Probing the Interface in Vapor-Deposited Bimetallic Pd-Au andPt-Au Films by CO Adsorption from the Liquid PhaseDavide Ferri, Bahar Behzadi,, Peter Kappenberger, Roland Hauert,

Karl-Heinz Ernst, and Alfons Baiker*,

Department of Chemistry and Applied Biosciences, ETH Zurich, HCI, CH - 8093 Zurich, Switzerland,and Nanoscale Materials Science, Empa, Swiss Federal Laboratories for Materials Testing and Research,

CH - 8600, Dubendorf, SwitzerlandReceiVed August 8, 2006. In Final Form: October 27, 2006

Bimetallic Pd-Au and Pt-Au and monometallic Pd, Pt, and Au films were prepared by physical vapor deposition.The resulting surfaces were characterized by means of XPS, AFM, and CO adsorption from the liquid phase (CH2Cl2)monitored by ATR-IR spectroscopy. CO adsorption combined with ATR-IR proved to be a very sensitive methodfor probing the degree of interdiffusion occurring at the interfaces whose properties were altered by variation of thePd and Pt film thickness from 0.2 to 2 nm. Because no CO adsorption was observed on Au, the evaporation of Pt-groupmetals on Au allowed us to study the effect of dilution on the adsorption properties of the surfaces. At equivalentPd film thickness, the evaporation of Au reduced the amount of adsorbed CO and caused the formation of 2-foldbridging CO, which was almost absent in monometallic surfaces. Additionally, the average particle size on Pd-Ausurfaces was smaller than that on monometallic Pd surfaces. The results indicate that a Pd/Au diffuse interface isformed that affects the Pd particle size even more drastically than the simple decrease in Pd film thickness in monometallicsurfaces. Pt-Au surfaces were less sensitive to CO adsorption, indicating that the two metals do not mix to a significantextent. The difference in the interfacial behavior of Pd and Pt in the bimetallic gold films is traced to the largely differentPd-Au and Pt-Au miscibility gaps.

IntroductionThe chemistry and structure of bi- or multimetallic surfaces

have gained considerable attention because of their importancein several technical applications, including catalysis, microelec-tronics, and sensors. In catalysis, the major significance of theiruse is that binary systems often exhibit deviations in terms ofcatalytic activity and selectivity compared to those of theindividual components1 and are thus suitable to optimize theperformance of catalytic materials. Gold and its combinationwith Pd and Pt are used in a range of liquid-phase oxidation2-5and hydrogenation reactions.6-10

Quasi-planar thin films of two metals or more generally modelcatalysts are valuable systems for gaining fundamental insightinto elementary surface processes occurring at real catalyticsurfaces on a molecular level.11 Physically deposited Pd-Aubimetallic films are often used for electrocatalytic purposes12,13and are prepared under similar deposition conditions as for thethin films prepared by us to study the adsorption of large chiral

molecules on noble metals.14,15 An important feature of thisdeposition method seems to be that alloys are readily obtainedwithout annealing the pristine films. Annealing is typicallyreported in the case of UHV studies and is efficient for studyingand adjusting the surface composition of alloys. Depending onthe annealing temperature, different extents of surface segregationare obtained.16 Bimetallic films deposited by consecutive electronbeam evaporation of Pd on Au show epitaxial growth attemperature as low as 150 K with Au interdiffusion increasingwith temperature.12 The formation of a bimetallic film (inter-diffusion) or of two distinct films (layer-by-layer epitaxial growth)represents the two main feasible scenarios (Figure 1) that shouldbe expected from the preparation of thin films obtained bysequential evaporation of Au and Pt-group metals without furthertreatment. Figure 1a shows the model for a possible bimetallicfilm in which the Pt-group metal diffuses through the Au filmupon deposition, thereby forming a corolla of mixed composi-tion around the ideal metal particle that denotes the contactbetween the two bulk metal films. It is likely that the occurrenceof one or the other scenario depends on the physical miscibilitybetween the two metals at the temperature at which depositionis performed. Because palladium and platinum exhibit vastlydifferent miscibility phase diagrams with gold,17 it is interestingto explore the surfaces of Pd-Au and Pt-Au bimetallic filmsprepared by physical vapor deposition without controlling thepossible alloy formation by temperature treatment. The films arepotential candidates for studies involving the adsorption of morecomplex organic molecules such as chiral modifiers on metal

* Corresponding author. E-mail: baiker@chem.ethz.ch. ETH Zurich. Swiss Federal Laboratories for Materials Testing and Research.(1) Sinfelt, J. H. Acc. Chem. Res. 1977, 10, 15.(2) Bianchi, C.; Porta, F.; Prati, L.; Rossi, M. Top. Catal. 2000, 13, 231.(3) Biella, S.; Castiglioni, G. L.; Fumagalli, C.; Prati, L.; Rossi, M. Catal.

Today 2002, 72, 43.(4) Corma, A.; Domine, M. E. Chem. Commun. 2005, 4042.(5) Hermans, S.; Devillers, M. Catal. Lett. 2005, 99, 55.(6) Bond, G. C.; Rawle, A. F. J. Mol. Catal. A: Chem. 1996, 109, 261.(7) Sarkany, A.; Horvath, A.; Beck, A. Appl. Catal., A 2002, 229, 117.(8) Venezia, A. M.; LaParola, V.; Pawelec, B.; Fierro, J. L. G. Appl. Catal.,

A 2004, 264, 43.(9) Claus, P. Appl. Catal., A 2005, 291, 222.(10) Pawelec, B.; Venezia, A. M.; LaParola, V.; Thomas, S.; Fierro, J. L. G.

Appl. Catal., A 2005, 283, 165.(11) Gunter, P. L. J.; Niemantsverdriet, J. W.; Ribeiro, F. H.; Somorjai, G. A.

Catal. ReV.sSci. Eng. 1997, 39, 77.(12) Koel, B. E.; Sellidj, A.; Paffett, M. T. Phys. ReV. B 1992, 46, 7846.(13) Schmidt, T. J.; Stamenkovic, V.; Markovic, N. M.; Ross, P. N. Electrochim.

Acta 2003, 48, 3823.

(14) Ferri, D.; Burgi, T.; Baiker, A. J. Phys. Chem. B 2001, 105, 3187.(15) Ferri, D.; Burgi, T. J. Am. Chem. Soc. 2001, 123, 12074.(16) Yi, C. W.; Luo, K.; Wei, T.; Goodman, D. W. J. Phys. Chem. B 2005,

109, 18535.(17) Okamoto, H.; Massalski, T. B. Phase Diagrams of Binary Gold Alloys;

ASM International: Metals Parks, OH, 1987.

1203Langmuir 2007, 23, 1203-1208

10.1021/la0623477 CCC: $37.00 2007 American Chemical SocietyPublished on Web 12/09/2006

surfaces, whose adsorption will be strongly biased by thecomposition of the surface of the as-deposited bimetallic films.18

The characterization of bimetallic surfaces using traditionalmethods is complicated by the nature of the metal-metalinteraction and the microscopic range of this interaction. Amongthe numerous classic methods available for the characterizationof bimetallic surfaces, the adsorption of carbon monoxide providesinformation on the surface structure including both morphologicaland electronic aspects.19,20 CO shows a characteristic dependenceof adsorption mode upon coordination, which markedly changesupon variations in the environment of the metallic atoms (i.e.,upon changing from metal to metal and from a metal surface toa composite metal surface). The strong interaction of CO withcatalytic metals can present a drawback for the characterizationof materials whose surfaces are reconstructed upon adsorption.21-23For bimetallic systems, reconstruction involves the preferentialenrichment of the surface with one component upon interactionwith the adsorbate.

The investigation of metal surfaces in contact with a solutionrequires the use of spectroscopic techniques enabling the efficientsubtraction of the large signals of the bulk solvent from the weakfeatures originating from species populating the surface. Inattenuated total reflection infrared (ATR-IR) spectroscopy,24,25the infrared radiation propagates through an infrared transparentmaterial serving as a waveguide (internal reflection element,IRE) into the absorbing medium in contact with it and reflectsat the interface, thus generating a short-range electric fielddecaying from the interface to the bulk medium. The detectionof species residing at the interface and in close proximity to theIRE/solution contact is enhanced because of this electric field.Owing to the short path length (up to few micrometers) of thetechnique, the contribution from the bulk phase in contact withthe IRE is minimized, and direct information about the solid-liquid interface is obtained. The technique becomes highlysurface-sensitive when the right combination of the IRE materialand a pseudoplanar metal surface is chosen. In this respect, ATR-IR spectroscopy has been applied to the characterization of mono-

and bimetallic surfaces in contact with solutions14,26-29 with someemphasis on the electrolyte/metal interfaces,30-37 which takeadvantage of remarkable surface enhancement induced by goldin the so-called Kretschmann configuration.38,39

The aim of this work is to elucidate the surface properties ofbimetallic films prepared using electron beam vapor depositionof Au, Pt, and Pd. The use of temperature treatments is ratherlimited and therefore not applied because serious damage canoccur to the Ge IRE used as a substrate for the films. Surfacecomposition and morphology have been investigated by X-rayphotoelectron spectroscopy (XPS) and atomic force microscopy(AFM), respectively. The adsorption properties of the surfacesare probed by ATR-IR spectroscopy using CO as the adsorbatein the presence of dichloromethane solvent. CO adsorption shouldprovide a qualitative measure at a molecular level of themorphological changes induced by the potential formation of atrue bimetallic interface in comparison with monometallicsurfaces.

Experimental SectionMaterials. Dichloromethane solvent (Baker) was stored over 5

molecular sieves. N2 (99.995 vol %) and H2 (99.999 vol %) gaseswere purchased from Pangas, and CO (0.5 vol % in argon) waspurchased from Sauerstoffwerk Lenzburg. Pt (99.99%), Pd (99.959%),and Au (99.99%) wires and Al2O3 (99.3%) tablets used as a targetfor electron beam vapor deposition (EB-VD) were supplied byUmicore.

Thin Film Preparation and Characterization. Thin model filmswere prepared by EB-VD onto the trapezoidal Ge internal reflectionelement (IRE, 52 20 2 mm3, 45 angle of incidence, Komlas)used for ATR-IR spectroscopy in a Balzers BAE-370 vacuum systemequipped with a turret allowing the sequential deposition of fourdifferent materials without breaking the vacuum. The methodologyhas been described in detail elsewhere.14 Briefly, the materials wereevaporated from a graphite crucible using an electron beam (8 kV,to 0.15 mA) at a base pressure of about 1.0 10-5 mbar and atevaporation rates of 0.5 and 1.0 /s for the metals and for Al2O3,respectively. The IRE was polished with 0.25 m diamond pasteafter use and thoroughly cleaned with ethanol. Typically, 100 nmAl2O3 was deposited on Ge, followed by the metal(s). The filmthickness was measured with a quartz crystal microbalance. Twoseries of samples were prepared for both Pt and Pd in which the filmthicknesses were 0.2, 0.5, 1.0, and 2.0 nm. The first series isrepresented by the monometallic samples, named M02, M05, M10,and M20, where M is the metal symbol (Pt or Pd). In the secondseries, Pt or Pd was deposited on 1 nm Au in addition to Al2O3(MAu02, MAu05, MAu10, and MAu20).

The dependence of the composition of the thin films on thethickness was investigated using XPS. Physical Electronics (PHI)Quantum 2000 equipment was used with X-rays generated by an Alsource operating at 23.5 W. The spectrometer energy scale was

(18) Behzadi, B.; Vargas, A.; Ferri, D.; Ernst, K. H.; Baiker, A. J. Phys. Chem.B 2006, 110, 17082.

(19) Ponec, V. AdV. Catal. 1983, 32, 149.(20) Hollins, P. Surf. Sci. Rep. 1992, 16, 51.(21) Somorjai, G. A. Catal. Lett. 1992, 12, 17.(22) Raval, R.; Haq, S.; Harrison, M. A.; Blyholder, G.; King, D. A. Chem.

Phys. Lett. 1990, 167, 391.(23) Zou, S.; Gomez, R.; Weaver, M. J. Surf. Sci. 1998, 399, 270.(24) Harrick, N. J. Internal Reflection Spectroscopy; Interscience Publishers:

New York, 1967.(25) Burgi, T.; Baiker, A. AdV. Catal. 2006, 50, 227.

(26) Zippel, E.; Breiter, M. W.; Kellner, R. J. Chem. Soc., Faraday Trans.1991, 87, 637.

(27) Ferri, D.; Burgi, T.; Baiker, A. J. Catal. 2002, 210, 160.(28) Burgener, M.; Ferri, D.; Grunwaldt, J. D.; Mallat, T.; Baiker, A. J. Phys.

Chem. B 2005, 109, 16794.(29) Ebbesen, S. D.; Mojet, B. L.; Lefferts, L. Langmuir 2006, 22, 1079.(30) Watanabe, M.; Zhu, Y.; Uchida, H. J. Phys. Chem. B 2000, 104, 1762.(31) Lu, G. Q.; Sun, S. G.; Cai, L. R.; Chen, S. P.; Tian, Z. W. Langmuir 2000,

16, 778.(32) Nakamura, R.; Sato, S. Langmuir 2002, 18, 4433.(33) Pronkin, S.; Wandlowski, T. Surf. Sci. 2004, 573, 109.(34) Yajima, T.; Uchida, H.; Watanabe, M. J. Phys. Chem. B 2004, 108, 2654.(35) Shiroishi, H.; Ayato, Y.; Kunimatsu, K.; Okada, T. J. Electroanal. Chem.

2005, 581, 132.(36) Yan, Y. G.; Li, Q. X.; Huo, S. J.; Ma, M.; Cai, W. B.; Osawa, M. J. Phys.

Chem. B 2005, 109, 790.(37) Chen, Y. X.; Ye, S.; Heinen, M.; Jusys, Z.; Osawa, M.; Behm, R. J.

J. Phys. Chem. B 2006, 110, 9534.(38) Hartstein, A.; Kirtley, J. R.; Tsang, J. C. Phys. ReV. Lett. 1980, 45, 201.(39) Osawa, M. Bull. Chem. Soc. Jpn. 1997, 70, 2861.

Figure 1. Schematic presentation of the formation of mono- andbimetallic films vapor deposited on Al2O3. (a) Diffusion of the noblemetal into the Au film and (b) layer-by-layer growth of the noblemetal film on Au.

1204 Langmuir, Vol. 23, No. 3, 2007 Ferri et al.

calibrated using the Au 4f7/2 line (84.0 eV). The beam diameter wastypically 20 mm for survey spectra and 10 mm for line scans. Surveyspectra were obtained over a binding-energy scale of 0-1100 eVusing an analyzer pass energy of 117.4 eV. High-resolution multiplexdata of the patterned samples were obtained at a pass energy of 58.7eV. The residual vacuum pressure was typically 2 10-9 mbar orlower during analysis. The binding energies of the photoelectronpeaks were referenced to the C 1s line at 284.6 eV. Data acquisitionand storage were accomplished using PHI-MultiPak software. Theatomic percentages of the elements were calculated using atomicsensitivity factors included with the instrument data system.

The morphology of mono- and bimetallic films was investigatedby atomic force microscopy (AFM). Measurements were performedon as-deposited samples at room temperature with a Dimension3100 from Digital Instruments using tapping mode operation. Astandard cantilever with a resonance frequency of about 284 kHzand a stiffness of 42 N/m was employed for all measurements.

ATR-IR Spectroscopy. In situ ATR-IR spectra of the solid-liquid interface were recorded by co-adding 200 scans at 4 cm-1resolution on an Equinox 55 spectrometer (Bruker Optics) equippedwith a liquid-nitrogen-cooled MCT detector. After mounting thecoated IRE, the homemade stainless steel flow-through cell wasplaced on a dedicated ATR accessory (Optispec), and the samplecompartment was closed to allow the removal of water vapor andCO2 by dry air overnight. The temperature was maintained at 10 Cthroughout the experiments for comparison with previous mono-metallic Pt14 and Pd27 films.

Measurements were carried out as follows. First, N2-saturatedCH2Cl2 provided from a gas bubble tank was placed into contact(1 mL/min flow rate) with the coated IRE to reach steady-stateconditions (2 h). Then the gas was changed from N2 to H2 in thesame glass bubble tank without interrupting the liquid flow, andH2-saturated CH2Cl2 was allowed to flow through the cell for 10 minto clean the surface. The last spectrum during cleaning served as thereference spectrum for CO adsorption (ca. 60 min) from CO-saturatedCH2Cl2 provided from a second glass bubble tank. Finally, H2-saturated CH2Cl2 was again allowed to flow through the cell followingdesorption. All spectra are presented in absorbance units and werecorrected for the contribution of water vapor where required.

Results and DiscussionX-ray photoelectron spectra of the mono- and bimetallic

samples indicated that all metals were in the reduced state (Au4f core level at 87 and 83.3 eV,40 Figure 2) but no significantshift was observed in the core-level binding energies of PdAuand PtAu samples, which would have indicated electronicmodification of either the Au or the Pt-group element.16Broadening on the high-binding-energy side suggested that afraction of the Pt-group metals were in the oxidized state,14,27likely forming a thin oxide film as confirmed by the slowerattenuation of the concentration of O compared to that of Al atincreasing film thickness. Contamination by adventitious carbonalso increased with film thickness.

XPS was crucial to the estimation of the thickness of the metalfilms evaporated on Au/Al2O3.41 The Au 4d5/2, Pt 4f5/2, and Pd3d core-level lines were used for this purpose. The Pt filmthickness was underestimated in most of the cases, in the worstcase by 16% (PtAu20), but excellent values were obtained forthe thinnest samples. The linear correlation (R ) 0.998) betweenthe expected (nominal) and the estimated film thickness confirmed

the reliability of the physical vapor deposition technique. On thecontrary, the thickness of Pd in PdAu surfaces was largelyoverestimated because of the overlap between the Au 4d and thePd 3d signals at all Pd contents.

Figure 2 shows atomic force microscopy images (500 500nm2) of selected monometallic and bimetallic thin films. The Aufilm is composed of a densely packed layer of particles exhibitinga variable width between about 9 and 30 nm. Comparison withthe morphology of the Al2O3 film14 allows the assignment ofthese entities to Au particles. Larger features (up to 65 nm) arealso observed, which might arise from impurities after exposureto air. The AFM images, which are similar to those of other thinmetal films (Cu,42 Pt,14,26 and Pd27), show the island characterof the electron-beam-deposited films The average particle sizeincreased when Pt or Pd was evaporated onto the Au film, whereasthe shape of the particles remained largely unaffected. For thePtAu05 sample, the smallest particles had an approximatediameter of 10-20 nm. The size of the individual particlesincreased to about 30 nm in the PtAu10 sample. However, thepresence of Au did not substantially change the particle size ofthe Pt-bimetallic films compared to that of the monometallicfilm as demonstrated for the Pt05 and PtAu05 samples. However,a dramatic difference is observed in the morphology of the Pd-containing samples. The smallest particles in PdAu10 are about

(40) Briggs, D.; Sheah, M. P. Practical Surface Analysis by Auger and X-rayPhotoelectron Spectroscopy; John Wiley & Sons: New York, 1983.

(41) Seah, M. P.; Spencer, S. J.; Bensebaa, F.; Vickridge, I.; Danzebrink, H.;Krumrey, M.; Gross, T.; Oesterle, W.; Wendler, E.; Rheinlander, B.; Azuma, Y.;Kojima, I.; Suzuki, N.; Suzuki, M.; Tanuma, S.; Moon, D. W.; Lee, H. J.; Cho,H. M.; Chen, H. Y.; Wee, A. T. S.; Osipowicz, T.; Pan, J. S.; Jordaan, W. A.;Hauert, R.; Klotz, U.; vanderMarel, C.; Verheijen, M.; Tamminga, Y.; Jeynes,C.; Bailey, P.; Biswas, S.; Falke, U.; Nguyen, N. V.; Chandler-Horowitz, D.;Ehrstein, J. R.; Muller, D.; Dura, J. A. Surf. Interface Anal. 2004, 36, 1269. (42) Ishida, K. P.; Griffiths, P. R. Anal. Chem. 1994, 66, 522.

Figure 2. AFM images (500 500 nm2) of selected Pd, PdAu, Pt,and PtAu samples together with the Au sample. The images arereferred to as deposited samples. The inset depicts the Au 4f core-level signals of the Au10 sample.

Probing the Interface in Vapor-Deposited Films Langmuir, Vol. 23, No. 3, 2007 1205

15-20 nm, whereas their size is approximately doubled in theabsence of Au (Figure 2).

The adsorption of CO on mono- and bimetallic films isdemonstrated in Figure 3 (Au and Pd) and Figure 4 (Pt) forexperiments carried out under identical conditions and spectraacquired after 60 min of contact with a CO-saturated solvent.CO adsorbed on Pd20 and Pt10 exhibited two pairs of signals:one above 2000 cm-1 (COL; Pt, 2047 cm-1; Pd, 2040 cm-1)assigned to on-top species and the second below 2000 cm-1(COB; Pt, 1817 cm-1; Pd, 1905 cm-1) indicating the populationof multicoordinated species.43 The COB species on the Pd filmexhibited a rather broad signal with at least two to threecomponents, indicating high surface heterogeneity and differentadsorption sites. The CO frequency above and below 1900 cm-1can be assigned to CO adsorbed predominantly on bridge (2-foldcoordination) and hollow sites (3-fold coordination), respectively.High intensity for the signal corresponding to 3-fold-bondedspecies warrants relatively large domains of reduced metal becausethe species occupies more surface atoms than 2-fold-coordinatedspecies. The assignment agrees with the relatively large size ofthe Pd particles determined by STM27 and AFM (Figure 2) andwith the low intensity of the signal of on-top species.

Figure 3a reveals that CO did not adsorb on Au. UHV studiesindicate that CO is completely desorbed at the temperature atwhich the present ATR-IR experiments have been performed,44although a signal at 2109 cm-1 (COL) was observed for a 20-nm-thick polycrystalline Au film at ambient temperature underspectroelectrochemical conditions.33 Owing to the relatively largeparticle size, it is likely that the polycrystalline Au film shown

in Figure 2 is not corrugated enough to allow CO adsorption.Despite the ability of CO to bind to gold in different geometries,45the particle size appears to be a crucial factor for CO adsorptionon Au,46 which occurs preferentially on low-coordinated Auatoms47 and hence on small clusters or on defects rather than onextended surfaces. This property makes Au nanoparticles excellentcatalytic materials for CO oxidation compared to bulk Au.48 Theabsence of adsorbed CO on the Au film has the benefit that itallows the investigation of the effect of dilution of Pd and Ptmetals in bimetallic surfaces using infrared spectroscopy andCO as a probe molecule without the interference from COon Au.

In the Pd and Pt monometallic samples, the decrease in filmthickness from 2 to 0.2 nm was accompanied by a decreased COadsorption, as expected. The frequency of the CO signals in thePd films was perturbed by the decrease in film thickness. Pd10and Pd05 show signals at 2040 and 1870 cm-1. A shoulder atca. 1800 cm-1 is evident in Pd10. A considerable blue shift isobserved in correspondence to Pd02, exhibiting signals at 2068and 1887 cm-1 for COL and COB, respectively. Similarly, a shiftof 35 cm-1 is observed in the COB signal when changing fromPd10 to Pd20 (1905 cm-1). Very weak signals were observedfor Pd05 and Pd02. The decrease in film thickness more likelymodifies the distribution of sites available for the adsorption ofCOB species and the size of the Pd domains.49 This is best shown

(43) Hoffmann, F. M. Surf. Sci. Rep. 1983, 3, 107.(44) Rainer, D. R.; Xu, C.; Holmblad, P. M.; Goodman, D. W. J. Vac. Sci.

Technol., A 1997, 153, 1653.

(45) Blizanac, B. B.; Arenz, M.; Ross, P. N.; Markovic, N. M. J. Am. Chem.Soc. 2004, 126, 10130.

(46) Lemire, C.; Meyer, R.; Shaikhutdinov, S. K.; Freund, H. J. Surf. Sci.2004, 552, 27.

(47) Boccuzzi, F.; Chiorino, A.; Tsubota, S.; Haruta, M. J. Phys. Chem. 1996,100, 3625.

(48) Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. J. Catal. 1989, 115,301.

(49) Wolter, K.; Seiferth, O.; Kuhlenbeck, H.; Baumer, M.; Freund, H. J. Surf.Sci. 1998, 399, 190.

Figure 3. ATR-IR spectra of CO adsorption from CH2Cl2 solventat 10 C on Pd (bold lines) and PdAu surfaces. Pd film thicknesses:(b) 0.2, (c) 0.5, (d) 1, and (e) 2 nm. Spectra a have been collectedfor CO adsorption on the monometallic Au surface (1 nm) underequivalent conditions. Spectra b and c have been magnified 3 and4 times, respectively, for ease of comparison. The inset shows amagnification of spectra b and c.

Figure 4. ATR-IR spectra of CO adsorption from CH2Cl2 solventat 10 C on Pt (bold lines) and PtAu surfaces. Pt film thicknesses:(a) 0.2, (b) 0.5, (c) 1, and (d) 2 nm. Spectra a and b have beenmagnified 5 and 3 times, respectively, for ease of comparison.

1206 Langmuir, Vol. 23, No. 3, 2007 Ferri et al.

in Figure 3c,d, where the component at ca. 1800 cm-1 completelydisappears from Pd10 to Pd05.

A slight blue shift ( ) 5 cm-1) accompanied by attenuationof the intensity was observed for the COL signal when the Pt filmthickness was decreased from 2 nm (Pt20, 2045 cm-1) to 1 nm(Pt10, 2049 cm-1) (Figure 4). Similar to the situation for Pd, thePt02 sample exhibits signals for COL and COB at higher frequencythan do the other Pt-containing surfaces (i.e., at 2057 and 1827cm-1, respectively). The overall intensity of the COL signal forPt20 is likely diminished due to the bipolar appearance of theenvelope. As a result, the maximum of the signal also shiftstoward lower frequency. This is likely also occurring for Pd20but to a lesser extent. The typical second derivative-like shapeassociated with the strongest signals of CO on both Pd and Ptis attributed to the perturbation of the optical properties of thefilm probed by the IR radiation propagating through the IREupon adsorption of a strong absorber.42,50 The perturbation hasa stronger effect on the appearance of the spectra associated withthicker films. Further indication of changes in the opticalproperties of the metal film was given by the increasingbackground absorption on all surfaces on contact with CO owingto the strong interaction between the CO adsorbate layer and themetal.33

Evaporation of 1 nm Au on the Al2O3 film prior to the Pt-group metal induced important changes in the intensity of theCO signals virtually only in the case of Pd. This reflects importantchanges in the adsorption properties of Pd compared to those ofmonometallic surfaces. CO adsorption occurred to a large extentonly for PdAu20 and PdAu10, whereas for films thinner than 1nm extremely weak signals were obtained. The difference withthe corresponding monometallic films and the clear attenuationof the signals with decreasing film thickness are obvious andprevent us from discussing possible enhancement effects inducedby Au. The adsorption sites available for CO also appeared tobe modified by the addition of Au. Comparison between Pd10and PdAu10 reveals that the component of the COB envelopelocated at ca. 1800 cm-1 is strongly attenuated or disappears inthe bimetallic surface. Moreover, the maxima of both COL andCOB signals are blue shifted to 2059 ( ) 19 cm-1) and 1908cm-1 ( ) 38 cm-1), respectively. The shift could be smallerbecause of the second derivative shape of the signals in Pd10,as discussed above, but the value is so large, at least for theCOB signals, that the contribution of this optical effect can beneglected. To support this argument, it should be noted that theCOB signal is found at the same position in Pd10 and Pd05despite the band shape observed in Pd10. Figure 3 also clearlydemonstrates that signals due to adsorbed CO are extremelyweak in the PdAu05 and PdAu02 samples, whereas CO is clearlyvisible in the corresponding monometallic surfaces.

The increase in frequency of the signal of COB species observedfor the bimetallic surfaces indicates a change in the structure ofthe Pd domains induced by the presence of Au, most probablyaccompanied by adsorption on 2-fold bridging sites. At very lowPd content (PdAu02), CO adsorption is inhibited to a largerextent, and the ratio between on-top and bridging species is closeto 1, indicating the possible diluting effect of Au and the absenceof large Pd domains.

Only minor changes appear in corresponding PtAu surfaces(Figure 4). Generally, the frequency of the COL signal remainsunchanged with decreasing Pt thickness. However, a blue shiftof ca. 13 cm-1 is observed for the COB signal, which changesfrom 1812 to about 1825 cm-1. Additionally, the adsorption

ability of the bimetallic Pt samples is comparable to that of themonometallic samples in contrast to that of Pd-containing films.

We note that there are two important differences between thePdAu and PtAu samples. In contrast to Pd-containing surfaces,CO adsorption occurs noticeably on all PtAu samples and is stillsignificant on PtAu02. Moreover, the time at which signalsaturation is achieved is identical for Pt mono- and bimetallicsurfaces (not shown) whereas it is clearly different in Pd-containing surfaces, the maximum intensity of the CO signalbeing achieved at a shorter time on stream on bimetallic samples(Figure 5). These observations confirm that the extent ofinteraction between Au and the two Pt-group metals, as probedby CO adsorption, is substantially different.

The combination of information obtained from the AFM images(lower average particle size in bimetallic surfaces compared tothat of Pd films deposited on Al2O3 at identical nominal thickness)with that obtained from the ATR-IR spectra of CO adsorbed onPd indicates that the growth of the Pd film on Au does not followa layer-by-layer mechanism (i.e., the Pd-Au bimetallic surfaceis not simply composed of two distinct Pd and Au films). Thedecrease in the intensity of the CO signals and the disappearanceof some signals with decreasing Pd film thickness on themonometallic surfaces is interpreted as being the result of aparticle size effect, decreasing the amount of available free metalfor adsorption and affecting the sites to which CO can adsorb(Figure 1). Figure 3 clearly shows that the decrease in the intensityof the CO signals is much more significant for the bimetallicsurfaces at equivalent Pd film thickness and that Au inducesstrong effects in adsorbed CO at low Pd film thickness (0.2 and0.5 nm).

The results shown in Figures 3 and 5 indicate the enrichmentof the bimetallic surface with Au and the generation of a diffusePd/Au interface, which can be represented as in the cartoon inFigure 1a. The diffuse interface can be defined as a bimetallicslice of undefined extension and composition located at theboundary between the two monometallic phases, Au and the Pddomains (likely large particles for thick films) exposed to theflowing solution. The size of the monometallic Pd domains inthe bimetallic PdAu films decreases with Pd content. The sizeand composition of the slice are probably also film thickness-dependent, and it is possible that small Pd domains decorate theslice, although no indication can be found in the ATR-IR spectraof CO adsorbed on such features. Although CO-induced surfacereconstruction with consequent surface enrichment with one ofthe metals composing the bimetallic material is common, theintensity decrease with decreasing Pd content observed in Figure(50) Burgi, T. Phys. Chem. Chem. Phys. 2001, 3, 2124.

Figure 5. Kinetic of adsorption of CO on Pd and PdAu surfaces.Pd film thicknesses: () 0.2, () 0.5, (0) 1, and (9) 2 nm.

Probing the Interface in Vapor-Deposited Films Langmuir, Vol. 23, No. 3, 2007 1207

3 is more consistent with CO adsorption on Pd domains notinfluenced by surrounding Au atoms. The ATR-IR spectra ratherindicate that the diffuse interface produces a sort of geometriceffect that influences the adsorption properties of the domainsshowing only Pd character. Electronic effects51 often ac-companying the formation of true bimetallic interfaces are difficultto disentangle on the basis of these data from the geometriceffect accounting for our interpretation. Although rather sig-nificant, the blue shifts observed in the frequency of the COsignals on both the PdAu and the PtAu surfaces (Figures 3 and4) cannot be attributed to electronic effects because similar shiftshave been also found for the monometallic surfaces.

The geometric effect created by the formation of the diffuseinterface is best seen in the formation of prevalently 2-fold-adsorbed CO, meaning that more defects are present because ofthe reduced size of the metal domains, and in the faster decreasein the signal of COB compared to that of COL. The reduced sizeof the Pd domains is also reflected in the shorter time at whichthe maximum in the COL signal intensity is reached in bimetallicsurfaces compared to that in monometallic surfaces. Hence, atequivalent nominal Pd thickness, the Pd domains in the PdAusamples are smaller (Figure 1a) than those that would be obtainedhypothetically following layer-by-layer growth (Figure 1b) (nodiffusion of Pd into Au) and those obtained in the monometallicsamples.

AFM and ATR-IR spectroscopy both indicate that the diffusionof Pt into Au is less probable than in the case of Pd. The onlyeffect observed on Pt and PtAu surfaces is that the decrease inthe Pt film thickness results in smaller amounts of adsorbed CO.With the exception of the PtAu10 sample, monometallic andbimetallic surfaces exhibit approximately the same distributionof sites for CO adsorption. PtAu02 and Pt02 are significantexamples. Hence, the evaporation of Pt on Au generates twonearly independent metallic films.

In principle, the differences observed for Pd and Pt in the easeof formation of the diffuse interface can be associated with thecombination of two factors. First, the present surfaces have beencharacterized as prepared (i.e., without thermal treatment

following preparation, thus under conditions limiting interdif-fusion). Second, the different electronic structures of Pd and Pt,reflected by the often observed different catalytic properties ofthe two metals for identical reactions, certainly play a role. Pdand Au are miscible in almost all ratios,16,52-54 whereas thesolubility of Pt and Au in solid solutions is limited by a largemiscibility gap between 15 and 100 atom % Pt.17,55 Theimmiscibility between Pt and Au together with the fact thatdiffusion is not favored by the deposition conditions explains theabsence of differences in the signals of CO on Pt and PtAusamples and the large discrepancy with the results obtainedwith Pd.

ConclusionsPd-Au and Pt-Au bimetallic surfaces have been prepared by

electron beam deposition, and their structure has been character-ized using XPS, AFM, and CO adsorption combined with ATR-IR spectroscopy. The changes observed in the infrared frequencyand in the shape of the CO signals upon adsorption from theliquid phase indicated that morphological changes occur in thePd films when decreasing the film thickness from 2 to 0.2 nmand when introducing a 1 nm Au film. In comparison,corresponding Pt-Au surfaces were less sensitive toward COadsorption, which can be attributed to the large difference inmiscibility among Pd, Pt, and Au. A diffuse Pd/Au interface isformed when Pd is deposited on Au, which results in theinterdiffusion of Pd into Au and the enrichment of the surfacewith Au for very thin Pd films. The overall effect of the formationof the diffuse interface is a decrease in the metal particle sizeinduced by Au addition, which cannot be obtained by simplydecreasing the film thickness of vapor-deposited Pd in themonometallic surfaces.

This study shows that CO adsorption combined with ATR-IRspectroscopy provides a very sensitive and useful probe forinvestigating the mixing of metals at metal/liquid interfaces.By applying other suitable probe molecules, the method couldbe extended to other bimetallic systems that are relevant totechnical devices surrounded by a liquid phase.

Acknowledgment. We kindly acknowledge the financialsupport from the Swiss National Science Foundation and theFoundation Claude and Giuliana.LA0623477

(51) Burch, R. Acc. Chem. Res. 1982, 15, 24.(52) Jablonski, A.; Overbury, S. H.; Somorjai, G. A. Surf. Sci. 1977, 65, 578.(53) Shih, H. D.; Bauer, E.; Poppa, H. Thin Solid Films 1982, 88, L21.(54) Sellidj, A.; Koel, B. E. Phys. ReV. B 1994, 49, 8367.(55) Bouwman, R.; Sachtler, W. M. H. J. Catal. 1970, 19, 127.

1208 Langmuir, Vol. 23, No. 3, 2007 Ferri et al.