Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Electronic and chemical interactions on Rb(0001) between boron and carbon monoxide Jo& A. Rodriguez, Charles M. Truong, W. Kevin Kuhn, and D. Wayne Goodman@ Department of Chemistry, Texas A&M University, College Station, Texas 77843-325.5 (Received 22 July 199 1; accepted 18 September 199 1) The interaction between B and CO on Ru (Oool ) has been studied by means of thermal desorption mass spectroscopy, Fourier transform infrared reflection absorption spectroscopy, and x-ray photoelectron spectroscopy. Boron adatoms poison CO chemisorption approximately on a one-to-one basis. No reaction or direct bonding between B and CO was observed. The Be * *CO interaction is repulsive due to the electron-acceptor nature of both adsorbates. Boron adatoms modify the electronic and chemical properties of first and second nearest-neighbor metal atoms. In the presence of B, the CO desorption temperature decreases, whereas the 0( 1s) binding energy and C-O stretching frequency increase. These trends are a consequence of ( 1) a reduction in 2a* back donation caused by competition for metal electrons between CO and B and (2) repulsive electrostatic interactions between the negative charges on CO and B. The infrared results indicate that metal atoms strongly affected by B are only occupied when no more unperturbed Ru sites are available on the surface, or when the high temperature of the system favors CO migration (due to entropic effects) onto these energetically less favorable sites. I. INTRODUCTION The addition of impurity atoms to a metal surface can produce dramatic changes in its catalytic activity for reactions that involve the conversion of CO.’ There have been a number of recent studies which indicate that the coadsorption of CO with “additive” atoms on metal surfaces can have a profound influence on the structure of the adsorbed molecule and on the kinetics of surface reactions. l-5 For example, coadsorption with alkali metals substantially increases the heat of adsorption of CO molecules and reduces the C-O stretch frequency.2-5 Opposite phenomena are observed when CO is coadsorbed with highly electronegative species such as oxygen and sulfur.1*3.5 A key question in coadsorption studies concerns the relative importance of local (or steric) vs long-range (or electronic) effects. “Additive” atoms that are pure site blockers exert only a short-range local effect in which adsorption sites that are within - 1 bond length of the surface impurity are poisoned.6*7 In the other extreme are “additive” atoms that induce electronic perturbations over large distances and modify the chemical properties of several surface atoms.le5 In the present study, we investigate the interaction between boron and CO on Ru (000 1) by means of thermal desorption mass spectroscopy (TDS), x-ray photoelectron spectroscopy (XPS), and Fourier-transform infrared reflection absorption spectroscopy (FT-IRAS). This work adds to the few studies in which the coadsorption of boron and CO on metal surfaces has been examined with the modern techniques of surface science.’ The investigation of the surface interactions between B and CO is interesting because B has a larger electronegativity than alkali metals, but a lower electronegativity than S, Cl, and 0. a) Author to whom correspondenceshould be addressed. The adsorption of CO on Ru (000 1) has been previously investigated by TDS,9 low-energy electron diffraction (LEED),” XPS,” ultraviolet photoelectron spectroscopy (UPS), ‘* electron stimulated desorption ion angular distribution (ESDIAD), ‘* high-resolution electron energy loss spectroscopy ( HREELS ) , l3 and FT-IRAS. l4 Carbon monoxide adsorbs on Ru (000 1) molecularly, bonded through its C-end, with the molecular axis perpendicular to the surface and with a saturation coverage of 0.68 molecules per Ru surface atom. It desorbs without any dissociation in the temperature range between 300 and 500 K. TDS and XPS have been employed to study the interaction between boron and Ru(OOO~).~~ The boron adatoms were produced by dosing B, H, to the metal surface (B2 H6,g +2B, + 3H,,*; .T surf = 500 K) . The saturation coverage of B on Ru( 0001) is - 1.1 ML. At this coverage, part of the B adlayer can be removed from Ru ( 000 1) by heating to 1250 K, with a large fraction of the B atoms still on the surface at 1450 K.15 II. EXPERIMENTAL The experiments were carried out in two separate ultrahigh vacuum chambers. One of the instruments (base pressure < lo- lo Torr) has capabilities for Auger electron spectroscopy (AES), XPS, TDS, and LEED.“j The B(ls), O( Is), and Ru(3d) spectra of Sec. III were recorded with Al Ka radiation. The variations in the binding energies of the B( 1s) and 0( 1s) XPS regions were determined by referencing against the Ru( 3d,,, > and Ru( 3p,,* ) peaks for clean Ru ( 000 1) , which were set at binding energies of 280 and 483 eV, respectively. 11*17Detection was normal to the surface in AES and XPS. The second UHV chamber used in this study was equipped with a Mattson Cygnus 100 Fourier-transform in- @ 1991 American institute of Physics 0021-9606/92/010740-08!$06.00 J. Chem. Phys. 96 (l), 1 January 1992 740 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp Rodriguez &al.: Boron and carbon monoxide on Ru(0001) frared spectrometer with a liquid nitrogen-cooled mercury cadmium telluride detector that covers the frequency range between 4000 and 800 cm- ’. ‘* The FT-IRAS spectra were acquired at a resolution of 4 cm- ‘. The Ru( 0001) crystal was cleaned following procedures reported in the literature:” oxidation in oxygen ( 1 X 10 - ’ Torr, 800 K) and anneal in vacuum to 1600-1700 K. The cleanliness and long-range order of the surface were verified by means of AES, XPS, and LEED. The crystal was mounted on a manipulator capable of resistive heating to 1500 K, electron beam heating to 2500 K, and cooling to 90 K. A W-5%Re/W-26%Re thermocouple was spot-welded to the sample edge for temperature measurements. A heating rate of 5 K/s was used in the thermal desorption experiments. In this work, adsorbate coverages are reported with respect to the number of Ru( 0001) surface atoms ( 1.57 X lOI5 atoms/cm*). One adatom (or admolecule) per substrate surface atom corresponds to 19= 1.0 ML. It has been reported that a saturated monolayer of CO on Ru(OOO1) yields an absolute coverage of 0co = 0.68 ML.14 This value was used to calibrate CO coverages determined by measuring the area below the CO-TDS peaks or in the O( 1s) XPS signal. The B coverages were estimated from the B( ls)/Ru( 3d) XPS and/or B( 179)/Ru( 273) AES intensity ratios following the procedure described in Ref. 15. Ill. RESULTS A. Boron poisoning of CO chemisorption Figure 1 shows the effect of B precoverage upon the saturation coverage of CO on Ru(OOO1). Boron adatoms poison the metal surface toward CO chemisorption. In a simple Langmuir-like adsorption mode1,19*20the saturation coverage of the CO molecule varies with B precoverage as 0, = 0 “Co( 1 - me, )“, where 0 so is the saturation coverage of CO on clean Ru(OOO1 ), n is the number of sites required for adsorption of the molecule, and m is the number of sites poisoned by a boron adatom. The reported saturation 741 coverage for CO on clean Ru (C)001), 0 “,, = 0.68 ML,14 implies that - 1.5 Ru surface atoms are required for each CO adsorption site. The TDS and XPS data in Fig. 1 are fit well by the following mathematical expression: &, = 0.68( 1 - 1.458,) I.‘. This implies that on average each boron adatom poisons - 1.5 metal surface atoms. Therefore, we can conclude that B exerts a short-range local effect on the chemisorption of CO, in the sense that only surface metal atoms that are close to the impurity atoms are poisoned. The TDS, XPS, and FT-IRAS results presented below indicate that a boron adatom also modifies the electronic and chemical properties of first and second nearestneighbor Ru atoms. B. CO/B/Ru(OOOl): Thermal desorption spectra A series of thermal desorption spectra for CO adsorbed on B/Ru(OOOl) is shown in Fig. 2. In a typical experiment the boron adlayer was generated by dosing diborane to Ru(OOO1) at 500 K,” then the surface was cooled to 90 K and saturated with CO. With no boron present, CO desorbs from Ru(OOO1) at temperatures between 300 and 500 K. The line shape and desorption temperatures in the spectrum are in good agreement with results presented in the literature for the CO/Ru(OOOl ) system.9,21 The presence of boron adatoms induces a reduction in the area below the CO-TDS peaks and a significant decrease in the desorption temperature of the CO molecule. Figure 3 displays the evolution of the CO thermal desorption spectra for different B precoverages and a fixed CO coverage of - 0.15 ML. On clean Ru (000 1) , the CO-TDS peak is centered at -485 K. The width of the CO desorption peak increases with increasing B coverage. At 0, = 0.48 ML, peaks at - 300 and 395 K are observed. Assuming first- CO/B/Ru(OOOl) - 0~0 = SATURATION CO/B/Ru(OOOl) BORON COVERAGE, ML FIG. 1. Effect of B coverage upon the saturation coverage of CO on Ru(0001) at 90 K. The amount of CO present on the surface was determined by measuring the areasbelow the CO-TDS or O( 1s) XPS peaks.The solid curve represents theoretical values calculated from the expression: e, =0.68(1 - 1.458,)“. 1 150 I 250 , I 350 I I 450 TEMPERATURE, I I 550 I I 650 K FIG. 2. TDS for CO adsorbedon Ru(OC01) surfaceswith different B precoverages.The spectra correspond to saturation CO coveragesat 90 K. J. Chem. Phys., Vol. 96, No. 1, I January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp Rodriguez eta/.: Boron and carbon monoxide on Ru(0001) 742 CO/B/Ru(0001) e,,.15~ $ z 3 :” s II 2 E a g 8 1 TEMPERATURE, K FIG. 3. TDS acquired after dosing 0.15 ML of CO to Ru(ODO1)surfaces with different precoveragesof B at 90 K. 538 534 BINDING 532 530 528 ENERGY, eV FIG. 4.0( 1s) XPS spectra for saturation coveragesof CO on B/Ru(OOOl ) surfaces.Temperature = 90 K. order desorption kinetics with the pre-exponential factor of lOI s- ’ reported for low coverages of CO on Ru(OOO~),~ a Redhead analysis” yields desorption activation energies of 22.2 and 29.4 kcal/mol for the new CO adsorption states induced by boron. These values are considerably smaller than the activation energy of 36.3 kcal/mol observed for 0.15 ML of CO on Ru(OOO1). Thus, the Be * *CO interaction reduces the adsorption energy of carbon monoxide on Ru(OOO1). C. CO/B/Ru(OOOl): X-ray photoelectron spectra 0( Is) XPS spectra for CO adsorbed on Ru( 0001) and B/Ru(OOOl) (0, = 0.12,0.28, and 0.42 ML) are shown in Fig. 4. The spectra were taken at 90 K after saturating the surfaces with CO. The intensity of the O( 1s) signal decreases with increasing B precoverage. Simultaneously, there is a shift in the peak position toward higher binding energy. The O( 1s) feature for CO on clean Ru(OOO1) appears at - 1.5 eV lower binding energy than for the adsorption of CO on a surface covered with 0.42 ML of B. No O( Is) signal was observed after annealing the CO/B/Ru( 0001) surfaces to 500 K. This result indicates that CO does not dissociate on clean Ru(OOO1) or on B/Ru( 0001) surfaces under ultrahigh vacuum conditions. Boron-oxygen adlayers are stable on Ru(OOO1) up to temperatures above 1000 K.15 Figure 5 illustrates the effect of CO adsorption on the B( 1s) binding energy of B layers adsorbed on Ru(OOO1). For 0, ~0.3 ML, the presence of CO induces a small shift ( -0.4 eV) toward higher binding energy in the B ( 1s) peak position. At $ = 0.42 ML, CO produces both a broadening of the B( Is) signal and features at higher binding energy than those in the corresponding spectrum of B/Ru (000 1) . The effects of CO adsorption on the B ( 1s) peak position of the B/Ru (000 1) surfaces were reversible. The B ( 1s) binding energy found after desorption of the CO molecule in most cases was identical to that seen before adsorption. D. CO/B/Ru(OOOl): FT-IRAS spectra Our FT-IRAS results for CO on clean Ru (000 1) agree well with published results for this system.14 Figure 6 shows infrared spectra for different coverages of CO on clean Ru (000 1) . The spectra were acquired at a temperature of 220 K. After each IR spectrum was recorded, the relative CO coverage was determined from integration of the flash B(ls) X P S CO/B/Ru(0001) Bco = SATURATION 192 190 BINDING 108 104 188 ENERGY, eV FIG. 5. Effect ofC0 adsorption on the B( 1s) XPS spectra of B/Ru(OCOl) surfaces.Temperature = 90 K. J. Chem. Phys., Vol. 96, No. 1,l January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp Rodriguez Hal.: Boron and carbon monoxide on Ru(0001) 743 FT-IRAS CO/B/Ru(OOOi) 0co = SATURATION IO , '"yq 2050 1600 2doo 21bo 2100 0.11 WAVENUMBER, 2000 WAVENUMBER, woo cm-’ 1950 cm*’ CO COVERAGE, ML FIG. 7. FT-IRAS spectra for CO adsorbed on Ru(OCO1)surfaces with different precoveragesof B. Temperature = 220 K. FIG. 6. Left: infrared spectra for different coverages of CO on clean Ru(0001) at a temperature of 220 K. Right: IR-frequency and integrated IR-intensity for CO/Ru(OOOl) at 220 K as a function of CO coverage. desorption curve. In the spectra, the CO stretch frequency appears between 2070 and 2000 cm - ‘, the range usually expected for terminal (or linear) CO on metal surfaces. In agreement with previous HREELS and FT-IRAS studies for CO/Ru(OO01),‘3*‘4 we did not see any evidence for existence of bridge-bonded CO in this adsorption system. In the FT-IRAS spectra of CO on clean Ru(OOO1 ), there was a slight temperature dependence in the peak position. At all temperatures, there was a monotonic shift of the peak position to higher wave numbers with increasing CO coverage. This shift is mainly a consequence of dipole-dipole coupling and a reduction in electron back-donation from Ru into the CO(2?r*) orbitals. l4 In general, there was not a linear relationship between the integrated IR intensity and the CO coverage. In Fig. 6(b), a maximum in the IR-intensity is observed at -0.35 ML of CO. The subsequent decrease in IR intensity with increasing &, has been attributed to strong lateral interactions between the molecules in the adlayer (i.e., dipole-dipole coupling, dipole screening, and changes in the dynamic polarizability ) .14 FT-IRAS spectra for CO adsorbed on Ru( 0001) surfaces with different precoverages of B ( Br, = 0.07,0.18,0.33, and 0.40 ML) are displayed in Fig. 7. The spectra were recorded at 220 K, after saturating the surfaces with CO. On clean Ru(OOO1 ), a single absorption band is observed at 2062 cm - ‘. As the B precoverage is increased above 0.15 ML, a peak at 2072 cm - ’ appears. The intensity of this new absorption band increases with the B coverage and is the dominant feature in the spectrum at 8, = 0.4 ML. For 0.15 < 8, < 0.35 ML, the infrared spectra are characterized by the presence of at least two peaks. The peak at 2072 cm - ’ is probably associated with CO bonded to Ru sites that are strongly affected by the presence of B adatoms, whereas the peak with a frequency below 2060 cm - ’ likely corresponds to CO on Ru atoms weakly perturbedor unaffectedby B. At 0, = 0.4 ML, a large fraction of the Ru surface atoms is blocked by boron and the remaining free sites show large perturbations in its chemical and electronic properties. The IR results of Fig. 7 are consistent with a model in which boron adatoms exert only short-range electronic perturbations, modifying the chemisorption properties of their first and second nearest-neighbor metal atoms. Figure 8 shows the integrated IR intensities for saturation coverages of CO on B/Ru (0001) surfaces. The intensities have been normalized to the value found at eB = 0. 0co was determined by measuring the area below the CO-TDS peaks. The IR intensity is enhanced as B is added, until a maximum is reached at 0.18<8, ~0.25 ML. This increase can be attributed to a reduction in the saturation coverage of 8co ,.3, 0.51 0.0 0;" SATURATION, ML 0;5 0.4Q-,92 I I 0.2 BORON 0; I I 0.4 COVERAGE, I I 0.6 ML FIG. 8. Integrated IR-intensities ( 1950-2150 cm - ’) for saturation coveragesof CO on B/Ru (000 1) surfaces at 220 K. The CO coveragewas determined from thermal desorption experiments. The intensities were normalized to the value found at 0s = 0 and 0cc = 0.68 ML. J. Chem. Phys., Vol. 96, No. 1,l January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp 744 Rodriguez et&: r- FT-IRAS: CO/B/Ru(0001) I 2072 2042 2020 0.10 A I 2100 Boron and carbon monoxide on Ru(0001) I 2600 WAVENUMBER, 1900 cm-’ FIG. 9. FT-IRAS spectra acquired after dosing different CO coveragesto a Ru(OOO1)surface precovered with 0.26 ML of B. Temperature = 220 K. The CO coveragewas determined from thermal desorption experiments. CO (see Fig. 6). For 8, > 0.25 ML, there is a linear decrease in the IR intensity with increasing B coverage. Under these conditions the CO coverage is low and the effects of CO. * *CO interactions upon the IR cross section are negligible. Figure 9 shows FT-IRAS spectra acquired after dosing different amounts of CO to a surface with 0.26 ML of B at 220 K. For all CO coverages the IR absorption band shows two peaks. Both features shift to higher frequency when the CO coverage increases, probably as a consequence of an enhancement in dipole-dipole coupling and a reduction in rback-donation. In all the cases examined, the measured peak positions are at higher frequency (2040 cm-‘) than the corresponding values for CO on clean Ru( 0001) . For &o < 0.2 ML, the peak at lower frequency presents the larger intensity, whereas at saturation (0co = 0.31 ML) the high frequency peak dominates in the IR spectrum. This type of behavior indicates that the CO adsorption state weakly perturbed by boron adatoms is populated first. This is not surprising if one takes into consideration that such a state probably has the largest heat of adsorption on the B/Ru (0001) surface (see trends in Figs. 1 and 2). Figure lO( a) illustrates the effect of temperature upon the FT-IRAS spectrum of a Ru (0001) surface with 0.12 ML of B and 0.51 ML of CO (saturation). The B/Ru(OOOl) surface was saturated with CO at 90 K, and the spectra were acquired in a sequential way after annealing to the indicated temperatures for 2 min. For 0, = 0.12 ML and 19,~ = 0.51 ML only minor changes are observed in the IR spectrum when the surface temperature is increased from 90 to 170 K. The peak position is centered around 2055 cm - i. At 220 K a shoulder toward higher frequency starts to appear and the main peak shifts to 2050 cm-‘. The spectrum at 270 K shows a new peak at 2070 cm-‘, with the main peak centered now at 2046 cm - i. Thus, the IR data indicate the existence of an inhomogeneous surface on which the population of the different adsorption states changes with temperature. When the temperature is increased, the CO molecules mi- I 2100 2000 WAVENUMBER 2100 2000 (cm-‘) FIG. 10. Effectoftemperatureupon theCO-ET-IRASspectraofRu(OOO1) surfaceswith 0.12 and 0.35 ML of B. The surfaces were saturated with CO at 90 K, and the spectra were acquired after annealing to the indicated temperatures during 2 min. grate (due to entropic effects) from Ru sites weakly affected by B (peak at 2050-2055 cm- i ) onto metal atoms largely perturbed by the “additive” (peak at - 2070 cm - ’) . In Fig. 10(b), the spectrum for 6a =0.35 ML and 0co = 0.20 ML (saturation) shows again different line shapes at 90 and 220 K. For the lowest temperature, the absorption band displays two unresolved peaks, with the one at higher frequency centered at - 2066 cm - ‘. As the temperature is raised, intensity shifts from the peak at lower frequency into the peak at higher frequency. Above 220 K, two well-defined peaks are observed with frequencies of 2072 and - 2050 cm - I. In general, the FT-IRAS results of Figs. 9 and 10 show that at low CO coverages and/or low surface temperatures, Ru atoms which are moderately perturbed by B are preferentially populated. Metal atoms strongly affected by B are only occupied when no more unperturbed Ru sites are available on the surface or when the high temperature of the system allows CO adsorption on these energetically less favorable sites due to entropic effects. IV. DISCUSSION The results of Fig. 1 indicate that boron adatoms poison the adsorption of CO on Ru (0001) more or less on a one-toone basis. This agrees well with data for the CO/B/Mo( 100) system,’ in which half a monolayer of boron reduces the saturation coverage of CO by half (t9a = 0.5 ML, e,, = 0.5 ML). Thus it appears that the poisoning effects of boron are short-range in nature, very different from the long-range poisoning found when additives with large electronegativities (O,S,CI) were adsorbed on several transition metal surfaces.’ On Ru (000 1) , however, no longrange poisoning of CO chemisorption has been observed for different “additives” investigated in the past.21V23-26Cu at- J. Chem. Phys., Vol. 96, No. 1,l January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp Rodriguez eta/: Boron and carbon monoxide on Ru(0001) tenuates CO chemisorption over Ru (0001) via a simple site blocking mechanism. 21 In thermal desorption experiments, there is a monotonic decrease upon addition of Cu of the CO structure associated with Ru and an increase of the structure corresponding to Cu. a For CO/O/Ru(OOOl ),23 a lowering in the saturation amount of absorbed CO by 50% is found at 0, = 0.25 ML. An oxygen-saturated Ru(OOO1) surface coo = 0.5 ML) adsorbed 35% as much CO as a clean surface.23 Finally, sulfur precoverages of 0.18 and 0.33 ML reduced the saturation coverage of CO on Ru( 0001) by 29% and 73%, respectively. 24 Carbon monoxide did not adsorb onto a Ru(OOO1) surface with 0.5 ML of sulfur. The TDS, XPS, and FT-IRAS results presented in the previous section show that boron adatoms can affect the chemical properties of the neighboring metal atoms. This phenomenon can be a consequence of electronic perturbations induced by the “additive” on the Ru surface atoms (through-metal interaction) or a product of direct electrostatic interactions between B and CO (through-space interaction). In order to understand the effects of B on the surface chemistry of CO on Ru(OOO1 ), we must examine first the nature of the Ru-CO and Ru-B adsorption bonds. The bonding mechanism of CO on metal surfaces involves B donation of electron density from CO into the unoccupied metal orbitals, and r-back-donation from occupied metal orbitals into the lowest unoccupied molecular orbital (27r*) of the CO mo1ecule.27-30Theoretical results with the constrained space orbital variation (CSOV) method show that, in this type of synergistic bond, r-back-donation is energetically more important in determining the character of the bond than is a donation. 28 Inverse photoemission results have supported the predominant importance of 2zr*-backdonation in CO chemisorption on Ru.~’ The population of the antibonding 2~ orbitals of CO upon adsorption on Ru(OOO1) is consistent with results of FT-IRAS, which show that the C-O stretching frequency of adsorbed CO is lower than that of free CO [ 2143 cm- ’ (Ref. 32) 1. It is wellknown that CO acts as a net electron acceptor when adsorbed on transition metals.33*34Theoretical calculations for CO adsorbed on typical transition metals indicate a charge transfer of 0.2-0.6e - from the substrate to C0.33,34 Experimental results for CO/Ru( Of!01 ) show an increase of 0.6 eV in the work function of Ru(OOO~),~~ suggesting a partial negative charge on adsorbed CO. Recently, we have examined the bonding interactions between atomic boron and Ru(OOO1) by performing quantum-chemical calculations based on the molecular orbital self-consistent field method referred as IND0/1.36 Previous studies have shown that INDO is a very good quantumchemical method for addressing the bonding mechanism of species adsorbed on metals.” The Ru(OOO1) surface was modeled by two-layer clusters containing between 18 and 15 metal atoms, and boron was adsorbed on the threefold hollow sites.36 In general, the results of the calculations showed an adsorption bond dominated by the interaction between the B (2~) orbitals and the valence 46 orbitals of Ru. For bulk iron borides (Fe, B and FeB), photoemission experiments and ab initio calculations show that the Fe-B bonds are mainly a consequence of the mixing of B (2~) states with 745 iron 3d valence states. 38 For B/Ru ( 000 1) , the percentage of the adsorption bond in which each type of Ru orbital participated was approximately as follows:36 5s, 15%; 5p, 25%; and 4d, 60%. From the viewpoint of the adsorbate, the B( 2p) orbitals were involved in - 70% of the chemisorption bond, whereas the B( 2.~) orbitals participated in only 30%.36 Adsorption of B induced a strong hybridization between the B( ;Is) and B( 2p) levels.36 Boron adatoms acted as moderate electron acceptors on Ru( 0001) with a charge of - - 0.37e calculated for adsorbed B.36 The direction of charge transfer predicted by INDO/l for the B/Ru( 0001) system is in good agreement with experimental measurements that show an increase in the work function of Mo( 100) when boron is adsorbed.’ Our results of Sec. III B show that coadsorption with boron significantly decreases the desorption temperature of CO on Ru(OOO1). At 0, = 0.48 ML, the activation energy for desorption of CO is reduced by - 10 kcal/mol. We suggest that this is a product of ( 1) a reduction in 2rr*-backbonding caused by a competition for metal electrons between CO and B, and (2) repulsive electrostatic interactions between the negative charges on CO and B. Both types of phenomena seem to be of short range nature, affecting only CO molecules bonded to Ru atoms that are first and second nearest-neighbor to B atoms. The XPS data of Sec. III C show that the 0( 1s) feature of CO appears at higher binding energy for adsorption on boron-covered Ru(OOO1) than for adsorption on clean Ru(OOO1). An identical trend was observed for the CO/Mo( 100) and CO/B/Mo( 100) systems.’ We can tentatively attribute the shift toward higher binding energy seen in the 0 ( 1s) XPS spectra of CO on B/Ru (000 1) to a simple initial-state stabilization of the 0( 1s) electronic energy level caused by a reduction in electron density on the adsorbed molecule. In a similar way, the CO-induced shift in B( 1s) peak position observed in Fig. 5 is probably a consequence of a decrease in the negative charge on the boron adatoms. When boron and CO are coadsorbed on Ru( 0001 >, they compete for metal electrons, and this produces adsorbates that have a lower electron density and higher core level binding energies than those found for CO/Ru(OOOl) and B/Ru(OOOl). The infrared data of Figs. 7, 9, and 10 show that for a given CO coverage the C-O stretch frequency is larger on B/Ru ( 0001) than on Ru (0001) . The strong dependence of the frequency with coverage (see Fig. 6) makes it difficult to separate quantitatively the effects of B. The CO coverages reported in Figs. 7-10 represent an average over the whole surface, and the real local coverages are probably different. It is not difficult to imagine that the CO molecules will pack preferentially on Ru sites not strongly perturbed by B, resulting in local coverages that are larger than the “average” coverage determined from TDS. However, it is clear that a large part of the B-induced shift in C-O stretch frequency is due to changes in the Ru-CO bond: the frequency of 2072 cm - ’seen for a surface with 0.4 ML of B and 0.17 ML of CO (saturation) is larger than that of 2062 cm - ’found for saturation CO on clean Ru(OOO1) (&, = 0.68 ML). This increase is probably caused by a reduction in 2rr*-back-dona- J. Chem. Phys., Vol. 96, No. 1,i January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp 746 Rodriguez eta/.: Boron and carbon monoxide on Ru(0001) tion from Ru to CO and is consistent with the XPS and TDS results. Boron is less electronegative than 0, S, and Cl, but more electronegative than the alkali metals.39 The type of interaction seen in the TDS, XPS, and FT-IRAS spectra for CO/B/Ru( 0001) is similar to that found for coadsorption of CO and electronegative “additives” [0, S, Cl (Refs. 23, 24, and 29)], and contrary to that seen when CO is coadsorbed with electropositive “additives” [ Li, Na, K (Refs. 25, 26, and 29) 1. In general, electronegative “additives” produce a substantial reduction in the heat of adsorption of CO and an increase in the C-O stretch frequency.23,24,29’4042 The standard picture usedz9 to explain these effects is an extension of the basic Blyholder model for CO chemisorption on metals. *’ Electronegative “additives” withdraw electrons from the metal, becoming partially negatively charged. As a result, r-back-donation to the 2n* orbitals of CO is reduced. This weakens the metal-CO bond and increases the strength of the C-O bond. In a few coadsorption cases, this general picture has been to some extent substantiated by calculations with different quantum-chemical methods. 29*43-46In another type of mode1,47 the coadsorption effects of the “additives” are a consequence of changes in the local density of occupied states near the Fermi level rather than a product of variations in the total charge density at the adsorption site. According to this mode1,47 the “additive”-induced total charge density vanishes beyond the immediately adjacent metal atom sites due to the effects of screening by the metal electrons. However, the Fermi-level density of states, which is not screened, and which governs the ability of the surface to respond to the presence of other species, is substantially reduced by electronegative “additives” even at nonadjacent sites.47 Finally, an alternate model, which produces the same final results, involves an electrostatic, through-space (as opposed to through-metal) repulsive interaction between the charge distribution of the coadsorbed species.34 Our results indicate that the B- . -CO interaction is repulsive due to the electron-acceptor nature of both adsorbates. In a previous study, l5 we observed that B increases the adsorption energy of ammonia (an electron-donor species) on Ru (000 1) . The B * * *NH, interaction was attractive.15’36 Adsorbed boron atoms create Ru sites on the surface with partial positive charge, which facilitates electron donation from ammonia to these metal sites. In addition, there is a stabilizing through-space interaction between the partial negative charge on the boron adatoms and the partial positive charge on the ammonia molecules. We did not see any evidence that indicates direct bonding or reaction between B and CO on Ru( 0001) . A similar result was found for the NH, /B/Ru (000 1) system under UHV conditions.” It appears that B behaves as an inert “additive” when coadsorbed with species that have a large chemical stability. A totally different picture emerges when boron is coadsorbed with reactive molecules.8~‘5~48-51 B oron adatoms react extensively with N, H, (Ref. 49) and C, N, (Ref. 48) to produce species with strong B-N bonds. Boron enhances the ability of the Ru( 0001) surface to chemisorb oxygen. l5 For 0, /B/Ru( 0001 ), XPS results show the formation of boron oxides that are stable on the surface up to temperatures above 1400 K.15 V. CONCLUSIONS ( 1) Boron adatoms poison CO chemisorption on Ru (000 1) approximately on a one-to-one basis. At 100 K, CO does not adsorb on Ru (000 1) surfaces with B coverages above 0.7 ML. No reaction or bonding between B and CO was observed. (2) Boron atoms modify the chemical properties of first and second nearest-neighbor metal atoms. The CO desorption temperature decreases, whereas the O( 1s) binding energy and C-O stretching frequency increase. (3) The Be * *CO interaction is repulsive due to the electron-acceptor nature of both adsorbates. The trends in TDS, XPS, and FT-IRAS are a consequence of ( 1) a reduction in 2r*-back-donation caused by competition for metal electrons between CO and B and (2) repulsive electrostatic interactions between the negative charges on CO and B. ( 4) The FT-IRAS results for CO on B/Ru (0001) show that at low CO coverages and/or low surface temperatures, Ru atoms which are weakly perturbed by B are preferentially populated. Metal atoms strongly affected by B are only occupied when no more unperturbed Ru sites are available on the surface, or when the high temperature of the system favors CO migration (due to entropic effects) onto them. ACKNOWLEDGMENTS We acknowledge with pleasure the support of this work by the Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences. ’J. A. Rodriguez and D. W. Goodman, in Studies in Surface Science and Cu~ulysis edited by L. Guczi (Elsevier, Amsterdam, 1991), Vol. 64, Chap. 3. *H. P. Bonzel, Surf. Sci. Rep. 8,43 (1988). ’ (a) D. W. Goodman, Act. Chem. Res. 17,194 ( 1984); (b) M. Kiskinova, in Ref. 1, Chap. 2. 4 J. A. Rodriguez, W. D. Clendening, and C. T. Campbell, J. Phys. Chem. 93, 5238 (1989), and referencestherein. 5J. M. White and S. Akhter, CRC Crit. Rev. Solid State Mater. Sci. 14, 130 (1988). 6C. T. Campbell, J. M. Campbell, P. J. Dalton, F. C. Henn, J. A. Rodriguez, and S. G. Seimanides,J. Phys. Chem. 93,806 (1989). ‘C. T. Campbell, Annu. Rev. Phys. Chem. 41,775 ( 1990). ‘T. B. Fryberger, J. L. Grant, and P. C. Stair, Langmuir 3, 1015 ( 1987). 9 (a) H. Pfniir, P. Feulner, H. A. Engelhardt, and D. Menzel, Chem. Phys. Lett. 59,481 (1978); (b) H. Pfniir, P. Feulner, and D. Menzel, J. Chem. Phys. 79,4613 (1983). “(a) E. Williams and W. H. Weinberg, Surf. Sci. 82, 93 (1979); (b) G. Michalk, W. Moritz, H. Pfniir, and D. Menzel, ibid. 129,92 (1983). “J. C. Fuggle, T. E. Madey, M. Steinkilberg, and D. Menzel, Surf. Sci. 82, 521 (1975). “T. E. Madey, Surf. Sci. 79,575 (1979). 13G. E. Thomas and W. H. Weinberg, J. Chem. Phys. 70,1437 ( 1979). “H. Pfniir, D. Menzel, F. M. Hoffmann, A. Ortega, and A. M. Bradshaw, Surf. Sci. 92,431 ( 1980). Is 3. A. Rodriguez, C. M. Truong, J. S. Comeille, and D. W. Goodman, J. Phys. Chem. (in press). 16R. A. Campbell and D. W. Goodman, Rev. Sci. Instrum. (in press). “C . D . Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder, and G. E. Muilenberg, Handbook of X-ray Photoelectron Spectroscopy ( Perkin-Elmer, Eden Prairie, MN, 1978). J. Chem. Phys., Vol. 96, No. 1,l January 1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp Rodriguez eta/.: BoronandcarbonmonoxideonRu(0001) “L.-W. H. Leung, J.-W. He,andD. W. Goodman, J.Chem. Phys.93,8328 (1990). I9M. T. Paffett, C. T. Campbell, and T. N. Taylor, J. Vat. Sci. Technol. A 3, 812 (1985). 2oM. Boudart and G. DjCga-Mariadassou,Kinetics of Heterogeneous Catalytic Reactions (Princeton University, Princeton, 1984). ” J. E. Houston, C. H. F. Peden, D. S. Blair, and D. W. Goodman, Surf. Sci. 167,427 (1986). “P. A. Redhead, Vacuum 12,203 (1962). “H. I. Lee, G. Praline, and J. M. White, Surf. Sci. 91, 581 (1980). “J. A. Schwarz and S. R. Kelemen, Surf. Sci. 87,510 ( 1979). “T. E. Madey and C. Bcnndorf, Surf. Sci. 164,602 (1985). 16F. P. Netzer, D. L. Doering, and T. E. Madey, Surf. Sci. 143, L363 (1984). r’ (a) G. Blyholder, J. Phys. Chem. 6&2772 (1964); (b) 79,756 (1975). ‘* (a) P. S. Bagus, K. Hermann, and C. W. Bauschlicher, 3. Chem. Phys. 81, 1966 (1984); (b) K. Hermann, P. S. Bagus, and C. J. Nelin, Phys. Rev. B 35,9467 (1987). “J. A. Rodriguez and C. T. Campbell, J. Phys. Chem. 91,216l (1987). ‘OR. A. van Santen and A. de Koster, in Ref. 1, Chap. 1. )’V. Dose and J. Rogozik, Surf. Sci. 176, L847 (1986). ‘*K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Strut(VanNostrand-Reinhold, New ture: IVConstantsofDiatomicMolecules York, 1979). “E. Shustorovitch and R. C. Baetzold, Science 227, 876 (1985). “(a) J. K. Ndrskov, S. Holloway, and N. D. Lang, Surf. Sci. 137, 65 ( 1984); (b) N. D. Lang, S. Holloway, and J. K. Ndrskov, ibid. 150, 24 (1985). 747 3sF. Nitschke, G. Ertl, and J. Kiippexs, J. Chem. Phys. 74,591l ( 1981). 36(a) J. A. Rodriguez, C. M. Truong, and D. W. Goodman (in preparation); (b) J. A. Rodriguez, C. M. Truong, and D. W. Goodman, J. Vat. Sci. Technol. A (in press). 373. A. Rodriguez, Surf. Sci. 234,421 ( 1990)) and referencestherein. 38(a) D. J. JoynerandR. F. Willis,Philos. Mag. A43,815 (1981); (b) D. J. Joyner, 0. Johnson, D. M. Hercules, D. W. Bullett, and J. H. Weaver, Phys. Rev. B 24,3122 (1981). “9T. Moeller, Inorganic Chemistv (Wiley, New York, 1982), p. 82. “‘E. L. Hardegree, P. Ho, and J. M. White, Surf Sci. 165,488 (1986). ” J. L. Gland, R. J. Madix, F. W. McCabe+and C. M. DeMaggio, Surf. Sci. 143,46 (1984). ‘*M. Trenary, K. Uram, and J. T. Yates, Surf. Sci. 157,512( 1985). 43N. K. Ray and A. B. Anderson, Surf. Sci. 125,803 (1983). *R. C. Baetzold, Phys. Rev. B 30,687O ( 1984). 4sM. C. Zonnevylle and R. Hoffmann, Langmuir 8,452 (1987). &J. M. MacLean, D. D. Vvedensky, J. B. Pendry, and R. W. Joyner, J. Chem. Sot. Faraday Trans. 183,1945 (1987). “(a) P. J. Feibelman and D. R. Haman, Surf. Sci. 149, 48 (1985); (b) Phys. Rev. L&t. 52, 61 (1964). ‘*F. Solymosi and L. Bugyi, Appl. Surf. Sci. 21, 125 (1985). 49C. M. Truong, J. A. Rodriguez, and D. W. Goodman, J. Phys. Chem. (in press). s°F. Solymosi and J. Kiss, Chem. Phys. Lett. 110, 639 (1984). “F. Solymosi, A. Berko, and I. T. Tarmoczi, Appl. Surf. Sci. 18, 233 (1984). J.Chem.Phys.,Vol.96,No.1,1 January1992 Downloaded 15 Apr 2004 to 165.91.178.213. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp