J . Phys. Chem. 1990, 94, 1496-1501
1496
Pt( 111) surface. This stabilization can occur either by potassium increasing the activation barrier to the formation of ethylidyne or by potassium stabilizing the ethylidene surface species. Both explanations could be important. We have argued6 that the increased electron density at the surface due to charge donation by the potassium adatoms disrupts the strong di-a bonding of ethylene to the surface, resulting in the formation of a weakly interacting, n-bonded ethylene species. Thus ethylidyne, which forms three covalent Pt-C a bonds, could be even more destabilized than ethylene on Pt( 11 1). It is reasonable then for potassium coadsorption to destabilize ethylidyne to a greater extent than ethylidene (which forms two covalent Pt-C u bonds), thereby relatively stabilizing the ethylidene intermediate and increasing the activation barrier for ethylidyne formation. In a recent paper,3s we have also argued that the gas-phase electron affinities of surface hydrocarbons can indicate which adsorbates will be stabilized by the presence of potassium, and these results indicate that ethylidene will be stabilized to a greater extent than ethylidyne due to its larger electron affinity. In that paper, we also report equilibrium thermodynamic estimates of the energetics of ethylene decomposition on Pt( 1 1 l), by calculating the heats of formation of several adsorbed hydrocarbons and using the measured activation energies for decomposition. These calculations support our proposed mechanism involving ethylidene on the unpromoted Pt( 11 1) surface. Summary
=
Potassium initially (0, 0.05) causes an increase of 2-3 kcal mol-' in the activation energy for C-H bond breaking, allowing (38) Carter, E. A,; Koel, B. E. Surf. Sci., in press.
us to trap ethylidene (CHCH3) as a stable intermediate in ethylene decomposition. Ethylidene dehydrogenates to form ethylidyne, which decomposes (as on the unpromoted Pt( 111) surface) to form C,H species above 420 K and C, sTcies above 550 K. Higher potassium coverages increase the activation energy for C-H bond cleavage further to 3 1 kcal mol-' (1 3 kcal mol-' higher than on Pt(l11) without K), stabilizing ethylidene over the range 300-420 K. Complete dehydrogenation of the surface species occurs over a small temperature range, yielding a single hydrogen desorption peak at 570 K. Thus, under these conditions, sequential C-H bond breaking reactions must have nearly the same activation energies. Ethylene decomposition on the K-promoted Pt( 11 1) surface proceeds through an ethylidene intermediate. We propose that this intermediate is also important in ethylene decomposition on the unpromoted Pt( 11 1) surface. In regard to the origin of the effect of K on ethylene surface chemistry, we conclude that geometric (ensemble size) effects due to K blocking sites required for ethylene decomposition are largely responsible for the decrease in the amount of dehydrogenation occurring (irreversible adsorption), but that the change in activation energy for C H bond cleavage is clearly an electron effect due to the chemical modification of the Pt electron structure. Acknowledgment. Acknowledgment is made to the U S . Department of Energy, Office of Basic Energy Sciences, Chemical Sciences Division, for support of this work. We thank Professor J. M. White and Professor F. Zaera for useful discussions and for making data available to us prior to publication. We also gratefully acknowledge Professor E. A. Carter for helpful comments on the paper. Registry No. K, 7440-09-7; C2H2,74-85-1; Pt, 7440-06-4; C2H6, 74-84-0; CHCH,,4218-50-2.
Copper Overlayers on Rhenium(0001) J.-W. He and D. W. Goodman* Department of Chemistry, Texas A & M University, College Station, Texas 77843 (Received: April 1 1 , 1989; In Final Form: July 26, 1989)
The interaction of Cu overlayers with a Re(0001) surface has been studied by low-energy electron diffraction (LEED), Auger electron spectroscopy (AES), thermal desorption spectroscopy (TDS), and electron energy loss spectroscopy (ELS). The thermal desorption spectra of Cu from Re(0001) indicate two desorption peaks, designated as 8, and p2 states. The state begins to appear upon saturation of the p2 state. Both states follow approximate zero-order kinetics. The activation energy for the and b2 states is estimated from TD spectra to be -74 and -93 kcal/mol, respectively. AES measurements show that with Cu coverages less than 2 monolayers (ML), the Cu to Re AES ratios remain constant upon annealing to 1000 K. At higher coverages of Cu, the Cu/Re AES ratios decrease when annealed between 1IO and 900 K, consistent with the formation of 3-D Cu clusters. Some Cu nucleation appears to take place during the deposition at 1 I O K. Several surface phases are observed by LEED following Cu deposition at 1 IO K and subsequent annealing. At a submonolayer coverage of Cu, a faint 2 X 2 structure is observed upon annealing to -470 K whereas further annealing to 1070 K produces satellite patterns. Cu deposition to 2 ML at 1IO K induces a 10 X 1 phase, which upon annealing to -500 K forms a new fractional pattern with hexagonal symmetry. This pattern appears to correspond to Cu clusters with (1 11) facets, superimposed on the pattern from the substrate. A Cu coverage of -3 ML, annealed to 600 K, yields a rectangular pattern, due to an incommensurate structure of Cu with respect to the Re substrate. Further annealing nucleates the Cu atoms into clusters with (1 11) facets. ELS results support the above conclusions that the first layer of Cu is stable while multilayers of Cu coalesce into 3-D clusters when annealed to temperatures above 600 K.
-
1. Introduction
Recently considerable attention in the surface science cornmunity has been directed toward the study of bimetallic surfaces. In a typical study, one metal is deposited onto a single-crystal substrate of a second metal. Structural and electronic properties of the interface as well as the interaction of gases with the bimetallic surface can then be studied by surface analytical techniques, such as Auger electron spectroscopy (AES), low-energy electron diffraction (LEED), and thermal desorption spectroscopy 0022-3654/90/2094- 1496$02.50/0
(TDS). Such studies are fundamentally important to the understanding of bimetallic catalysts and ihin-klm materials. Previous work in our laboratories has addressed metal overlayers The On substrates Of w(loo)* w(110)~''2and Ru(0001).3-s ( 1 ) Berlowitz, P. J.; Goodman, D. W. Surf. Sci. 1987, 187, 463. (2) Berlowitz, P. J.; Goodman, D. W. Lungmuir 1988, 4, 1091. (3) Houston, J. E.; Peden, C. H . F.; Blair, D. S.; Goodman, D. W. Surf. sei. 1986, 167, 427. (4) Campbell, C. T.; Goodman, D. W. J . Phys. Chem. 1988, 92, 2569.
0 1990 American Chemical Society
Copper Overlayers on Re(0001) conclusions of these studies can be summarized as follows: ( I ) In general, at a coverage of 1 monolayer (ML), the metal overlayer exhibits a 1 X 1 LEED pattern coincident with that of the substrate. The AES ratio of the metal overlayer to the substrate at submonolayer coverages is not changed by a subsequent annealing (below the desorption temperature of the overlayer metal). These studies have concluded that the first metal monolayer adsorbs pseudomorphically with respect to the substrate. The overlayer lattice usually needs to be strained with respect to the underlying substrate due to different lattice constants of the substrate and the bulk of the deposited metal. (2) Upon multilayer deposition, the metals grow layer by layer (a Frank-Van de Merwe mechanism) as indicated by breaks in the AES ratio of the overlayer metal to the substrate. Annealing the multilayer structure causes a significant decrease in the overlayer/substrate AES ratio. This decrease is caused by either nucleation of the deposits into three-dimensional c l ~ s t e r s ' - ~or. ~ the formation of intermetallic compounds by diffusion of the overlayer metal into the s ~ b s t r a t e . ~ (3) Chemisorptive properties of gases such as H2 and CO on the bimetallic surfaces are found to be significantly different from either the clean substrates or bulk crystals of the overlayer metal.'-s As a continuation of these studies, we recently have investigated metal overlayer growth on Re(0001) as well as the adsorption of hydrogen, CO, and nitrogen on these bimetallic surfaces. Re, a neighbor of W in the periodic table, exhibits a hexagonal closed packed (hcp) structure, compared to the bcc structure of W. It is also noteworthy that Re is an important additive in reforming catalysts and recently has been found to be a promising catalyst for ammonia synthesis.68 Previous work of metal overlayer growth of Pt on Re(0001), reported by Godbey and Somorjai? found Pt to grow layer by layer in what was believed to be a hcp structure. X-ray photoemission spectra indicated the formation of alloys at the Re-Pt( 11 1) interface. With regard to the alloying properties of Cu-Re, it is known that Cu and Re are virtually immiscible, the solubility of Re being no more than 1-2 ppm in Cu.Io In this paper, we report the results of our study of Cu overlayers on Re(0001) using AES, LEED, TDS, and electron energy loss spectroscopy (ELS). The results of the interaction of H2, CO, and N2with the various Cu covered Re(0001) surfaces will be presented and discussed in the following paper in this issue.
2. Experimental Section The experiments were carried out in a conventional UHV chamber equipped with LEED, a double-pass cylindrical mirror analyzer (CMA) for AES and ELS, and a mass spectrometer for TDS. This system has been described in detail previously.'O The sample was spot-welded to two 0.5-mm Ta wires that allow resistive heating to 1500 K and sample cooling to 110 K. A W/5% ReW/26% Re thermocouple was spot-welded to the sample edge for temperature measurement. The Re(0001) surface was cleaned by a relatively simple procedure described in ref 6. Surface carbon was removed by maintaining the surface at 1200 K in oxygen (IO-' Torr) for several hours. After the oxidation process, the only surface impurity remaining was oxygen, which was then removed by flashing the sample to 1900 K in vacuum. After this cleaning procedure, AES indicated a clean surface (2.2 ML, the diminution of the Cu/Re AES ratios upon heating begins at 115 K, the Cu deposition temperature. We conclude then that the Cu agglomeration to form three-dimensional islands starts to occur even at 115 K, suggesting a Cu growth mechanism of the Stranski-Krastanov type, i.e., uniform overlayer growth followed by the formation of three-dimensional clusters. The observation of ordered LEED structures induced by the Cu deposition at 115 K (see section 3.3) also suggests that the Cu deposits are mobile on the surface at 115 K. The growth mode of Cu on Re(0001) at 115 K appears to be different from that of previous where the deposited metals have been found to grow layer by layer at a low temperature. This probably is due to a smaller activation energy for Cu diffusion on the Re(0001)
(13) He, J.-W.; Norton, P. R. Surf. Sci. 1988, 195, L199. Chem. Phys. Leu. 1988, 151, 27. ( 1 5 ) Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL,
-
(14) He, J.-W.
1983-1 984. (16) Gallon, T. M. Surf.Sci. 1969, 17, 486.
(17) Zhou, X.-L.; Yoon, L.; White, J. M. Surf. Sci. 1988, 203, 53. (18) Somorjai, G. A. Chemisrry in Two Dimensions;Cornell University Press: Ithaca. NY, 1981; p 40.
Copper Overlayers on Re( 000 1)
Figure 5. LEED pattern of a 2.2 ML of Cu-covered Re(0001) surface at 1 15 K. The surface was not annealed following the Cu deposition. A primary incident electron energy of 135 eV was used. Some spots of the substrate have been indexed.
surface compared to the corresponding diffusion energies of the previously studied bimetallic systems.’-5 Second, when samples are annealed to 900 K, the AES Cu/Re ratios at Cu coverages of >3 ML decrease to values below those found for a coverage of 2.2 ML of Cu annealed to the same temperature. Furthermore, the Cu/Re ratios with coverages of >3 ML increase upon annealing between 900-1090 K. The TD spectra indicate that in this temperature range Cu in the PI state begins to desorb. We interpret the initial decrease in the AES Cu/Re ratios upon annealing to arise from the clustering of Cu atoms originally present as a uniform layer. The subsequent increase in the AES Cu/Re ratios could be due to the redispersion of a fraction of Cu atoms within the Cu clusters during the desorption process. The reason for the dispersion is not understood. This decrease followed by an increase might be caused by alloying and dealloying between Cu and Re. The lack of miscibility of Re in Cu makes this mechanism much less plausible.10 Furthermore, if Cu-Re alloying were to occur, we would expect such alloying to occur invariably at the Cu-Re interface for multilayer Cu. This behavior would lead to a decrease in the Cu/Re AES ratio. The AES ratio for BCu < 2 ML indicates no such change following an anneal to T < 1100 K. 3.3. Low-Energy Electron Diffraction (LEED). 3.3.1. < I M L of Cu Coverage. Deposition of Cu to
