The Nanostructuring of Atomically Flat Ru(0001) upon Oxidation and Reduction
The O/Ru(0001) system is widely studied due to its rich phase variety of various stoichiometry and atomic arrangements, including the formation of a RuO2/Ru(0001) oxide layer. Apart from homogeneous ruthenium surfaces in certain oxidation states, also strongly heterogeneous surfaces can exist due to oxidation state’s variation at the nanoscale. We report on a scanning tunneling microscopy (STM) study of the nanostructuring of the oxidized Ru(0001) surface as a result of its interaction with molecular oxygen at elevated temperatures and subsequent reduction of a resulting RuO2 film by CO or HCl molecules from the gas phase in high-vacuum environment.
KeywordsRuthenium Surface Oxidation Reduction Scanning tunneling microscopy
Scanning tunneling microscopy
X-ray photoelectron spectroscopy
Ruthenium is an important material in modern catalysis, as recently reviewed in , where the catalytically active phase has shown to consist of a surface layer of ruthenium dioxide. The latter is electronically conducting thus being employed as electrode material and contact material in super caps . Another interesting aspect of ruthenium is the ease of its (0001) crystal face to support two-dimensional (2D) materials in a laterally nanostructured state, as demonstrated for boron nitride , graphene , and their mixtures . These systems can act as nanotemplates for the growth of various catalytically active metal nanoparticles [5, 6], while the preparation of the 2D/Ru(0001) nanotemplate itself is strongly dependent on the previous preparation of the ruthenium sample in molecular oxygen at elevated temperatures. This “technological” step drives carbon impurities (always present in bulk Ru) out of the subsurface region toward the surface thus strongly affecting the resulting stoichiometry within the 2D BxNyCz mixture on top of Ru(0001).
The oxygen treatment of Ru(0001) usually starts with the ordered oxygen overlayer on top of the metal surface without oxygen penetration below the surface. As the oxygen coverage increases, the adsorbed O atoms form a 2 × 2 superstructure at 0.25 ML (monolayer) and further a 2 × 1 superstructure at 0.5 ML [7, 8]. Toward 1 ML coverage, there appears a (2 × 2)-3O [9, 10] and finally a simple (1 × 1)-O overlayer . At even higher oxygen exposures and elevated temperatures of around 650 K, the Ru(0001) surface reveals a self-limiting growth of a 1.6-nm thick heteroepitaxial RuO2(110) film .
The RuO2(110)/Ru(0001) system was investigated by means of scanning tunneling microscopy (STM) and ab initio density functional theory (DFT) calculations enabling its exact atomic structure clarification . Such kind of surface exposes parallel rows of Ru atoms and so-called “bridge” O atoms. The RuO2(110) surface is an efficient catalyst for CO oxidation even at temperatures as low as 300 K . There is evidence that the RuO2(110)/Ru(0001) system can be laterally inhomogeneous, as observed for instance by low-energy electron microscopy (LEEM) in the form of (1 × 1)-O in coexistence with RuO2(100) and RuO2(110) [15, 16]. However, due to insufficient lateral resolution of the LEEM technique, the nanoscale structure of the RuO2(110) could not be fully resolved. Therefore, the STM technique was applied to study the strongly inhomogeneous RuO2(110) films, such as at initial nucleation stage [17, 18, 19] or after partial thermal decomposition of the oxide .
Despite some nanoscopic insights into the RuO2(110) growth on Ru(0001) already documented in the literature, our view on this process is still far from being complete. There is a certain knowledge gap at one of the most interesting stages, roughly halfway in between the nucleation onset and the final continuous state of the oxide film. Clearly, at this stage, a pronounced nanostructuring of the surface can be observed without resorting to any kind of time-consuming nanolithography techniques thus being potentially interesting for a range of applications, e.g., nanocatalysis, nanobiotechnology, and nanoelectronics. In the present study, we report on the nanostructuring of the RuO2(110)/Ru(0001) system and its extensive characterization employing STM and X-ray photoelectron spectroscopy (XPS). In the following, we have chosen a rather narrow oxidation temperature range of 600–650 K, as a trade-off between requirements not to drive the system into a spatially homogeneous equilibrium state and to maintain a high enough mobility of the species for effective oxidation.
All experiments were conducted in a home-built ultra-high vacuum (UHV) multi-chamber system with a background pressure in the lower 10−10 mbar range. The initial preparation of the single crystal Ru(0001) surface (sample size 5 mm × 5 mm × 5 mm, delivered by Mateck) always included sputtering with 1.5 keV Ar+ ions (Ar purity of 99.999%, delivered by Linde), the sample being held at 1100 K for continuous healing of the ion beam-induced damages of the crystalline structure of ruthenium. An undesired side effect of ion sputtering is the departure from the uniformly smooth surface if viewed at the micrometer scale. Then, the general specifics of the sample are the coexistence of extended atomically flat terraces with abrupt height changes of several dozen nanometers (essentially step-bunching areas). A previous photoelectron spectromicroscopy investigation has demonstrated a substantially faster oxidation kinetics of such step-bunching areas in comparison with atomically flat ones , later directly confirmed on the atomic scale by the STM technique .
The ion sputtering was followed by exposing the surface to molecular oxygen (purity 99.999%, delivered by Linde) at pressures in the 10−8–10−7 mbar range for several dozen minutes and the sample temperature kept at ~1100 K. Such a treatment typically removed carbon contamination from the near-surface region of the sample: carbon segregates from the bulk/selvedge region of the sample to the surface and is subsequently converted into gas phase CO or CO2 molecules. The RuO2 growth (oxidation) was performed during several dozen minutes at elevated sample temperatures (600–650 K) in molecular oxygen at constant pressure in the 10−5–10−4 mbar range. Afterwards, RuO2 can partially be reduced by interaction with reducing molecules such as CO (purity 99.999%, delivered by Linde) or HCl (purity 99.9%, delivered by Linde) at typical pressures of 10−7–10−8 mbar.
The surface analysis was performed using the XPS setup delivered by PSP Vacuum Technology. It features a dual X-ray source (Mg Kα radiation hν = 1253.6 eV, 200 W power) together with a hemispherical analyzer (20 eV constant pass energy). The surface topography of the samples after various treatments described above was investigated in situ by means of STM in constant current mode (VT-STM, delivered by Omicron). All measurements were performed at the background pressure in the UHV range and always after the sample has cooled to room temperature. We have used metallic probe tips hand cut from the Pt80%Ir20% Ø 0.25 mm wire (purity 99.9%, delivered by Sigma Aldrich).
Results and Discussion
Figure 1c presents an STM image of the 860 nm × 860 nm surface area of the partially oxidized sample with RuO2(110) present on the periphery and non-oxidized atomically flat Ru(0001) in the center (parallelogram-shaped patch). This morphology is in line with the corresponding LEEM results, showing the growth of RuO2(110) domains with three distinct orientations along the high symmetry directions of the Ru(0001) surface [15, 16]. The parallelograms can arise when the surface is not yet completely covered by the oxide and four domains of two different orientations (in alternating order) form a closed contour around an internal metallic patch.
A cross-section along the black dashed line in Fig. 1c is given in the lower left part of the figure showing the height jump of ~0.6 nm between the oxide and the metal. This value is not identical to the oxide film thickness because the oxygen penetration into the ruthenium during the growth process will shift the oxide/metal interface below the unreacted metal surface. Another interesting feature, that can be deduced from this cross-section is that the basements of the holes within the oxide film often reach down to the non-oxidized Ru(0001). This means that such holes are essentially nanoscale perforations of the RuO2(110) film; thus, the overall surface is strongly heterogeneous (metallic vs oxidic) both at the micro and the nanoscale.
When the oxide film appears on top of the substrate (Fig. 2b), it has nanoscale perforations, as already discussed in Fig. 1b, c. As the amount of oxide increases, the perforations become smaller in lateral dimensions (Fig. 2b–d). At this growth stage, the film is essentially nanostructured in the form of apertures up to ~10 nm in lateral size and up to ~1 nm in depth. Finally, the perforations vanish, and the film becomes continuous (Fig. 2e). At the final growth stage, there is a smoothening of the surface, leading to the formation of extended atomically flat RuO2(110) terraces (Fig. 2f), and the film is no longer nanostructured. The corresponding atomic scale STM image is presented in the inset in Fig. 2f. Visible here are rows of the so-called bridging O atoms, which are the topmost atoms of the given stoichiometric RuO2(110) surface . These parallel rows run along the  direction at a distance of 0.64 nm and are clearly resolved in the image. The neighboring O atoms within a row are only 0.31 nm apart and are barely resolved as the atomic corrugation along the rows is roughly at the noise level of our measurement.
In Fig. 3c, the initial state is non-continuous (nanoperforated film), and the result of a similar exposure (10 L CO at 600 K) is shown in Fig. 3d. Now, the major effect is lateral widening of the holes (some of them now reaching several dozen nanometers); in other words, we observe lateral etching, which proceeds isotropically in the surface plane. Much higher exposure (100 L CO at 600 K) had an effect of complete oxide film reduction, as demonstrated by the sequence from Fig. 3e to Fig. 3f. The latter STM image shows the surface with numerous single atomic layer islands and pits of several nanometers in their lateral size, observed in the right part of Fig. 3f. However, in the left part of Fig. 3f, we observe the same morphology and structure as in Fig. 2a, indicating the presence of the oxide-free patch prior to reduction. It seems like a natural explanation that the exact area imaged in Fig. 3f consists of both an oxide-covered patch (right) and oxide-free patch (left) prior to reduction. Again, the pairs of black arrows express the correspondence between these two types of areas before and after the reaction with CO. The border between these two patches is a section of the parallelogram’s perimeter (of the one similar as in Fig. 1c or the one partially imaged in Fig. 3e). If viewed on atomic scale, both types of areas in Fig. 3f show the same 2 × 2 superstructure as the inset in Fig. 2a.
In the middle group of curves, a spectrum code-named as “partially oxidized” was obtained on the sample shown in Fig. 2b. Here, the O 1 s emission is present in the spectrum at the level typical for the non-continuous (nanoperforated) RuO2(110) film. The sample with a continuous atomically flat oxide film (shown in Fig. 2f) produced a photoelectron spectrum shown by the upper curve in Fig. 4, designated as “fully oxidized”. Its hallmark is the O 1 s emission of the intensity higher than the Ru 3p peaks and comparable with those from the Ru 3d core levels.
Further, we have acquired the XPS spectrum after partial reduction of the oxide film by CO, designated as “partially reduced by CO” in the middle group of curves in Fig. 4 and corresponding to the surface depicted in Fig. 3d. Here, the O 1 s emission has a similar intensity as that from the as-grown “partially oxidized” sample, consistent with the incomplete coverage of the surface by the oxide in the corresponding STM images. Finally, the spectrum obtained after complete reduction of the oxide in CO can be found in the lower group of curves (marked “fully reduced by CO”). Notably, this spectrum is basically identical to the “initial substrate” in the lower group of curves. It was obtained on the surface shown in Fig. 3f thus completing the full circle: metallic surface–metal partially covered by oxide–metal fully covered by oxide–metal partially covered by oxide-metallic surface. Under the “metallic surface,” we mean the Ru(0001) surface with chemisorbed 0.25 ML of oxygen forming the 2 × 2 superstructure. The return to this surface condition is evidenced by STM (Fig. 3f) and identical O 1 s emissions in the lower group of curves in Fig. 4. Overall, we observe an excellent correlation between the intensity of the O 1 s emission and the amount of oxide-related morphological features on the surface of our Ru(0001) sample.
Finally, we compare the above described reduction process induced by CO with that caused by interaction with HCl molecules. For this purpose, we have exposed the sample shown in Fig. 2e to 10 L (1 × 10−7 mbar for 100 s) HCl at substrate temperature of 700 K. The corresponding XPS spectrum is marked as “partially reduced by HCl” in the middle group of curves in Fig. 4. It is worth noting that no Cl 2p emission is visible in the spectrum after partial reduction by HCl. A separate analysis has indicated that this emission is actually present but its intensity is very small. Thus, we can neglect the presence of chlorine on the surface, which stands in line with its release into the gas phase during the RuO2 reduction above 650 K .
In conclusion, we have identified by means of STM investigation a variety of Ru(0001) oxidation and RuO2(110)/Ru(0001) reduction processes leading to nanostructured (and/or heterogeneous) surfaces when started from atomically smooth morphology of the chemically homogeneous sample. The oxide growth at 650 K in O2 proceeds through the nanostructured stage later reaching an atomically flat surface of the saturated RuO2(110) film. The nanostructuring appears in the form of irregular nanoscale perforations within the oxide layer, while their lateral sizes can be further manipulated (increased) by partial oxide reduction in CO (at 600 K) or HCl (at 700 K) gaseous environment. Our preliminary investigations demonstrate the “etching” of RuO2 by HCl, which proceeds anisotropically in the surface plane, as opposed to isotropic etching by CO.
Both authors have contributed equally to the present work. Both authors read and approved the final manuscript.
The authors declare that they have no competing interests.
- 9.Kim YD, Wendt S, Schwegmann S, Over H, Ertl G (1998) Structural analysis of the pure and cesiated Ru(0001)–(2 × 2)-3O phase. Surf Sci 418:267–272Google Scholar
- 14.Madhavaram H, Idriss H, Wendt S, Kim YD, Knapp M, Over H, Aßmann J, Loffler E, Muhler M (2001) Oxidation reactions over RuO2: a comparative study of the reactivity of the (110) single crystal and polycrystalline surfaces. J Catal. doi: 10.1006/jcat.2001.3281
- 24.Zweidinger S, Hofmann JP, Balmes O, Lundgren E, Over H (2010) In situ studies of the oxidation of HCl over RuO2 model catalysts: stability and reactivity. J Catal 272:169–175Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.