Step-modified phase diagram for O/Ni(977)
 |
| Figure 1 |
Introducing steps to this system has significantly altered the phase diagram of chemisorbed oxygen. It is not surprising that steps would have such an effect, as related results have been documented for hydrogen chemisorbed on nickel. There are three noteworthy step-induced effects. First, unlike H/Ni(977), the transition temperature of the p(2´2)-O phase is unaffected, but a coexisting Ni[8(111)´(100)]-2(1d)-O structure not observed on Ni(111) is anchored by the steps (Figure 1).
In the limit of low coverage, less than 2% of a monolayer, oxygen atoms adsorb to the bottom of the Ni(977) step edges due to the favorable four-fold coordination available from the (100) step faces. For adsorption at room temperature in this low coverage limit, the oxygen adsorbed at the steps forms a previously unobserved ordered (n´2) structure exhibiting weak diffraction. With further exposure at room temperature, oxygen forms an ordered p(2´2) superstructure at a coverage of 0.25 monolayers. As the substrate is heated, this phase undergoes an order-disorder transition at 438-440 K both for the flat Ni(111) and stepped Ni(977) (Figure 3b). However, the step-localized (n´2) phase, also seen at lower coverage, begins to appear together with the p(2´2) around room temperature and persists well above the p(2´2) disordering transition temperature. The step-stabilized ordered overlayer remains locked in place until it is finally incorporated into the bulk around 565 K. This dissolution temperature is dependent on the oxygen dissolution history of a particular crystal. Lower temperatures for oxygen incorporation into nickel have been reported for crystals with less oxidation history. Note that the LEED intensities that are tracked in Figure 3b only hold true if the sample temperature is increased due to irreversible dissolution. If the sample is cooled to room temperature after reaching 565 K, where diffraction for the overlayer disappeared, the only visible spots are from the substrate terrace, indicating complete dissolution of surface oxygen into the bulk
| Figure 2 |
 |
The second major effect induced by the array of steps is the observation that the oxygen LEED spots become increasingly oblate with increasing temperature (Figure 2). Anisotropic diffraction of this nature suggests the existence of direction-dependent coherence lengths. We believe this to be evidence for ordered domains persisting locally in the vicinity of step edges above the terrace order-disorder transition temperature. Unfortunately, verifying this inhomogeneous disordering by analyzing the dimension-dependent coherence lengths for the adlayer is hindered by the steps themselves. There is streaking in the <211> direction associated with the stepped surface; the spot widths in this direction involve a convolution of domain coherence length and step diffraction. Further studies with an STM would be most difficult given the high temperatures and binding sites involved, but high energy helium diffraction may have the potential to resolve this issue.
| Figure 3 |
 |
| (a) |
 |
| (b) |
A third manifestation of step edge influence is a change in dissolution behavior. The generation of a step-registered structure after the 0.25 ML overlayer disorders and loses oxygen to dissolution suggests that there exist multiple dissolution stages. On Ni(111), oxygen will not dissolve below 500 K, but we observe a progressive dissolution into Ni(977) beginning even below 400 K (Figure 3a). These temperatures, however, are strongly influenced by the oxidation history of the crystal in terms of oxygen population of selvedge and bulk regions of the material. We believe that oxygen dissolves anisotropically: it dissolves into the terraces and remains near the step edges at higher temperatures. We propose a kinetic difference in dissolution of oxygen between terrace (111) and step (100) crystallographies resulting from the difference in bonding between four-fold and three-fold hollow sites found on the step and terrace faces of this surface, respectively. The O-Ni bond strength derived from heat of adsorption measurements is 5.39 eV for the (100) face and 4.87 eV for (111). Assuming a similar barrier height for dissolution, this difference produces a higher effective kinetic barrier for absorption via the steps. A Ni vicinal surface with identical geometry on the steps and terraces would not exhibit such two-step oxygen dissolution behavior.
We have studied the interaction between ordered overlayers of oxygen and a stepped Ni(977) surface
in order to isolate the effect of a regular array of steps on the adsorbate phase diagram. Our
results demonstrate that introducing steps produces three significant effects, namely: (i) adding
a previously unknown step-stabilized and -anchored ordered oxygen phase,
Ni[8(111)´(100)]-2(1d)-O, that is stable at much higher temperatures than
ordered phases on the flat surface; (ii) promoting spatially inhomogeneous disordering where oxygen
disorders in the center of the terraces before doing so in the proximity of the step edges; and
(iii) introducing multiple stages for oxygen dissolution into the bulk. All of these effects are
intimately related to the step-doubling and -resingling reconstructions that occur on this stepped
surface as a function of temperature and oxygen coverage. The stabilizing influence of the steps
on adsorbate overlayers provides a direct competition in the free energy balance that governs other
step phenomena such as step edge mobility and step reconstructions leading to faceting.
91. "Step-modified region of phase diagram for chemisorbed oxygen on nickel"
