Applied Surface Science 256 (2009) 371–375

Oxygen induced facet formation on Rh(2 1 0) surface

Govind a,*, Wenhua Chen b, Hao Wang b, T.E. Madey b

a Surface Physics & Nanostructures Group, National Physical Laboratory, Dr. K.S. Krishnan Road, New Delhi 110012, Indiab Department of Physics & Astronomy and Laboratory for Surface Modification, Rutgers, The State University of New Jersey, 136 Frelinghuysen Road, Piscataway, NJ 08854, USA

A R T I C L E I N F O

Article history:

Available online 27 May 2009

Keywords:

Faceting

Rhodium

Auger

LEED

STM

A B S T R A C T

Oxygen induced nanometer-scale faceting of the atomically rough Rh(2 1 0) surface has been studied

using Auger electron spectroscopy (AES), low energy electron diffraction (LEED), and scanning tunneling

microscopy (STM). The Rh(2 1 0) surface completely covered with nanometer-scale facets when

annealed at �550 K in the presence of oxygen. LEED studies reveal that the pyramidal faceted surface is

characterized by three-sided nanoscale pyramids exposing (7 3 1), (7 3 �1) and (1 1 0) faces. A clean

faceted surface was prepared through the use of low temperature surface cleaning method using the

reaction with H2 while preserving (‘‘freezing’’) the pyramidal facet structure. The resulting clean faceted

surface remains stable for T � 600 K and for higher temperatures; the faceted surface irreversibly relaxes

to the planar surface. STM measurements confirms the formation of nanopyramids with average

pyramid size ranging from 12 to 21 nm depending upon the annealing temperature. The nanopyramidal

faceted Rh surface may be used as a potential template for the growth of metallic nanoclusters and for

structure sensitive reactions.

� 2009 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Applied Surface Science

journa l homepage: www.e lsev ier .com/ locate /apsusc

1. Introduction

Atomically rough clean metal surfaces generally have lowersurface atom density and higher surface free energy than close-packed surfaces of the same metal. A variety of atomically roughclean metal surfaces can be prepared as stable orientations, but thepresence of (strongly interacting) adsorbate can cause changes insurface morphology through mechanisms such as reconstructionand facet formation [1–6]. These morphological changes areusually explained in terms of changes in surface free energy due tothe presence of adsorbate [7,8]. There are many overlayer/substrate systems that exhibit faceting, including oxygen andmetal covered e.g., bcc W(1 1 1), Mo(1 1 1), fcc Cu(2 1 0), Ir(2 1 0),Pt(2 1 0) and Ni(2 1 0) and hcp Re(1 2 �3 1) and Re(1 1 �2 1)surfaces [3–11]. It has been observed that changes in morphologycan be accompanied by changes in electronic structure and surfacereactivity [12,13].

In the present study, we discuss the oxygen inducedmorphological change of Rh(2 1 0) surface which is an importantsubstrate for catalytic processes and has potential applications insurface chemistry. Previous studies on fcc(2 1 0) surfaces [5–9]indicate that the facets formation can be induced by the presenceof various gaseous adsorbates for example oxygen and nitrogencan induce faceting on Cu(2 1 0) and Ni(2 1 0) while oxygen and

* Corresponding author. Tel.: +91 11 45608403; fax: +91 11 45609310.

E-mail address: govind@mail.nplindia.ernet.in (Govind).

0169-4332/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.apsusc.2009.05.042

CO provide favorable conditions to restructure the Pt(2 1 0)surface to form facets. Ir(2 1 0) [8] shows the formation ofpyramidal type facets with {3 1 1} and (1 1 0) facets, whenannealed in the presence of oxygen above 600 K. In a recent study,Govind et al. [14] has studied the oxygen induced faceting ofRh(2 1 0) and revealed the various condition for the formation offaceted surface.

Rhodium is an fcc metal of the platinum group with meltingpoint (Tm � 2239 K) and an ideal model of its surface is illustratedin Fig. 1. The bulk truncated surface is atomically very open andrough, with four exposed layers of atoms. The top layer has C2v

symmetry (1808 rotation, with the [2 1 0] vector as the z axis andprincipal axis, as well as one vertical reflection plane defined by the[2 1 0] and [�1 2 0] vectors). The z-periodicity of the structure is 10layers – the atoms of the 11th layer are in the same x and y

positions as the atoms in the top layer. In this paper, motivation isto study the oxygen induced morphological instability of atomicrough Rh(2 1 0) metal surface and study the oxygen induced facetformation, destruction and freezing of the clean faceted Rh(2 1 0)surface. The paper is arranged in following manner: the experi-mental procedure is presented in Section 2 while results of AES,LEED and STM are presented in Section 3. Further the conclusionsof the present study are presented in Section 4.

2. Experimental

The experiments were carried out in two different ultrahighvacuum (UHV) chambers denoted as LEED and STM chambers,

Fig. 1. Hard sphere model of fcc(2 1 0) surface (top and side view) showing top four

exposed layer.

Fig. 2. The oxygen uptake curve on Rh(2 1 0) at room temperature measured by AES.

Govind et al. / Applied Surface Science 256 (2009) 371–375372

respectively. All LEED images are obtained in the LEED chamberthat also contains a quadrupole mass spectrometer (QMS) forresidual gas analysis and a cylindrical mirror analyzer for AES. STMexperiments were performed in a chamber at room temperatureusing a hybrid variable temperature Omicron STM with tungstentips. The Rh(2 1 0) crystal is cut from a single crystal Rh (99.99%)rod � 10 mm in diameter, �1.5 mm thick, aligned within 0.58 ofthe (2 1 0) orientation and polished to a mirror finish. The sample issupported by two rhenium leads where a high current (up to 30 A)can be passed through the leads to achieve temperatures up to1600 K. A C-type (W–5 at.% Re/W–26 at.% Re) thermocouple is spotwelded directly to the rear of the sample for accurate temperaturemeasurement. The sample support assembly also includes atungsten filament for electron bombardment heating; a tempera-ture up to 2000 K can be achieved by flashing the sample in UHV.The sample was cleaned by repeated cycles of 1 KeV Ar+ ionbombardment initially at 300 K and then at increasingly highertemperature (<650 K), annealing in UHV at 1200 K, annealing in O2

(2 � 10�8 Torr) at 1000–1200 K followed by rapid flashes to�1600 K in UHV to desorbs the excess oxygen from the surface.The oxygen gas deposition was achieved by back-filling thechamber by O2.

Fig. 3. (a) LEED pattern of Rh(2 1 0) clean surface at incident electron beam ene

3. Results

We study the adsorption of oxygen on planar Rh(2 1 0) at roomtemperature (Fig. 2). The curve shows the O/Rh Auger peakintensity ratio as a function of oxygen dose at room temperature;the oxygen dose is expressed in units of Langmuir (L, 1L = 1 �10�6 Torr s = 1.33 � 10�4 Pa s). The O/Rh Auger ratio increaseswith oxygen dose and reach at the saturation at 10L of oxygen dose,which indicate the formation of one physical monolayer of oxygenat room temperature.

The low energy electron diffraction pattern obtained from cleanRh(2 1 0) is shown in Fig. 3a. The observed LEED pattern isconsistent with the presence of an unreconstructed bulk fcc(2 1 0)plane. When electron beam energy (Ee) is increased the electronwavelength and, consequently, the diffraction angles decrease, andan apparent motion of the diffraction beams toward the specularlyreflected (0, 0) beam is observed (Fig. 3b). For the clean Rh(2 1 0)surface the (0, 0) beam is perpendicular to the macroscopic surfaceplane and is in the center of the LEED pattern. The behavior of theLEED beams indicates that the surface is microscopically planar. Noadditional beams appear in the LEED pattern as a result of oxygenadsorption at room temperature. However, an increase in thebackground intensity is observed upon oxygen deposition whichcan be attributed to additional diffuse scattering from the oxygenoverlayer and oxygen induced disorder in the topmost Rh layer.

rgy 160 eV. (b) LEED pattern from planar Rh(2 1 0) with energy 20–80 eV.

Fig. 4. (a) LEED pattern from faceted O/Rh(2 1 0) with incident electron beam energy (a) 66 eV; (b) 20–42 eV. The position of specular beam has been shown by circles.

Govind et al. / Applied Surface Science 256 (2009) 371–375 373

To study the oxygen induced faceting of Rh(2 1 0) surface, theRh surface was pre-dosed with 10L oxygen followed by annealingto elevated temperature in presence of oxygen atmosphere(2 � 10�8 Torr). The behavior of the oxygen-covered surfacechanges gradually with sample annealing temperature. As thesample temperature is increased the LEED beams of the initialplanar (1 � 1) pattern (shown in Fig. 3a) slowly become diffuse anda set of new beam appeared in the LEED pattern is startedapproximately at 450 K. With further temperature increase the(1 � 1) beams become progressively fainter, and the emerging newbeams become gradually brighter and sharper. This processcontinues until the sample temperature reached to 550 K, whereall traces of the (1 � 1) beams disappeared. The observed patternshow a completely different geometry and have roughly threetimes more beams replaces the original one (Fig. 4a). The geometryof the LEED pattern does not change with further (above 550 K)temperature increase but changes of the spot size can be observed.Subsequent cooling of the sample to room temperature does notreverse the process and the new pattern remains stable in theentire temperature range. On increasing electron beam energy, themotion of the new LEED pattern differs from that of the (1 � 1), i.e.,the diffracted beams move in directions that converge to threedistinct points (Fig. 4b). They can therefore be identified asspecular reflection beams originating from surfaces that are tiltedwith respect to the (2 1 0) macroscopic surface plane. Thecomplete LEED pattern is the superposition of three distinct LEEDpatterns, whose respective specular directions are each inclined atunique angles determined by the crystallographic orientationof the surfaces from which they originate. Subsequently, theemergence of these patterns has been interpreted to be aconsequence of the formation of nanometer-size facets on thesubstrate surface. The LEED patterns of the faceted surface exhibitreflection symmetry with respect to the plane defined by the[2 1 0] and [�1 2 0] vectors which is also the plane of reflectionsymmetry of the bulk crystal. This fact, combined with thepresence of three specular spots in the LEED pattern, indicates thatthe facets are the faces of three-sided nanopyramids, withreflection symmetry. The faceted Rh(2 1 0) surface found to bestable upto 850 K in UHV and irreversibly relax to planar surface onannealing to higher temperature.

For experiments performed on Rh surface the cleanliness iscritical: carbon contamination can obstruct or even completelyavert the facet formation. It is observed that after numerouscleaning cycles, the only contaminant detected was carbon, inamounts that were comparable to the noise level and therefore

difficult to estimate, but are of the order of one percent of amonolayer or less. The oxygen absorbed on the surface interactswith carbon contaminant and reduced from the surface due tovariety of chemical reactions. The new LEED pattern develops evenfor a moderately carbon-contaminated surface and remainspresent provided that the sample is in an oxygen atmosphere orits temperature is comfortably under 600 K. In cases of highercarbon concentration, the faceted pattern promptly and irrever-sibly disappears (the surface relaxes into its original planar state)upon heating at temperatures as low as 600 K. Under conditions ofsuch carbon contamination facets can be only formed by heatingthe sample in oxygen atmosphere, not by heating oxygen-coveredsample in UHV. In view of this all the experiment described hereare comprise of 10L dose of oxygen on planar Rh(2 1 0) followed byannealing the surface to elevated temperature in presence ofoxygen atmosphere (2 � 10�8 Torr) followed by cooling to roomtemperature.

The crystallographic orientation of these facets, the position ofeach individual specular beam, can be determined by calculatingthe corresponding facet tilt angles with respect to (2 1 0) plane,and the azimuthal orientation of facets with respect to each other.It is observed that out of three facet planes, two facet planes areidentical and tiled to similar angle to the plane normal to (2 1 0).The measured angle between one facet plane corresponding to thespecular beam and the normal direction of the (2 1 0) plane whichcomes out to be 8.2 � 0.18 and due to the symmetry, the tilt anglebetween the other facet planes and the (2 1 0) plane is also 8.2 � 0.18.The azimuthal angle between these two faceted planes can be directlymeasured and is F = 132 � 48. The position of specular beamcorresponding to third facet shown as within broken circle inFig. 4(a and b), is difficult to determined as the specular beam isblocked by the leads of the sample support assembly. In this case theposition of the specular beam is determined by extrapolating themotion of the diffraction spots corresponding to the face withchanging the electron beam energy. The estimated tilt angle betweenthe facet and the (2 1 0) plane is 18 � 0.28. Based on the tilt angles andazimuthal angle between the two symmetric facets, the Miller indicesof these two facets are identified as (7 3 1) and (7 3 1) planes and thethird facet as (110) [14]. The observation of {7 3 1} and (1 1 0) facet isalso supported by early study by Tucker [15] on Rh(2 1 0) surface.However, Tucker observed diffraction spots corresponding to {7 3 1}facet on Rh(2 1 0) when annealing the surface in oxygen background1 � 10�7 Torr at 573 K and these diffraction spot are completelyreplaced by (1 1 0) plane when annealing in oxygen pressure1 � 10�6 Torr at the same temperature.

Fig. 5. LEED pattern from the faceted Rh(2 1 0) at electron energy 66 eV (a) oxygen-covered faceted surface; (b) clean faceted surface; (c) illustrate the cleaning reaction.

Govind et al. / Applied Surface Science 256 (2009) 371–375374

For overlayer-induced faceted surface the important issue is theproduction of clean faceted surface. The clean faceted surface is anexcellent system for studying the sensitivity of chemical reactionsto the surface morphology and provides a simple system on whichone can experimentally probe and subsequently theoreticallymodel the facet relaxation phenomena. The production of a cleannanoscale facets or freezing of the facets is a new aspect tooverlayer-induced faceting experiments of the faceted rhodiumsurface. From the literature the most straightforward method ofremoving oxygen from the rhodium surface is through thermaldesorption, however, to produce a clean faceted rhodium surface isdifficult because the faceted surface reverts to planar surface at850 K (in the absence of oxygen background). This temperature issignificantly lower than those required to completely remove theoxygen due to desorption (�1400 K) and making it impossible toremove the oxygen by desorption without destroying the facetedsurface. However, our investigations have revealed two ‘‘lowtemperature’’ procedures for circumventing this problem, bychemically removing the oxygen overlayer. The first such methodtakes advantage of catalytic CO oxidation and the second by H2

interaction with the overlayer oxygen. The catalytic CO oxidationcan successfully produce a clean faceted Rh surface; however the

Fig. 6. STM scan of the fully faceted Rh (2 1 0) surface (a) The STM image of the nanopyra

image of single pyramid of faceted Rh (2 1 0) surface (22 nm � 25 nm).

CO interaction on faceted Rh surface can lead to destruction offacets and hence required an extreme control on substratetemperature as well on CO coverage. In order to evade thepossibility of facet destruction hydrogen has been used to removesurface oxygen, which can be signified as the following reaction:

H2ðgÞ þOðaÞ ! H2OðgasÞ (1)

In this case, H2 is temporarily introduced into the chamber(usually backfilled to a pressure of 4 � 10�8 Torr for 400 s,resulting in a 16L exposure) while keeping the oxygen-coveredfaceted Rh(2 1 0) surface at room temperature. The H2 issubsequently pumped out, and the sample kept at 400 K in UHVfor additional 120 s before cooling to room temperature. Thistemperature is sufficient to ensure desorption of both hydrogenand water. This procedure results in a clean faceted surface, withthe added benefit that the sample temperature remains comfor-tably lower than 600 K at all times. The surface remains ‘‘frozen’’ inits faceted state at temperatures up to 600 K. A schematic diagramof the reaction mechanism and LEED images before and after cleanfacet formation is shown in Fig. 5.

The surface morphology of faceted Rh(2 1 0) surface isconfirmed by STM. Fig. 6(a) shows a typical STM image of faceted

midal faceted surface scan area 150nm � 150 nm (V = 1.2 mV, I = 1 nA) (b) 3D-STM

Govind et al. / Applied Surface Science 256 (2009) 371–375 375

Rh(2 1 0) surface where the surface was prepared by pre dosing of10L oxygen at 300 K followed by annealing at 850 K for 2 minfollowed by cooling to 300 K in the presence of oxygen(4 � 10�8 Torr). The faceted Rh surface is fully covered with wellformed nano pyramids (three-sided facets) with similar shape,which implies that they expose faces of identical crystalorientation. A three dimensional image of single pyramid isshown in Fig. 6(b), where the crystal orientation all the facets onRh(2 1 0) surface are displayed. The orientation of the facetsidentified by LEED is confirmed by measuring azimuthal anglesbetween the edge lines of individual pyramids. The averagepyramid size for these nanopyramids formed under describedcondition was found to be �18 nm. However, the size of thenanopyramids can vary from 12 nm to 21 nm, depending upon theannealing temperature.

4. Conclusion

In the present work, we studied the adsorbate inducedmorphological instability of the Rh(2 1 0) substrate. For cleanmetal surfaces the anisotropy in surface free energy is generallyto small for an atomically rough surface to spontaneously form afaceted surface. The presence of adsorbate can reduce the surfacefree energy, which results in an increase in anisotropy of surfacefree energy and guide to facet formation. We observed oxygeninduced pyramidal faceting of Rh(2 1 0) surface with {7 3 1} and(1 1 0) facets when annealing Rh(2 1 0) in oxygen background atand above 550 K. A clean faceted surface is prepared using lowtemperature surface cleaning using H2 reaction while preservingfreezing the pyramidal facet structure. The resulting cleanfaceted surface remains stable for T � 600 K and for tempera-tures above this value, the surface irreversibly relaxes to theplanar state. STM studies confirms the formation of nanopyr-amids on Rh(2 1 0) surface with average pyramid size ranges

from 12 to 21 nm, which can be controlled by changing theannealing temperature. Theoretical studies are necessary to givedetailed energetic descriptions of the bonding characteristicsbetween oxygen and Rh on different facets and the change in thesurface energy anisotropy. The faceted Rh surfaces providepossible model systems to study structure sensitivity in Rh basedcatalytic reactions as well as potential nanotemplate to grownanoclusters.

Acknowledgement

Dr. Govind thanks Department of Science & Technology, Govt.India, New Delhi, India for BOYSCAST fellowship. This work hasbeen supported by the U.S. Department of Energy (DOE), Office ofBasic Energy Sciences (Grant DE-FG-02-93ER14331).

References

[1] G. Ertl, H. Knozinger, J. Weitkamp, Handbook of Heterogeneous Catalysis, Wiley,New York, 1997.

[2] Q. Chen, N.V. Richardson, Surf. Prog. Surf. Sci. 73 (2003) 59.[3] T.E. Madey, J. Guan, C.H. Nien, C.Z. Dong, H.S. Tao, R.A. Campbell, Surf. Rev. Lett. 3

(1996) 1315.[4] T.E. Madey, C.H. Nien, K. Pelhos, J.J. Kolodziej, I.M. Abdelrehim, H.S. Tao, Surf. Sci.

438 (1999) 191.[5] R.E. Kirby, C.S. McKee, M.W. Roberts, Surf. Sci. 55 (1976) 725.[6] R.E. Kirby, C.S. McKee, L.V. Renny, Surf. Sci. 97 (1980) 457.[7] A.T.S. Wee, J.S. Foord, R.G. Egdell, J.B. Pethica, Phys. Rev. B 58 (1998) R7548.[8] I. Ermanoski, K. Pelhos, W. Chen, J.S. Quinton, T.E. Madey, Surf. Sci. 549 (2004) 1.[9] M. Sander, R. Imbihl, R. Schuster, J.V. Barth, G. Ertl, Surf. Sci. 271 (1992) 159.

[10] H. Wang, W. Chen, T.E. Madey, Phys. Rev. B 74 (2006) 205426.[11] H. Wang, A.S.Y. Chan, W. Chen, P. Kaghazchi, T. Jacob, T.E. Madey, ACS Nano 1 (5)

(2007) 449.[12] W. Chen, I. Ermanoski, Q. Wu, T.E. Madey, H. Hwu, H.J.G. Chen, J. Phys. Chem. B 107

(2003) 5231.[13] W. Chen, I. Ermanoski, T.E. Madey, J. Am. Chem. Soc. 127 (2005) 5014.[14] Govind, Wenhua Chen, Hao Wang, T.E. Madey (communicated).[15] C.W. Tucker Jr., Acta Metall. 15 (1967) 1465.