indicates that the CO2production rate at the terraces of the stepped Pt is lower than at a flat (111) surface and alsodepends on the step structure. One of the origins of these behaviors may be that the latticecontraction in the microterrace induced by the step lowers the probability of activated O2chemisorption and causes reactivity lower than that of a flat (111) surface.AcknowledgementsThe author acknowledges Professor Ludo Juurlink for his instruction regarding how to pre-pare a clean, curved Pt(111) crystal surface. This work was supported by JSPS KAKENHIGrant Numbers 20H02623 and 24K01349, as well as the Iketani Science and TechnologyFoundation.References(1) Van Spronsen, M. A.; Frenken, J. W.; Groot, I. M. Surface science under reactionconditions: CO oxidation on Pt and Pd model catalysts. Chemical Society Reviews2017, 46, 4347–4374.(2) Hopster, H.; Ibach, H.; Comsa, G. Catalytic oxidation of carbon monoxide on steppedplatinum (111) surfaces. Journal of Catalysis 1977, 46, 37–48.(3) Sander, M.; Imbihl, R.; Ertl, G. Kinetic oscillations in catalytic CO oxidation on acylindrical Pt single crystal surface. The Journal of Chemical Physics 1992, 97, 5193–5204.(4) Garcia-Martinez, F.; Garćıa-Fernández, C.; Simonovis, J. P.; Hunt, A.; Walter, A.;Waluyo, I.; Bertram, F.; Merte, L. R.; Shipilin, M.; Pfaff, S. et al. Catalytic Oxidation20of CO on a Curved Pt (111) Surface: Simultaneous Ignition at All Facets through aTransient CO-O Complex. Angewandte Chemie 2020, 132, 20212–20218.(5) Garćıa-Mart́ınez, F.; Rämisch, L.; Ali, K.; Waluyo, I.; Bodero, R. C.; Pfaff, S.; Villar-Garćıa, I. J.; Walter, A. L.; Hunt, A.; Pérez-Dieste, V. et al. Structure matters: asym-metric CO oxidation at Rh steps with different atomic packing. Journal of the AmericanChemical Society 2022, 144, 15363–15371.(6) Schiller, F.; Ilyn, M.; Pérez-Dieste, V.; Escudero, C.; Huck-Iriart, C.; Ruiz del Arbol, N.;Hagman, B.; Merte, L. R.; Bertram, F.; Shipilin, M. et al. Catalytic oxidation of carbonmonoxide on a curved Pd crystal: spatial variation of active and poisoning phases instationary conditions. Journal of the American Chemical Society 2018, 140, 16245–16252.(7) Blomberg, S.; Zetterberg, J.; Zhou, J.; Merte, L. R.; Gustafson, J.; Shipilin, M.;Trinchero, A.; Miccio, L. A.; Magaña, A.; Ilyn, M. et al. Strain dependent light-offtemperature in catalysis revealed by planar laser-induced fluorescence. ACS Catalysis2017, 7, 110–114.(8) Walter, A. L.; Schiller, F.; Corso, M.; Merte, L. R.; Bertram, F.; Lobo-Checa, J.;Shipilin, M.; Gustafson, J.; Lundgren, E.; Brión-Ŕıos, A. X. et al. X-ray photoemissionanalysis of clean and carbon monoxide-chemisorbed platinum (111) stepped surfacesusing a curved crystal. Nature communications 2015, 6, 8903.(9) Roorda, T.; Auras, S. V.; Juurlink, L. B. Chiral Surface Characterisation and ReactivityToward H–D Exchange of a Curved Platinum Crystal. Topics in Catalysis 2020, 63,1558–1568.(10) Van Lent, R.; Auras, S. V.; Cao, K.; Walsh, A. J.; Gleeson, M. A.; Juurlink, L. B.Site-specific reactivity of molecules with surface defects-the case of H2 dissociation onPt. Science 2019, 363, 155–157.21(11) Cao, K.; van Lent, R.; Kleyn, A. W.; Kurahashi, M.; Juurlink, L. B. Steps on Ptstereodynamically filter sticking of O2. Proceedings of the National Academy of Sciences2019, 116, 13862–13866.(12) Auras, S. V.; Juurlink, L. B. Recent advances in the use of curved single crystal surfaces.Progress in Surface Science 2021, 96, 100627.(13)(14) Auerbach, D. J.; Tully, J. C.; Wodtke, A. M. Chemical dynamics from the gas-phaseto surfaces. Natural Sciences 2021, 1, e10005.(15) Kleyn, A. Molecular beams and chemical dynamics at surfaces. Chemical Society Re-views 2003, 32, 87–95.(16) Luntz, A.; Williams, M.; Bethune, D. The sticking of O2 on a Pt (111) surface. TheJournal of Chemical Physics 1988, 89, 4381–4395.(17) Jacobse, L.; den Dunnen, A.; Juurlink, L. B. The molecular dynamics of adsorption anddissociation of O2 on Pt (553). The Journal of Chemical Physics 2015, 143, 014703.(18) Ueta, H.; Kurahashi, M. Dynamics of O2 Chemisorption on a Flat Platinum SurfaceProbed by an Alignment-Controlled O2 Beam. Angewandte Chemie International Edi-tion 2017, 56, 4174–4177.(19) de Willigen, M. J.; Kurahashi, M.; Juurlink, L. B. Alignment and impact angulardependence to O2 sticking and dissociation on Pt (111) and close-packed steps. PhysicalChemistry Chemical Physics 2022, 24, 18227–18235.(20) Kurahashi, M.; Yamauchi, Y. Steric effect in O2 sticking on Al (111): Preference forparallel geometry. Physical Review Letters 2013, 110, 246102.(21) Kurahashi, M. Oxygen adsorption on surfaces studied by a spin-and alignment-controlled O2 beam. Progress in Surface Science 2016, 91, 29–55.22(22) Kurahashi, M.; Yamauchi, Y. Production of a single spin-rotational state [(J, M)=(2,2)] selected molecular oxygen (3Σ−g ) beam by a hexapole magnet. Review of ScientificInstruments 2009, 80, 083103.(23) King, D. A.; Wells, M. G. Molecular beam investigation of adsorption kinetics on bulkmetal targets: Nitrogen on tungsten. Surface Science 1972, 29, 454–482.(24) Henzler, M. LEED-investigation of step arrays on cleaved germanium (111) surfaces.Surface Science 1970, 19, 159–171.(25) Ueta, H.; Kurahashi, M. Steric effect in CO oxidation on Pt (111). The Journal ofChemical Physics 2017, 147, 194705.(26) Burnett, D. J.; Capitano, A. T.; Gabelnick, A. M.; Marsh, A. L.; Fischer, D. A.;Gland, J. L. In-situ soft X-ray studies of CO oxidation on the Pt (111) surface. SurfaceScience 2004, 564, 29–37.(27) Campbell, C.; Ertl, G.; Kuipers, H.; Segner, J. A molecular beam study of the catalyticoxidation of CO on a Pt (111) surface. The Journal of Chemical Physics 1980, 73,5862–5873.(28) Luo, J.; Tobin, R.; Lambert, D. K.; Fisher, G. B.; DiMaggio, C. L. CO adsorption siteoccupation on Pt (335): a quantitative investigation using TPD and EELS. SurfaceScience 1992, 274, 53–62.(29) Comsa, G.; Mechtersheimer, G.; Poelsema, B. He-beam scattering study of the dynam-ics of oxygen induced reconstruction of the Pt (997) surface: I. Dynamics of doublestep formation and destruction. Surface Science 1982, 119, 159–171.(30) Hoogers, G.; King, D. Adsorbate-induced step-doubling reconstruction of a vicinalmetal surface: oxygen on Rh {332}. Surface science 1993, 286, 306–316.23(31) Mavrikakis, M.; Hammer, B.; Nørskov, J. K. Effect of strain on the reactivity of metalsurfaces. Physical Review Letters 1998, 81, 2819.(32) Temmel, S. E.; Fabbri, E.; Pergolesi, D.; Lippert, T.; Schmidt, T. J. Investigating therole of strain toward the oxygen reduction activity on model thin film Pt catalysts.ACS Catalysis 2016, 6, 7566–7576.24TOC Graphic25