Y.B. Zhao

Adsorption of CO on Pd1/W(110)

The properties of the system CO/Pd1/W(110) have been investigated. Unlike CO/Cuj/WfllO) there are no particular anomalies in ΔH or surface entropy Ss probably because Pd matches W in size much more closely than does Cu. Δ H decreases with coverage from 21 kcal/mol at low to 12 kcal/mol near saturation coverage. Ss is initially very high, 24 cal mol−1 K−1 and decreases with coverage. It is postulated that the high values at low coverage (and relatively high temperature) correspond to CO being a 2 D gas with additional low energy vibrational degrees of freedom. Isosteric desorption rate measurements yield Edes values lower than desorption enthalpies at θ ⩾ 0.35. The desorption measurements also yield first order frequency factors ν which show a very strong compensation effect: log ν versus Edes is linear over eleven orders of magnitude, i.e. from ν = 1018s−1 at low coverage, high Edes to ν ≈ 107s−1 high coverage, low Edes. The drop in Edes below desorption enthalpies is attributed to steep decreases in sticking coefficient s with increasing surface temperature. At moderate T, s ≈ 1 at low coverage, shows a plateau and then decreases steeply as θ increases. Work function rises nearly linearly with coverage to high values; Δφ = +620 meV at saturation coverage, CO/Pd = 0.77 or 1.1 × 1015 molecules cm−2. The He I UPS spectrum for CO on Pd1/W(110) shows a 5gs−1π peak near EF−8 eV, and is very similar to that of CO/Cu1/W(110). Differences in the binding energy of the O 1s peak for CO/W(110), CO/Cu1/W(110), and CO/Pd1/W(110) can be explained almost quantitatively by the differences in the work functions of these surfaces.

Adsorption of Hg on CO, O, and H adsorbed on W(110)

Abstract The adsorption of Hg on O/W(110), CO/W(110) and H/W(110) was studied by work function, UPS, XPS, LEED, energy loss and thermal desorption measurements. Hg adsorbs on these surfaces without causing geometric changes in the underlying chemisorbate. The monolayer coverage of Hg on these surfaces is the same as on clean W(110), and was measured to be ∼ 1 × 10 15 Hg atoms/cm 2 or Hg/W = 0.72. One monolayer of Hg screens the OW and HW dipole moments nearly completely and gives work functions almost identical to that of multilayer Hg deposits on clean W(110). In the case of CO screening is incomplete. In all cases the Hg core levels are unshifted relative to Hg/W(110), suggesting that the Hg layer is in electronic “communication” with tungsten and that a common Fermi level exists. An Hg surface plasmon at ∼ 7 eV, not seen on clean W(110) for less than 2.5 Hg layers, appears slightly above Hg monolayer coverage in all cases, but is shifted upward in energy from 7 to 8.5 eV on O/W(110). There is no wetting of higher than the first Hg layer for O/W(110), and only partial wetting of higher layers in the other cases. The Hg desorption temperature from Hg 1 /W(110) is 500 K, and is reduced to 250 K on O/W(110) and 200 K on CO/W(110), but is 500 K on H/W(110). For Hg/H 1 /W(110) the desorption spectrum of H 2 is strongly affected with 40% of H 2 desorbing at very low temperature, ∼ 200 K, the remainder more or less as from clean W(110). O 2 and CO are not adsorbed on Hg 1 /W(110) but H 2 apparently can diffuse through several Hg layers to become dissociatively adsorbed on tungsten. However for preadsorbed Hg the 200 K H 2 peak is suppressed and only the fraction corresponding to the 450–500 K peak, i.e., 60% of what is seen on clean W(110) is adsorbed. These results indicate that despite its closed 6s 2 shell atomic configuration Hg can behave very like bulk Hg metal in monolayer form on some, but apparently not all surfaces.

Adsorption, desorption and dissociation of N2 on W(110)

Abstract N 2 adsorbs molecularly on W(110) at T s ⩽ 100 K with high and nearly constant sticking coefficient ( s ≈ 0.7). Saturation coverage is estimated as N 2 / W = 0.7, from the p(4 × 1) LEED pattern seen at saturation and by analogy with CO adsorption. Desorption at ∼ 130–150 K seems to occur from more than a single binding state, even for partial coverages. The tightly bound component (75% of total N 2 adsorbed, independent of coverage) desorbs with first-order kinetics as shown by isothermal measurements with E = 10.3 kcal, v = 10 15 s −1 . Approximately 10% of total N 2 is converted to atomically bound N during desorption. The amount converted is independent of desorption temperature over the range 128–148 K, suggesting an activation energy equal to that of desorption of the tightly bound component, i.e. ∼ 10 kcal. This suggests that dissociation from preadsorbed N 2 occurs via a different channel than dissociation from impinging energetic N 2 molecules.

Adsorption of Pd on H1/W(110)

Abstract Some properties of the system Pd/H/W(110) have been examined. Adsorption of a monolayer of Pd on H1/W(110) leads to a work function of 5.18 eV, not yet that of Pd multilayers on W(110), 5.6 eV, but heading in that direction. A Pd excitation is seen at 8.5 eV; in the absence of H this energy is seen only for thicker Pd layers on W(110). It is concluded that Pd monolayers on H/W(110), and also on O/W(110) behave somewhat like multilayers of Pd on W(110), i.e., approach bulk Pd properties, although not nearly as effectively as Cu monolayers approach bulk Cu. Desorption of H2 through 1–3 Pd monolayers was observed and suggests that Pd lowers the binding of H at the WPd interface.

The adsorption of Ag on CO, O, and H covered W(110): I. Overlayer formation and stability

Abstract The adsorption of Ag mono- and multilayers on O/W(110), CO/W(110), and H/W(110) was investigated, from 90 K to Ag desorption temperatures. Auger measurements indicate that Ag growth occurs layer by layer at 90 K on all substrates. For Ag/O/W(110) dewetting, i.e. partial balling up of Ag occurs from 200 to 400 K. At T ⩾ 500 K segregation sets in, with O forced into p(1 × 1) islands, and Ag pushed into islands which can be several layers thick. This behavior is similar to that seen previously with O and Cu and O and Pd on W(110). There is also a slight amount of O left trapped under the Ag covered surface. In the case of CO partial dewetting also seems to set in at 200–300 K. Ag adsorption does not cause CO desorption. However, with an Ag overlayer there is a slight decrease from 400 to 350 K in the virgin-CO desorption peak, and the ratio of virgin to β-desorption increases with Ag coverage. This suggests that Ag competes with remaining CO for sites which could be used for dissociative, i.e. β-adsorption. CO is only very weakly adsorbed on Ag 1 /W(110); this explains the relatively small effect on v-CO desorption, since there is no driving force for CO-Ag interchange, as in the case of Cu/CO/W(110), where CO is still strongly bound on Cu 1 /W(110). The effect of Ag on H 2 desorption from H/W(110) is quite dramatic, with the appearance of a sharp low temperature peak at 150 K and a second sharp peak at 400 K, also lower than the broad main peak for H/W(110). This behavior is very similar to what was seen previously for Hg/H/W(110) and Pd/H/W(110) and may result from a modification of H binding sites on W(110) by the overlayer metal, even if the latter does not interact directly with H.

The adsorption of Ag on CO, O and H covered W(110). II: Metallicity

Abstract Monolayers of Ag adsorbed on O/W(110), CO/W(110) and H/W(110) give rise to work functions nearly identical to those of thick Ag deposits on clean W(110) and also show bulk and surface plasmon losses seen only for 3 or more Ag monolayers on clean W(110). UPS spectra show a 4d peak closer in energy to that of Ag multilayers on clean W than for a single layer. There are only very small (~ 0.1 eV) core level shifts relative to Ag 1 /W(110) or Ag 3 /W(110). These facts indicate that metal atoms capable of forming partially filled s-bands show metallicity even as monolayers, and that partial decoupling from the W substrate by intervening chemisorbates makes the properties of such monolayers closer to those of the bulk overlayer metal than when a single layer is in good physical and electronic contact with the substrate.

The adsorption of oxygen on Pd1/W(110)

The behavior of oxygen on Pd1/W(110) has been investigated from 25 to 200 K by thermal desorption, UPS, XPS, and work function measurements. At 25 K only dioxygen species are present. A weakly bound O2 layer, containing O2Pd = 0.31 or 4.4 × 1014O2 molecules/cm2 is desorbed at 35 K, leaving a coverage of O2Pd = 0.35 or 5 × 1014O2 molecules/cm2. Heating to 200 K results in desorption of molecular O2 as well as conversion to O, with O/Pd = 0.3. The molecular states, except the very weakly bound one, have high dipole moments with electron transfer to O2, and thus correspond to Superoxide and peroxide species. These have UPS spectra quite different from physisorbed O2. At 90 K adsorption is still mostly molecular with a sticking coefficient s near unity. At 200 K, adsorption is atomic with an initial s0 = 0.8. This must be contrasted with Cu1/W(110) where s0 is unity even at 300 K. The difference can be explained by the much better size match of Pd and W, than Cu and W which makes it easier for Cu to take up momentum of impinging O2 molecules. The behavior of oxygen on Pd1/W(110) is very similar to that on bulk Pd(111), suggesting that for oxygen the former surface resembles bulk Pd. This is not so for CO adsorption which is much weaker on Pd1/W(110) than on bulk Pd. The reasons for this difference are not presently understood.

Electron stimulated desorption of hydrogen from W(110)

Adsorption of H on W(110) at 90 K gives rise to H+ desorption under electron impact with high disappearance cross section, σ = 5 × 10−17cm2, while adsorption at 300 K does not yield H+. For saturated or nearly saturated layers formed at 90 K H + persists to 300 K. The amount of H desorbed by ESD is very small, as shown by thermal desorption and work function measurements. It is postulated that the H+ yielding state corresponds to adsorption at imperfections which are filled randomly at 90 K but for kinetic reasons not at 300 K. Once filled, however, they stay filled to 300 K for near-saturation coverages, but not when there are alternate sites available at lower θ, and stay empty on recooling. The cross section for H+ production was estimated from the H+/e− ratio as σH+ = (5 ± 1) × 10−23 (WH) cm2, where H/W is the ratio of H+ yielding hydrogen to W atom latter corresponding to 1.4 × 1015atoms/cm2. It was also found that submonolayer coverages of Ag post-adsorbed on H1/W(110) increase the yield and disappearance cross section of the H+ yielding state, but do not lead to appreciably more total electron stimulated H desorption. Small amounts of CO kill H+ production.

Electron stimulated processes for oxygen adsorbed on Pd1/W(110)

The effect of electron impact on oxygen adsorbed on Pd1/W(110) has been probed from 25 to 200 K. As previously shown [Y.B. Zhao and R. Gomer, Surf. Sci. 260 (1992) 129] oxygen at 25 and 90 K is present as dioxygen. The present work confirms this. Electron impact on a chemisorbed layer at 25 K leads to desorption of neutral O2 and O+ but not O2+. There may also be conversion to an intermediate state which yields O+ and is converted to adsorbed atomic O. This state could also be present ab initio. For θ = 0.5 and 25 K this state is either absent or not formed, and there is direct conversion of adsorbed O2 to adsorbed O, as well as O2 desorption. The initial desorption plus conversion cross section for O2 is very high, ∼ 10-16 cm2. Physisorbed O2 o top of the 25 K chemisorbed dioxygen layer desorbs as O2 with an even higher cross section, ∼5 × 10-16 cm2. For a layer prepared at 90 K dioxygen still seems the main species present initially but differs from the 25 K one. Electron impact leads to small amounts of neutral O2 desorption and mainly to conversion to atomic O with a cross section of 6.5 × 10-17 cm2. O+ is also seen and goes through a maximum, which can be explained by the postulated mechanism. It then decays with σ ≈ 2 × 10-8 cm2. At 200 K a monotonically decreasing O+ signal, decaying with a cross section of 1 × 10-18 cm2 is seen, indicating that only atomic O is present on the surface at this temperature.

CO oxidation on Pd1/W(110)

Abstract The oxidation of CO adsorbed on a monolayer of Pd, adsorbed on W(110) (Pd 1 /W(110)) occurs both via dioxygen (the dominant form on the surface at 90 K) and from atomic O (the dominant species at T ⩾ 200 K). In the first case CO 2 desorption peaks near 200 K, in the second the CO 2 desorption spectrum is very broad, extending from ∼ 100 to 400 K. CO preadsorption blocks dioxygen adsorption at 90 K, with O 2 adsorption absent on a saturated CO layer. CO can be adsorbed on O 0.6 /W(110) but is desorbed without oxidation. Thermodynamic arguments are given to explain this behavior. The low temperature of CO oxidation on Pd 1 /W(110) seems to result from the relatively weak binding energy of CO on this surface as well as the relatively low adsorption energy of dioxygen and the ease with which it can be converted to adsorbed atomic O and utilized in CO oxidation.