CO Chemisorption on Monodispersed Platinum Clusters on Si02

Mar 13, 1995 - Surface Science Center, Department of Chemistly, University of Pittsburgh, ... CO desorption from Pt(l1 l), Pt(l12), and bulk Pt films ...
1 downloads 0 Views 610KB Size
8730

J. Phys. Chem. 1995, 99, 8730-8735

CO Chemisorption on Monodispersed Platinum Clusters on Si02: Detection of CO Chemisorption on Single Platinum Atoms U. Heiz,"' R. Sherwood? D. M. Cox,' A. Kaldor,' and J. T. Yates, Jr.*rt Surface Science Center, Department of Chemistly, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, and Exxon Research and Engineering Company, Route 22E, Annandale, New Jersey 08801 Received: September 9, 1994; In Final Form: March 13, 1995" Monodispersed Ptl, Pt2, and Pt3 clusters have been deposited on an atomically clean Si02 film in ultrahigh vacuum to surface coverages of about 0.01 monolayer. CO chemisorptioddesorption experiments at 300 K exhibit a distinct 340 K CO desorption state with an activation energy of -16 kcaymol. Comparisons with CO desorption from Pt(l1 l), Pt(l12), and bulk Pt films on Si02 are made, showing that the 340 K CO desorption process is a unique feature of highly dispersed Pt atoms on SiO2. CO desorption from other Pt sites is also detected exhibiting higher desorption activation energies, but the nature of these sites cannot be ascertained in these experirnents with certainty.

I. Introduction It is now well established that small unsupported gas phase transition metal clusters exhibit pronounced variations in their physical, electronic, magnetic, and chemical properties as the number of metal atoms in the cluster is varied.'-4 These results suggest that supported clusters, i.e. clusters deposited onto surfaces, may also exhibit novel behavior, which depends strongly on the cluster size. Indeed, size dependent variations in electronic structure of deposited transition metal clusters have been r e p ~ r t e d . ~ Such - ~ considerations lead to speculations that one may be able to build "designer" catalysts by judiciously choosing the cluster size and substrate or support material. A key to finding the right cluster size and metal for a specific catalytic reaction is to understand the electrophilic character of the deposited cluster, i.e. the energy and symmetry of the HOMO and/or LUMO in a one electron picture. By varying cluster size and the metal itself, one can change the electronic properties in order to adjust the cluster chemistry. Gas phase studies have shown that the reactivity of a cluster may depend strongly on the oxidation state of the metal In a similar manner by changing the acidity of the substrate, one can alter the charge of supported clusters somewhat to optimize their reactivity.I2 Thus, experiments designed to deposit metal clusters of known selected monodispersed size onto atomically clean oxide supports of various types prepared under ultrahigh vacuum conditions are important. Such experiments offer the possibility of studying, in a carefully controlled environment, the influence of metal particle size and oxide support character on the properties of the metal clusters. The experiments reported here are a first attempt to deposit size-selected metal particles on a carefully prepared, clean Si02 surface and to study their chemisorption properties. The key to performing such studies is the provision of sufficient cluster deposition intensity (combined with maintenance of ultrahigh vacuum to maintain clean surface conditions) and to use highly sophisticated equipment to characterize the chemisorption and binding of a test molecule, which in this case is carbon monoxide. 11. Experimental Procedures The thermal desorption experiments on size-selected supported Pt clusters were performed in an ultrahigh vacuum system

' University @

of Pittsburgh. Exxon Research and Engineering Co. Abstract published in Advance ACS Absrracts, May 1, 1995.

0022-365419512099-8730$09.00/0

consisting of two separately pumped units: the sputter cluster source with the mass-separation unit (base pressure = 2 x Torr) and the analyzing chamber (base pressure = 2 x Torr) equipped for Auger electron spectroscopy (AES), high sensitivity temperature programmed desorption (TPD), and ion bombardment. The analyzing chamber also contains a conventional Pt evaporation source with a moveable shutter. The two units are connected with a gate valve. The sputter cluster source and mass-separation unit have been described previously5 and will be discussed only briefly. A cold reflex discharge ion source (CORDIS) provided the Xe+ beam at a typical current density of 5 mA/cm* at 25-keV kinetic energy. The CORDIS was mounted perpendicular to the cluster beam direction so that the primary Xe+ beam bombarded the Pt target at an incidence angle of 45". Ion extraction lenses, placed close to the target, focused the ionized particles into an energy analyzer (Bessel Box: Extranuclear 616-l), operated in the low resolution broad pass mode. The cluster ions were sizeselected with a quadrupole mass spectrometer (Extranuclear 7-162-8) and then transported to an atomically clean Si02 surface on a Si(100) crystal in the main chamber through an RF mode only quadrupole and two Einzel lenses. The beam current and beam profile were measured during deposition by a moveable detector consisting of four copper segments used as Faraday plates. For the CO desorption measurements from the Pt, particles on the Si02 surface, a differentially pumped, apertured, and multiplexed line-of-sight quadrupole mass spectrometer (UTI lOOC equipped with a CdBe multiplier) was used. A skimmer with a diameter of 5 mm was used as the aperture and located close to the surface for desorption experiments. The aperture was electrically isolated and kept at a voltage of -90 V during CO desorption experiments in order to avoid electron stimulated desorption and charging effects on the Si02 by electrons emitted from the filament of the mass ~pectr0meter.l~ For each experiment a Si(100) crystal (13 mm x 13 mm x 1.5 mm, 10 Gem, p-type, B-doped) was sputtered with Ar+ (1.5 kV) for 45 min with the ion gun. The crystal was annealed to 1200 K and then cooled slowly to 300 K. It was found that this procedure produces a clean, well-ordered, (2 x 1) surface.I4 The Si02 film is formed at 300 K by bombarding the crystal with Ar+ (0.8 kV) in an oxygen background of 2 x Torr for 20 min. The clean Si02 film was annealed to 900 K. From the ratio of the 0 (KLL) and Si (LVV) Auger peak-to-peak signals, a film thickness of about 40 A was e~timated.'~ High

0 1995 American Chemical Society

J. Phys. Chem., Vol. 99, No. 21, 1995 8731

CO Chemisorption on Monodispersed Pt Clusters on Si02

1

I

1

n

P2

3

f P

u

w

0

78 eV

505 eV

50

100

150

200

250

300

350

450

400

i

500

550

KINETIC ENERGY [eV] Figure 1. Auger spectrum of a clean SiO~lSi(100)film. The estimated Si02 thickness is about 40

resolution electron energy loss spectroscopy (HREELS) studies have shown that such films grown on a Si(100) single crystal reveal almost identical phonon frequencies as amorphous Si02 materials.16 The Auger spectrum of this substrate is shown in Figure 1, and it may be seen that it is free of carbon at the sensitivity limit of Auger spectroscopy (-1 at. %). The cleanliness of the Si02 film was also checked by measuring CO desorption after film formation, which reveals no CO desorption feature in the measured temperature range (300700 K). The oxidized Si crystal could be resistively heated by a homebuilt linear-programtemperature controller at a rate of 1.8 Ws.” The surface composition of the Si( 100) crystal, the Si02 film, and the Pt deposits were examined with a digital Auger electron spectrometer (Perkin Elmer 32- 150). Pt coverages for atoms, dimers, and trimers were typically 1% of a monolayer, as measured from the cluster beam current and beam profile. These low coverages were used to minimize interactions between Pt, clusters. Typical Pt,’ beam currents were in the range of a few nanoamperes. At a Pt target bias of +10 V, the actual beam energy was found to be 10 f 5 eV. The background pressure during deposition was 1 x Torr; the increase in the background pressure in the analyzing chamber by a factor of 5 is mainly due to the Xe sputter gas. A collimated and calibrated effusive beam doser was employed to achieve accurate CO fluences on the PtJSiO2 surfaces.18J9 TPD measurements were done with CO exposures of 2 x l O I 3 to 5.6 x l O I 5 molecules/cm2. We used isotopically labeled I3CO to improve the signal-to-noise ratio, avoiding the 28 amu background signal. 111. Results

A. CO AdsorptionDesorptionon Platinum clusters. This paper describes studies of the CO desorption from four different samples, a SiOz/Si substrate with 0.009 monolayer of Pt atoms deposited, a SiOZ/Si substrate with 0.009 monolayer of Pt dimers deposited, a SiOZ/Si substrate with 0.006 monolayer of Pt trimers deposited, and a clean SiOdSi substrate. Figure 2 shows results

1

I

1

400

1

A.

1

l

500

1

1

300

TEMPERATURE

1

400

1

1

500

1

600

(K)

Figure 2. TPD spectra of CO desorption for four different Pt cluster samples: (A) SiOZlSi substrate with 0.009 monolayer of Ptl deposited: (B) SiOz/Si substrate with 0.009 monolayer of Pt2 deposited; (C) SiOzl Si substrate with 0.006 monolayer of Pt3 deposited; (D) clean SiOdSi substrate.

from CO desorption from each of the four samples after an exposure of 5.6 x lOI5 CO molecules/cm2. No evidence for CO desorption is observed for the SiO2/Si substrate void of any platinum. Since subsequent Auger analysis showed no evidence of carbon buildup attributable to CO decomposition on the Si02, we conclude, as have many others, that CO does not adsorb onto Si02 at room temperature.20.2’ For the three samples containing Ptl, P t 2 , and Pt3 cluster deposits, CO is readily chemisorbed at 300 K. CO desorption spectra from the Pt clusters are shown in Figure 2. For Ptl,

8732 J. Phys. Chem., Vol. 99, No. 21, 1995

Heiz et al. I

I

I

I

0.

-

13c%

1

Tad

Pt2 300K

dT/dt = 1.8K/s

1

300

I

400

I

1

I

500

I

600

I

71

TEMPERATURE (K) Figure 3. TPD spectra of deposits from Pt2 at different CO exposures. The inset shows the relative coverage, 8/&, of CO as a function of exposure to CO, CCO. At ECO = 5.6 x Om = 1.

three desorption peaks are observed, one at 340 K and two lower intensity peaks at 400 and 440 K. For Pt2, the relative intensity of the two higher temperature peaks is enhanced compared to the 340 K peak. Pt3 shows three high temperature CO desorption peaks superimposed on the 340 K desorption peak. The relative yield of the higher temperature CO desorption product is further enhanced on Pt3 compared to Ptl or Pt2 deposits. Figure 3 shows a series of desorption spectra for different CO exposures on deposits made from Pt2. These experiments were done on the same Pt2 deposited layer, interrupting the temperature program at 550 K in each case, followed by a different exposure to CO for the next experiment. The experiments shown in Figure 3 indicate that the distribution between CO species in the 340 K desorption state and the higher temperature states does not change markedly as the CO coverage is increased, The inset in Figure 3 shows that the quantity of observed CO increases linearly with CO exposure, which suggests that sintering due to the heating process during TPD can be excluded. B. CO Isotopic Studies. To provide a better understanding of the desorption process, isotopic labeling experiments were performed as follows: The sample was prepared and then dosed at 300 K with 5.6 x IOl5 I3CO molecules/cm2. The sample was then heated to just above a specific desorption peak, cooled to 300 K, and dosed again with natural abundance I2CO. In the following TPD experiment, both I2CO and I3CO desorption signals were observed. Due to poor signal-to-noise ratios for I2CO, the experiment was repeated by dosing first with 12C0, followed by I3CO after the heating and cooling cycle. The results from two of these experiments are shown in Figure 4. The higher temperature curve (A.2 of Figure 4) results from first dosing with I3CO, heating to 353 K, cooling to 300 K, dosing with l2CO,and then monitoring 13C0during TPD. Note that almost no I3C0 is observed for T < 350 K. In addition the amount of I3CO desorbed is identical (within our experimental uncertainities) to that which would have been obtained in this high temperature range if I2CO had not been adsorbed. When the Pt2 deposit was initially dosed with I2CO, heated to

4Qo

SO0

6

TEMPERATURE (K) Figure 4. TPD spectra of I3COfrom deposits produced from Ptz (CO site exchange experiments using sequential isotopic CO adsorption): (Al) TPD after dosing with WO, annealing to 353 K, and readsorption of I3CO at 300 K; (A2) TPD after dosing with 13C0, annealing to 353 K, and readsorption of "CO at 300 K; ( B l ) TPD after dosing with T O , annealing to 398 K, and readsorption of I3CO at 300 K; (B2) TPD after dosing with I3CO,annealing to 398 K, and readsorption of T O at 300 K.

353 K, cooled, and dosed with I3CO,the I3CO desorption shown by A.l of Figure 4 was obtained. For Figure 4B the same procedure was followed except that the Pt2/Si02 sample was heated to 398 K. Similar results were obtained for CO desorption from Ptl. These results show that no exchange of CO occurs between the different binding sites and suggests that the different binding sites are independent. Figure 5 shows the CO desorption behavior from Pt( 111) (Figure 5A) and Pt( 112) (Figure 5B) single crystal surfaces as a function of exposure. The results are useful for comparison with thermal desorption spectra obtained on the Pt, cluster deposits. For Pt( 111) in Figure 5A, the CO desorption occurs at a peak temperature of 470 K in the low coverage limit. On the stepped Pt(112) crystal, CO desorption occurs at 520 K in the low coverage limit. As the CO coverage increases and repulsive CO- C O interactions occur, desorption at lower temperatures is seen for both Pt crystals, but on the Pt(112) surface a resolved high temperature CO desorption process is seen at all coverages. For Pt(ll1) all CO desorption occurs between 300 and about 470 K; for Pt( 112), the CO desorption range is shifted slightly higher from 340 to about 520 K, and two desorption processes are clearly resolved. Figure 6 shows results of CO desorption from three different Pt deposits on a Si02 film. The Pt is deposited from a conventional Pt evaporation source: CO desorption is shown from a thick Pt film, a medium (0.3 monolayer) and a low (0.01 monolayer) Pt coverage. The desorption kinetics from the thick film are quite similar to these from Pt( 11l), as may be seen by comparison with Figure 5A. As expected for highly dispersed Pt deposits on SiO2, the CO TPD from the 0.01 monolayer Pt/ Si02 preparation is nearly identical to that for the Pt layer prepared by deposition of Ptl+ ions. A comparison of the results in Figures 5 and 6 with those in Figures 2 and 3 shows that the CO desorption kinetics for Ptl,

CO Chemisorption on Monodispersed Pt Clusters on Si02

. I . Phys. Chem., Vol. 99, No. 21, 1995 8733 ~

'2c1

TI 5 ~ 1 0 - ~ A

~~

0.01 ML Pt/SiO,

0.3 M L Pt/Si02

80/Pt(l 111

e/e 1.00 0.70 0.50

0.30 0.20 0.1 5 0.10

0.04 6.02

300

400

500

1

I

I

I

I

I

1

I

2c180/Pt( iX10-7~

Tad

a

1 1 2)

220

K

dT/dt = 2K /s

e/e, 0.92 0.78 0.61 0.53 0.48 0.41 0.37 0.31 0.23

300

400

500

400 500 600

.

)

400 500 600 ;

TEMPERATURE (K)

Figure 6. CO desorption from three different evaporated Pt films on Si02lSi prepared using a conventional Pt filament evaporation source.

600

TEMPERATURE ( K ) I

---

600

TEMPERATURE ( K ) Figure 5. 1 2 C ' 8 0temperature programmed desorption as a function of increasing CO exposure: from Pt( 111) (A, top) and from Pt(l12) (B, bottom).

P t 2 , and Pt3 deposits on Si02 are completely different from CO

desorption from single crystal surfaces or from bulk films of deposited Pt on Si02.

IV. Discussion A. Background Information. To our knowledge these are the first adsorbate desorption measurements from metal particles

deposited from mass-selected metal cluster beams. Several studies in the past 10 years have investigated small supported metal particles formed by evaporating metal from a hot filament. All of these s t ~ d i e s ~on~ supported -~~ Pt particles produced adsorbate desorption profiles similar to the bulk metal, even for particles as small as 2 nm in diameter (about 30-40 atoms). The desorption behavior of these particles was compared to flat and stepped Pt surfaces; changes as a function of particle size were interpreted as a change in the population of bulklike adsorption sites of flat and stepped single crystals. For Pt single crystals several measurements26of the initial (zero coverage) desorption activation energy (E'd) of CO gave values of 36.5 and 39.5 kcal/mol for Pt(557) and Pt( 112) and 33 kcaVmol for Pt(ll1). Ab initio generalized valence band and correlation-consistent configuration interaction studies of CO interacting with Pt atomsz7lead to adiabatic Pt-CO dissociation energies of 15.4 kcaVmo1; a corrected estimate of 18.5 kcal/mol was obtained for E"d(C0). It is reported that co bonds to a single Pt atom by a a-donorln-back-bonding mechanism yielding linear geometries with u repulsive effects due to unpaired 6s electrons. Pt-Pt bonding in the case of the dimer and trimer involves primarily 6s electrons, which diminishes these u repulsive effects. CO primarily interacts with the attractive d9 configuration for bigger clusters and in the bulk limit, which leads to higher CO adsorption energies in comparison to the adsorption energy on single Pt atoms. B. Comparison of Experimental Results. The basic experiments reported here may be summarized as follows: (1) All highly dispersed PtJSiOZ surfaces exhibit a distinct -340 K CO desorption state which is not observed for CO on Pt single crystals or for bulk Pt films on Si02. The relative yield of CO in this 340 K state is highest for depositions made from Pt, and decreases systematically for deposits made from Pt2 and Pt3 clusters (Figure 2 ) . Using the Chan et al. method,28 a desorption activation energy of 16 kcal/mol is estimated for this CO desorption state on deposits made from Ptl clusters. (2) All Ptx/Si02 made from clusters (Ptl, Pt2, and Pt3) and from bulk sublimation exhibit one or more CO desorption states above -400 K (Figures 2 and 6).

8734 J. Phys. Chem., Vol. 99, No. 21, I995

(3) Isotopic exchange of CO does not occur between the 340 K CO state and the CO desorption states desorbing above -400 K, indicating that the states do not interconvert on heating, and therefore the species responsible for these states to not coexist on the same Pt, particle (Figure 4). (4) The CO desorption states fill at a constant rate independent of CO coverage, with little variation in the relative population of the 340 K CO and the higher temperature CO states over the entire CO coverage range (Figure 3). This linearity of the amount of desorbing CO with CO exposure indicates that there is no poisoning of the different CO sites, e.g. by CO dissociation. The fact that the relative population of the CO sites remains constant and no new CO sites appear shows that the particle size distribution (if present) does not change in the measured temperature range by sintering. (5) The desorption temperature of the different CO binding sites shows no variation over the entire CO coverage range. This indicates that C O * C O interaction between CO molecules in the different types of bonding sites does not occur, again suggesting that the desorption states do not originate as a result of CO- * C O interaction on a single Pt, particle. (6) Thick Pt films (bulklike Pt) deposited on the Si02 substrate closely resemble Pt( 111) in CO desorption kinetics (Figure 6). C. Two Possible Origins of Characteristic CO Desorption States. One explanation of the results is that the 340 K CO desorption state originates from single Pt atoms adsorbed on Si02. This is based the observation of this 340 K CO desorption state for Ptl, Pt2, Pt3, and low coverage Pt sublimationdeposition experiments on Si02. Furthermore, the measured activation energy of CO desorption (16 kcal/mol) agrees well with ab initio calculations for a Pt-CO specie^.^' The observation of the 340 K CO state for Pt2 and Pt3 suggests therefore that the cluster deposition onto Si02 is dissociative in some cases at the 10 & 5 eV impact energy used here. The higher temperature CO desorption states may be interpreted in different ways, as listed below: (1) The high temperature CO desorption states could originate from larger clusters of Pt, which behave more as bulklike Pt, clusters in their CO adsorptioddesorption properties. These clusters could be formed from Pt migratiodnucleation effects which might be induced by 5-10-eV Pt,+ collisions with Si02 at 300 K. (2) The higher temperature desorption states could originate from preserved Pt2 and Pt3 clusters, due to the possibility of bridged and 3-fold sites being available.29 A second interpretation is that the clusters land and remain intact upon deposition. In this case the different CO desorption states of each cluster size are proposed to originate from monodispersed Pt, species, which interact with different binding sites on the Si02 substrate to produce various Pt coordinations. These different binding sites on the Si02 substrate may be characteristic of our Si02 film or could be introduced by the impact of the Pt, clusters upon collision with the substrate.30 In this case the multiple CO desorption states observed for the Pt clusters may be due to differently bound Pt particles, with the 340 K CO desorption state originating from the most Pt atom-like CO adsorption site. This is in accordance with the good agreement between the 340 K CO desorption energy and the theoretical Pt-CO dissociation energy. Our experiments cannot completely discriminate between these two interpretations.

V. Conclusions The following conclusions may be made for these sizeselected metal cluster desorption experiments.

Heiz et al. (1) A unique site for CO adsorption is produced by Ptl, P t 2 , and Pt3 deposition. This site desorbs CO at -340 K with an activation energy of -16 kcal/mol and is thought to be due to single Pt atoms on SiO2. CO desorption kinetics of this kind are not observed on Pt(ll1) surfaces, on stepped Pt(ll1) surfaces, or on bulk WSiO2 films which resemble Pt( 111) in CO binding behavior. (2) Pt sites binding CO with higher energies are also observed from Ptl, Pt2, and Pt3 deposits. It is unclear whether CO desorption from stoichiometric clusters is being observed or whether larger clusters are being produced by migration during deposition or heating to 550 K. (3) The Pt atom sites and the additional sites binding CO with higher energies do not interact by site exchange of CO. In addition, site filling during adsorption occurs by a parallel process characteristic of independent sites, rather than by a process involving CO. * C O repulsive interactions, as is seen on Pt single crystals. (4) The role of Si02 site inhomogeneity, if any, could not be determined from this work.

Acknowledgment. We acknowledge with thanks the support of the Army Office of Research for the purchase of the Auger spectrometer used in this work under Contract No. DAAL0391-G-0323. U.H. acknowledges support of an A. W. Mellon postdoctoral fellowship from the University of Pittsburgh and a Swiss National Science Foundation postdoctoral fellowship. We also acknowledge the support of the Exxon Research and Engineering Co. in hosting U.H. and in supplying the major apparatus used in this work. References and Notes (1) Kaldor, A.: Cox, D. M.: Zakin, M. R. Adv. Chem. Phys. 1988, 70, 211. (2) Cox, D. M.; Kaldor, A.; Fayet, P.; Eberhardt, W.: Brickman, R.: Shenvood, R.; Fu, Z.; Sondericher, D. ACS Symposium Series 437; American Chemical Society: Washington, DC, 1990; Chapter 17, p 172. (3) Berry, R. S.; Beck, T. L.; Davis, H. L.; Jellinek, J. Adv. Chem. Phys. 1988, 70, 75. (4) Heiz, U.: Roethlisberger, U.;Vayloyan, A,; Schumacher, E. Isr. J. Chem. 1990, 30, 147. ( 5 ) Fayet, P.; Woeste, L. Spectrosc. Int. J . 1984, 3, 91; Surf. Sci. 1984, 156, 134; Z. Phys. 1986, 0 3 , 177. (6) Cox, D. M.; Kessler, B.: Fayet, P.: Eberhardt, W.; Fu, Z.; Sondericher, D.: Shenvood, R. D.: Kaldor, A. NanoStruct. Mater. 1992, I , 161. (7) Cox, D. M.: Kessler, B.: Fayet, P.; Eberhardt, W.; Sherwood, R. D.: Kaldor, A. MRS Symp. Proc. 1991, 206, 351. (8) Fayet, P.; Woeste, L. Sutf Sci. 1984, 156, 134. (9) Zakin, M. R.; Brickman, R. 0.:Cox, D. M.; Kaldor, A. J . Chem. Phys. 1988, 88, 3555, 6605. (10) Kaldor, A.: Cox, D. M. J . Chem. SOC.,, Faraday Trans. 1990, 86, 2459. (11) Cox, D. M.: Brickman, R.; Creegan, K.: Kaldor, A. Z. Phys. D 1991, 19, 353. (12) Tanaba, K. Solid Acids and Bases; Academic Press: New York, 1970. (13) Smentkowski, V. S.: Yates, J. T., Jr. J. Vac. Sci. Techno/. 1989, A7, 3325. (14) Colaianni, M. L.;Chen, P. J.; Gutleben, H.; Yates, J. T., Jr. Chem. Phys. Lett. 1992, 191, 561. (15) Yu, C. F.; Todorov, S. S.; Fossum, E. R. J . Vac. Sci. Technol. 1987, A5, 1569. (16) Chen, P. J.: Colaianni, M. L.: Arbab, M.: Yates, J. T., Jr. J . NonCryst. Solids 1993, 155, 131. (17) Muha, R. J.; Gates, S. M.; Basu, P.; Yates, J. T., Jr. Rev. Sci. Instrum. 1985, 56, 613. ( 1 8) Winkler, A.; Yates, J. T., Jr. J . Vac. Sci. Techno/. 1988, A6, 2929. (19) Bozack. M. J.; Muehlhoff, L.; Russell, J. N., Jr.: Choyke, W. J.; Yates, J. T., Jr. J . Vac. Sci. Technol. 1987, A5, 1. (20) Miura, H.: Gonzalez, R. D. J . Phys. Chem. 1982, 86, 1577. (21) Vannice, M. A.: Hasselbring, L. C.; Sen, B. J . C a r d 1986, 97,66. (22) Altman, E. I.; Gorte, R. J. Surf.. Sci. 1988, 195, 392. (23) Altman, E. I.; Gorte, R. J. Surf. Sci. 1986, 172, 71.

J. Phys. Chem., Vol. 99, No. 21, 1995 8735

CO Chemisorption on Monodispersed Pt Clusters on Si02 (24) Doering, D. L.; Poppa, H.; Dickinson, J. T. J. Vac. Sci. Technol.

1982,20, 827. (25) Zhu, Y.; Schmidt, L. D. Surf. Sci. 1983,129, 107. (26) Siddiqui, H. R.; Chen, P. J.; Guo, X.; Chorkendorff, I.; Yates, J. T., Jr. Unpublished results. (27) Smith, G. W.; Carter, E. A. J . Phys. Chem. 1991, 95, 2327. (28) Chen, C. M.; Aris, R.; Weinberg, W. H. Appl. Surf. Sci. 1987,1, 360. (29) [The only calculations found in the literature are density functional studies of CO chemisorbed on Rh4 and P&, which compare these different bonding sites. If one assigns the additional desorption peak of the dimer at 445 K to bridge-bound CO and the specific two peaks of the trimer at 377 and 474 K to bridge-bound and 3-fold-bound CO, respectively, one can observe the same trend as that for the P&CO complex, whose desorption energies were calculated as follows: 30.4 kcaVmol (top), 41.2 kcaYmol (bridge), and 54.2 kcal/mol (%fold). For Rh&O, little differences in the desorption energies of these three sites are postulated: 53.3 kcal/mol (top),

54.4 kcaYmol (bridge), and 49.5 kcaYmol(3-fold).] Goursot, A.; Papai, I.; Salahub, D. R. J. Am. Chem. SOC. 1992,114, 7452. (30) It is well established that, in a silicon dioxide film up to 50 8, in thickness, distorted amorphous Si02 exists containing four-, six-, seven-, eight-, and nine-membered rings of Si02 tetrahedra joined by bridging oxygen. It is possible that differently coordinated Pt species are present on the surface, which leads to different CO adsorption energies. Similar effects were observed for small Pt particles supported on alumina. A weak metal-substrate interaction for the smallest particles, and a growth mechanism on defects of alumina was suggested. Braun, W.; Kuhlenbeck, H. Sui$ Sci. 1987,180,279. Himpsel, F.J.; McFeely, F. R.; Taleb-Ibrahlmi, A,; Yarmoff, J. A. Phys. Rev. E 1988, 38, 6084. Grunthaner, F. J.; Grunthaner, P. J.; Vasquez, R. P.; Lewis, B. F.; Maserjian, J.; Madhukar, A. Phys. Rev. Lett. 1979,43, 1683. Masson, A,; Bellamy, B.; Hadjromdhane, Y.; Che, M.; Roulet, H.; Dufuour, G. Surf. Sci. 1986,173, 479. JP942424J