Matrix effects in the neutralization of He ions at a metal surface containing oxygen

Surface Science 609 (2013) 167–171

Contents lists available at SciVerse ScienceDirect

Surface Science journal homepage: www.elsevier.com/locate/susc

Matrix effects in the neutralization of He ions at a metal surface containing oxygen Philipp Kürnsteiner a, Roland Steinberger a, Daniel Primetzhofer a, b, Dominik Goebl a, Thorsten Wagner a, Zdena Druckmüllerova a, c, Peter Zeppenfeld a, Peter Bauer a,⁎ a b c

Institute of Experimental Physics, Atomic Physics and Surface Science, Johannes Kepler University Linz, 4040 Linz, Austria Department of Physics and Astronomy, Ion Physics, Uppsala University, 751 20, Sweden Institute of Physical Engineering, Brno University of Technology, 61669 Brno, Czech Republic

a r t i c l e

i n f o

Article history: Received 25 September 2012 Accepted 6 December 2012 Available online 14 December 2012 Keywords: Low energy ion scattering LEIS Charge exchange processes Surface composition analysis Metal oxide

a b s t r a c t Charge exchange between He ions and a Ni(111) surface containing oxygen was studied by Low-Energy Ion Scattering, using 1.25 keV He+ as primary ions. The energy resolved yield of positive ions was detected after backscattering from Ni or O for different exposures of Ni(111) to molecular oxygen. Pronounced changes in the neutralization efficiency due to the presence of oxygen are observed for both, the adsorbate phase at low oxygen dose, and the NiO phase at high dose. The presence of O in the surface makes resonant charge transfer in a close collision possible. Evidence for a strong matrix effect is found: O in NiO neutralizes much more efficiently than O in the adsorbate phase. Independently, the different interaction stages of Ni–O and the surface structure were monitored by Photoelectron-Emission-Microscopy and Low-Energy Electron Diffraction. © 2012 Published by Elsevier B.V.

1. Introduction In the interaction of slow ions with solid surfaces or gas atoms, charge exchange has been a very interesting physical phenomenon being studied ever since the early days of ion physics [1–3]. Electron capture and loss are of importance for a fundamental understanding of ion–solid interaction [4], in astrophysics [5], and in plasma physics [6]. Last but not least, it is of great relevance in manifold applications successfully employed to study morphology and composition of surfaces or near-surface structures [7,8]. In these contexts neutralization of noble gas ions at metal surfaces has been investigated extensively in the last decades [9–11]. The good qualitative and quantitative understanding of the underlying processes has led to a wide field of applications for techniques like Low-Energy Ion Scattering (LEIS) or Secondary Ion Mass Spectrometry (SIMS). However, for both systems matrix effects have been reported, which means that the efficiency of active charge exchange processes is influenced by the chemical environment [12,13]. Oxygen is an abundant constituent on surfaces of technological or scientific interest, no matter whether as a contaminant or as an inherent part of the system of interest, and it is known to change ion fractions in SIMS drastically [14]. In fact, a change in the O concentration of less than a percent can increase ion yields by an order of magnitude [15]. In LEIS, only insufficient knowledge on the role of oxygen in charge exchange of slow light ions is available. This represents a striking lack of basic

⁎ Corresponding author. Tel.: +43 732 2468 8516. E-mail address: [email protected] (P. Bauer). 0039-6028/$ – see front matter © 2012 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.susc.2012.12.003

knowledge about charge exchange processes, and is a handicap for quantification in applications of LEIS to oxide surfaces [7,16]. Thus, a deeper understanding of the role of oxygen in ion neutralization at surfaces is not only of fundamental interest, but could also break the ground for an even wider field of application of the surface analytical techniques mentioned before. In the present study we investigated charge exchange of He+ ions with Ni and O atoms for both, O chemisorbed on Ni(111), and a NiO surface oxide on Ni(111). To this aim we applied LEIS, Photo-Electron Emission Spectrometry (PEEM) and Low-Energy Electron Diffraction (LEED). The Ni(111)–O system represents an ideal case for our study, since it has been extensively studied in the past (see e.g. [17–21] and references therein). Thus, a profound understanding of the mechanisms of chemisorption and surface oxidation is available today. This allows us to extract information on the role of O in ion neutralization by metal atoms, and on charge exchange between helium ions and oxygen in different chemical states. For He ions with a kinetic energy of several keV, the dominant charge exchange processes in the interaction with surface atoms are well known. For most chemical elements, only two different types of charge exchange processes prevail: Auger neutralization and resonant charge transfer in a close collision [22]. Auger neutralization (AN) is nonlocal, acts along the trajectory, is due to interaction of the projectile ground state with the valence-conduction band of the sample, and is governed by a transition rate ΓA. Via the rate equation for the ion fraction P + this leads to P + = exp(− vc/vz), where vc = ∫ΓAdz is a measure for the AN efficiency, and vz is the perpendicular component of the ion velocity [23]. If AN is the only process, P + represents a survival probability which is ≪ 1 even for scattering at the

168

P. Kürnsteiner et al. / Surface Science 609 (2013) 167–171

outermost surface atoms, leading to a superb surface sensitivity of ≈1 atomic layer, for which LEIS is known. Resonant processes refer to a situation where the He ground state level is resonant with occupied or unoccupied states of the conduction band: Only a small number of elements like Ge, In or Pb exhibit electrons with similar binding energies than the unperturbed He 1s level. For the prevailing part of the elements, resonant charge transfer is only possible in a close collision – at an interatomic distance, r, smaller than a critical value R0 – as a consequence of electron promotion of the projectile level due to interaction with atomic levels of the target atom. Resonant charge transfer may lead to (resonant) neutralization (RN) when He + enters the collision, or even to (re)ionization (RI) when He 0 is scattered [24,25]. For a fixed scattering angle, the probabilities for these processes, PRN and PRI, depend on the ion energy, and vanish for energies lower than a threshold energy Eth, which is characteristic for the specific choice of projectile and scattering center. For instance, for He and Ni, Eth = 1500 eV, while for He and O, a value of 700 eV has been reported for Eth [25–27]. For E >Eth, the ion fraction consists of two contributions: “survivals” and “reionized projectiles”. Survivals are projectiles that did not undergo any charge exchange process; reionized projectiles have first been neutralized and then reionized on their way out in a collision close to the surface. In this regime, the information depth may exceed one atomic layer, but for polycrystalline samples, excellent surface sensitivity can be maintained if the LEIS signal is properly evaluated [28]. For typical LEIS energies, an information depth of less than 2 or 3 atomic layers can be obtained, details depending on the investigated system. 2. Experiment We performed the experiments using the ESA-LEIS setup MINIMOBIS [29] at the Institute of Experimental Physics at the Johannes Kepler University Linz. In this setup a mass selected beam of He ions with primary energies in the range of 0.2–4 keV can be directed towards the sample surface in perpendicular incidence. Primary beam currents range from 0.05 to 2 nA. Ions that are backscattered by θ =136° pass a cylindrical mirror analyzer (CMA) with an azimuthal acceptance angle of 2π and are detected by micro channel plates (MCP). As a sample, a polished Ni(111) single crystal with a roughness below 0.03 μm and a precision of the orientation of ±0.1° was used. We prepared a clean crystalline surface by cycles of Ar+-sputtering and annealing (up to 700 °C). Surface crystallinity was checked by Low-Energy Electron Diffraction (LEED). After preparation, LEIS spectra did not show any surface impurities. To study the influence of oxygen on charge exchange, the sample was exposed to molecular oxygen at room temperature. LEIS spectra were recorded for 1.25 keV He+ ions for oxygen exposures in the range from 0.1 to 10,000 L. Photo-electron Emission Microscopy (PEEM) and LEED were used to study changes in electronic properties and surface structure; PEEM and LEED images were recorded for Ni(111) as a function of oxygen exposure, in another UHV-setup but with identical sample preparation. For the excitation of photoelectrons a D2 lamp with a MgF2 window was used. This lamp delivers photons up to 11 eV with a major emission line in the deep UV at 7.7 eV (160.8 nm). LEED pictures were taken at an electron energy of 83.6 eV.

Fig. 1. Normalized electron emission yield extracted from PEEM images during exposure of a Ni(111) surface to molecular oxygen. Also shown are LEED images recorded for selected oxygen doses.

LEED pattern recorded for an exposure of 2 L. For higher exposures, the PEEM intensity remains constant, whereas the LEED pattern continuously changes due to formation of more complex adsorbate structures [18]. For exposures above ~ 80 L, the PEEM intensity increases again, due to the onset of surface oxide formation, and a corresponding decrease in the work function. The final LEED pattern shows again a hexagonal surface structure, as expected for NiO on Ni(111). A comparison of the patterns for clean Ni(111) and NiO shows an increase by 15% in the size of the unit cell [32]. To summarize, all observations described so far are in accordance with literature, proving proper sample preparation. As a next step, the Ni(111)–O system was investigated by LEIS, using 1.25 keV He+ ions as projectiles. Spectra obtained for different O exposures are shown in Fig. 2. The peaks around 950 and 480 eV correspond to ions which were backscattered in a single collision from Ni or O (if present), respectively. The integral peak areas, i.e. the ion yields, can be evaluated independently for Ni and O, yielding YNi+ and YO +, respectively. From YNi + and the corresponding surface atom density of Ni(111), the ion fraction P + can be deduced, i.e. the percentage of positively charged helium ions when backscattered from a Ni atom in a Ni(111) surface. For the clean Ni(111) surface and 1.25 keV He ions, this results in P + = 0.18. A detailed description of the evaluation

3. Experimental results PEEM intensities averaged over a field of view of ~ 145 μm are plotted in Fig. 1 for oxygen exposures from 0.1 to 3000 L. The figure also includes characteristic LEED patterns recorded in the course of the experiment. These results permit to identify the individual stages of the interaction of Ni(111) with oxygen. At exposures of about 1 to 5 L, the decrease of the normalized intensity corresponds to the adsorption of O in a (2 × 2) superstructure, which is known to lead to an increase in work function from about 5.3 to 6.0 eV [30,31]. This interpretation is also in accordance with the

Fig. 2. Energy spectrum of He+ ions backscattered from Ni(111) exposed to O2 for different doses of O2 (see inset) and normalized for the same primary ion fluence. The primary ion energy was 1.25 keV.

P. Kürnsteiner et al. / Surface Science 609 (2013) 167–171

procedure is given elsewhere [33,34]. To concentrate on the relevant physical processes rather than on quantitative figures, we will use the normalized ion yields YNi + and YO + instead of P + in our discussion. YNi+ = 1 refers to the yield of ions backscattered from the clean Ni(111) surface; YO + = 0.25 refers to the yield of ions backscattered from oxygen in the (2× 2) adsorbate structure on the Ni(111) surface. In Fig. 3, YNi+ is presented as a function of oxygen exposure. Similarly as for the PEEM intensity, three different regimes can be identified. During formation of the (2× 2) superstructure, the Ni signal is decreasing. When an oxygen coverage ΘΟ = 0.25 is reached, the ion yield remains more or less constant for increasing oxygen exposure. At an exposure of ~60 L surface oxide growth starts and the ion signal further decreases. At exposures >150 L, the surface oxide layer is completed and the ion yield becomes constant again. However, after surface oxidation, the ion intensity is lower than for clean Ni(111) by almost one order of magnitude. Some specific data points are plotted as full red symbols to allow for an easier recognition of the corresponding data in later plots. Whereas YNi+ decreases monotonously as a function of O exposure, the normalized ion yield YO+ for projectiles scattered from O shows an interestingly different behavior. In Fig. 4, YO + is presented as a function of exposure. As shown in Fig. 4, at low exposures, i.e. during formation of the superstructure, YO + increases. During formation of the more complex adsorbate structure, YO+ slightly decreases. When surface oxide growth starts, YO + exhibits a strong decrease. For exposures above ~150 L, at which YNi + does not change anymore, the very same behavior is observed for YO+. After surface oxidation YO+ is lower than in the (2× 2) adsorbate structure by a factor of four. 4. Interpretation In order to obtain information on charge exchange, it is indispensible to assess the LEIS results, i.e. YNi+ and YO+. For this purpose we compare our YO+ results to the following models: first, the statistical growth model (Langmuir isotherm), and second, the growth model taking dissociative chemisorption into account [35]. In both cases we assume that charge exchange is not subject to matrix effects; in other words, when O adsorbs in the outermost atomic layer, YO+ does not depend on the total amount of oxygen present at the surface. Thus, the ion signal is expected to be a direct measure for the oxygen coverage ΘΟ, according to YO+ ~P+ · ΘΟ. For the investigated system, ΘΟ is expected to saturate at a value of 0.25 corresponding to a complete (2×2)O superstructure on top of Ni(111). The fits according to the two distinct growth models are shown in Fig. 4 as dashed green line and solid red, respectively. Both models are in good agreement with the experimental data. From

Fig. 3. Normalized ion yield YNi+ for 1.25 keV He+ scattered from Ni atoms in a Ni(111) surface as a function of exposure to O2.

169

Fig. 4. Normalized ion yield YO+ for 1.25 keV He+ scattered from O atoms as a function of O2 exposure of a Ni(111) surface. Open circles: experimental data; red line: Langmuir isotherm; and dashed green line: dissociative adsorption model [35].

the models it is possible to derive a value for the sticking coefficient for the impinging oxygen. For the Langmuir isotherm, a sticking coefficient of 0.3 is deduced. The dissociative adsorption model yields a sticking coefficient of 0.46. Thus, the values derived with both models agree well with the value of 0.23 found in a calorimetric experiment for oxygen chemisorption on Ni(111) [36]. From the fact that these models result in an ion yield YO + which is consistent with adsorption data, we conclude that in this coverage regime the ion fraction P + is a constant, i.e. no matrix effects are observed in the ion signal YO +. Based on this finding it is now possible to calculate an absolute ion fraction P + valid for scattering of He ions from O atoms in the (2 × 2)O superstructure. By use of an appropriate scattering potential [37] this yields a He + fraction of ~ 18% for the projectiles scattered from O in the (2 × 2) superstructure. For a binary compound, AxB1−x, a very sensitive method to check for matrix effects in ion neutralization is to plot the ion signal of constituent A, YA +, as a function of the ion signal of constituent B, YB + [7]. For constant values of PA+ and PB+, i.e. in the absence of matrix effects, a linear relationship YA +(YB+) is expected on the basis that Yi+ ~ Pi+ · Θi where i stands for A or B and ΘA + ΘB = 1. Fig. 5 shows a plot for YO + versus YNi+. In Fig. 5, a linear relationship between the normalized ion signals is indeed observed for YNi + ≥ 0.25 (see the dashed red line), even if in this case the surface is not a binary compound but consists of an adsorbate layer on top of the Ni(111) surface. Note that the signal obtained from Ni decreases faster than the O signal increases. In the experiment, the ion yield for Ni drops by a factor of 4 when the oxygen superstructure is formed, corresponding to 0 ≤ ΘO ≤ 0.25 with respect to the Ni lattice. This is an interesting observation, which indicates that the presence of oxygen leads to far more effective neutralization than expected from naïve geometrical arguments (assum+ ing ΘO = 0.25 and ΘNi = 0.75 one might expect YNi = 0.75). Since O atoms adsorb at three-fold hollow sites on top of Ni(111) they hardly influence trajectories of He ions scattered from Ni atoms [38]. Nevertheless, a significant increase in neutralization efficiency is expected for two reasons: first, adsorbed O atoms will generally increase the overall electron density in front of the surface, and thus increase Auger neutralization rates. Second, also the scattering geometry should be considered, which due to the exit angle of 44° with respect to the surface normal leads to a relatively close encounter with O atoms (below 2 a.u.) for the majority of He ions scattered from Ni. At distances between 1 and 2 a.u., however, Auger neutralization is

170

P. Kürnsteiner et al. / Surface Science 609 (2013) 167–171

Fig. 5. Ion signal YO+ obtained for backscattering from O surface atoms, as a function of the ion signal obtained YNi+ for backscattering from Ni surface atoms. Both ion signals are normalized to the surface density of Ni atoms in Ni(111).

known to be most effective [39], whereas no deflection of primary ions has to be expected. As can be seen from Fig. 6, three out of four Ni atoms in the p(2 × 2) cell are adjacent to adsorbed O atoms. When scattered from these three Ni atoms, He ions will experience increased neutralization probability when passing by adsorbate atoms. These arguments can be summarized in the simple conclusion that LEIS is much more sensitive to O atoms in the adsorbate layer than to Ni atoms in the atomic layer below. This is in good accordance with the claimed superior surface sensitivity of LEIS. Strictly speaking, this finding can be called a matrix effect, disregarding the changed surface geometry. At intermediate coverages, the ion yields for scattering from both, Ni and O atoms (corresponding to the ellipse in Fig. 5) exhibit

Oxygen

Nickel

Fig. 6. Schematic of the O p(2 × 2) phase on Ni(111). Small spheres correspond to oxygen atoms, large spheres correspond to Ni atoms. Ni atoms which are adjacent to O atoms are colored blue. The straight lines identify the p(2 × 2) cell on the Ni lattice.

plateaus. Based on LEED patterns and from previous results we can assume that mainly a rearrangement of oxygen atoms in the surface takes place. Apparently, the ion yields do not change drastically. In fact, literature suggests that in the adsorbate phase one monolayer of oxygen does not exceed ΘO = 0.29, with respect to the Ni lattice [18]. The minor changes observed for YO + may be attributed to the fact, that in a possible (√3 × √3)O superstructure the relative positions of oxygen atoms are changed and neutralization becomes even more efficient due to the influence of neighboring O atoms. In any case, the ion signal obtained for backscattering from O atoms at the surface is no longer a direct measure for the O surface concentration. This behavior represents a marked matrix effect in neutralization of He ions by oxygen. In the regime of O exposures > 60 L, where the onset of surface oxide growth is expected, strikingly both ion signals decrease with increasing exposure. For exposures > 60 L, the ion yield YO+ drops by a factor of ~ 3 and YNi + decreases by a factor of ~ 4. This is in contrast to naïve expectations since from the LEED pattern, it can be assumed, that the NiO layers grow as a film in (111)-orientation. Thus, in contrast to the situation in which a superstructure of O atoms is formed, the first bilayer of the surface oxide contains both, O and Ni atoms with equal abundance. Consequently, a bilayer of the surface oxide contains twice as much oxygen as the (2 × 2) adsorbate structure. Therefore, the simultaneous decrease of both ion signals is a clear matrix effect and may be traced back to changes in the electronic properties, i.e. in the band structure and in the electron density distribution. At the ion energy employed in the experiment, Auger neutralization (AN) is the only charge exchange mechanism active for clean Ni, i.e. there are no contributions from resonant charge exchange processes. In terms of AN it is obvious that chemisorbed oxygen will increase the AN efficiency, as a consequence of the fact that He + must pass by the adsorbed oxygen where it may be neutralized efficiently in an AN process. The presence of oxygen at the surface may, however, open the possibility for resonant processes (resonant neutralization, RN, and reionization, RI) in a close collision [7]. As outlined in [7], these resonant processes usually lead to efficient neutralization and allow for reionization of projectiles that penetrate into the sample, are backscattered in a deeper layer and reionized in a final collision close to the surface. As a consequence, a feature of reionization is a low energy background due to projectiles backscattered in deeper layers [28]. Such a reionization background has indeed been observed in our data in the adsorption regime in contrast to clean Ni, for which the background is very low and energy independent (see Fig. 2). From this we conclude that in this regime resonant charge transfer is possible in a collision with oxygen. Fig. 7 shows how the background (integrated between the single scattering peaks for O and Ni, i.e. 600 eV–800 eV) develops with increasing oxygen exposure. This energy range corresponds to trajectories which require backscattering from Ni, possibly in deeper layers; additional interactions with oxygen or nickel atoms may occur when leaving the surface. It can be clearly seen in Fig. 7 that the background exhibits significant variations, by up to a factor of 4 in the whole range of exposures. For exposures below 6 L, the increase of the background signal correlates remarkably well to the evolution of the oxygen yield (see Fig. 4). This emphasizes the importance of O for reionization of He projectiles. At exposures from ~ 6 L to ~ 100 L both, the background and YO + decrease and reach a plateau. Since in this regime a rather constant coverage of the surface by oxygen is maintained, this behavior may be traced back to changes in the electronic structure if the surface. The decrease in YO + and background corresponds either to a decrease in the reionization probability or to an increase in the neutralization efficiency (AN, RN). For coverages above 100 L, i.e. in the oxide phase, both YO + and + YNi decrease (Figs. 3 and 4) while the background increases. This regime differs from those discussed before in both, electronic structure

P. Kürnsteiner et al. / Surface Science 609 (2013) 167–171

171

Since the dynamics of the projectile ground state level is known to be strongly affected in the interaction with atoms, investigation of neutralization for other oxide materials and in a wide range of projectile energies will help to improve understanding of these band structure effects in charge exchange. At this stage of the investigations, also the demand for a well-founded theoretical description of the level dynamics in the scattering processes is obvious. On the one hand, the presented experiment exhibits the expected high surface sensitivity of LEIS. On the other hand, it proves that accurate quantification of surface composition may be a demanding task. In order to be successful, one has to know which is the chemical state of the surface and which are the relevant charge exchange mechanisms. Therefore, for successful quantification, support by other methods or detailed knowledge about the investigated system will be decisive. Acknowledgment Fig. 7. Integrated background signal in the energy range between 600 eV and 800 eV as a function of oxygen exposure. The spectra were normalized for ion fluence.

This work was supported by the Austrian Science Fund (FWF) under Contract. No. P20831.

and surface geometry. Therefore, even a qualitative discussion is difficult. Nevertheless, one may conclude that there neutralization and reionization due to oxygen are most efficient. In the following, the regime of high oxygen exposures is discussed. The formation of the surface oxide leads to major changes in the electronic structure: a wide valence band is formed extending to energies of ~ 15 eV below the Fermi level EF and the oxygen 2s level forms a pronounced peak in the electron density of states at ~ 20 eV below EF [40]. Resonant charge transfer involving these levels will be possible already at larger interaction distances, and therefore leads to very effective neutralization of He projectiles scattered from Ni or O atoms in NiO. The pronounced reionization background observed for scattering of He from NiO proves the strong increase in the relevance of resonant processes here (see Fig. 2). Thus, the observed reduction of both ion yields, YNi + and YO +, in the regime of surface oxide formation is rather due to the increased efficiency of resonant processes than due to a modification of the AN efficiency. Since it is known, that slow oxide growth continues at higher exposures (see also Fig. 1), it can be expected, that the film thickness finally will exceed the information depth of LEIS. The fact that ion fractions are low and ion yields obtained from the peaks in the spectra do not depend on the oxygen exposure points towards an information depth of 1 to 2 ML, in accordance with previous investigations [28].

References

5. Summary and conclusions The present investigation presents experimental evidence for a pronounced matrix effect in Auger neutralization of He ions at a Ni(111) surface containing oxygen in different chemical states. Charge exchange in the interaction of He ions with Ni and O atoms is very much different: while for Ni and adsorbed O Auger neutralization is the dominant process, for the surface oxide resonant charge exchange processes (RN and RI) prevail, leading to much lower ion yields than in the surface adsorbate phase. These matrix effects affect neutralization for ions backscattered from both, Ni and O surface atoms. The results support the assumption, that at least Auger neutralization of noble gas ions at solid surfaces may strongly be influenced by chemical interactions taking place at the surface. As expected for AN, changes in the density of states of the material may induce strong matrix effects in neutralization.

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]

E. Rutherford, Philos. Mag. 47 (1924) 277. S.K. Allison, Rev. Mod. Phys. 30 (1958) 1137. H. Niehus, W. Heiland, E. Taglauer, Surf. Sci. Rep. 17 (1993) 293. A. Arnau, M. Peñalba, P.M. Echenique, F. Flores, Nucl. Instrum. Methods B 69 (1992) 102. H. Balsinger, et al., Nature 321 (1986) 330. M. Lampe, V. Gavrishchaka, G. Ganguli, G. Joyce, Phys. Rev. Lett. 86 (2001) 5278. H.H. Brongersma, M. Draxler, M. de Ridder, P. Bauer, Surf. Sci. Rep. 62 (2007) 63. A. Benninghoven, Surf. Sci. 299 (1994) 246. J. Merino, N. Lorente, P. Pou, F. Flores, Phys. Rev. B 54 (1996) 10959. N.P. Wang, E.A. Garcia, R.C. Monreal, F. Flores, E.C. Goldberg, H.H. Brongersma, P. Bauer, Phys. Rev. A 64 (2001) 012901. D. Valdés, J.M. Blanco, V.A. Esaulov, R.C. Monreal, Phys. Rev. Lett. 97 (2006) 476011. D. Primetzhofer, S.N. Markin, E. Taglauer, P. Bauer, Phys. Rev. Lett. 100 (2008) 231201. H.W. Werner, Surf. Interface Anal. 2 (1980) 56. J.M. Eiler, C. Graham, J.W. Valley, Chem. Geol. 138 (1997) 221. T.L. Alford, et al., Fundamentals of Nanoscale Film Analysis, Springer, New York, 2007. H.H. Brongersma, P.A.C. Groenen, J.-P. Jacobs, Mater. Sci. Monogr. 81 (1995) 113. P.H. Holloway, J.B. Hudson, Surf. Sci. 43 (1974) 141. A.R. Kortan, R.L. Park, Phys. Rev. B 23 (1981) 6340. J.A. Slezak, B.D. Zion, S.J. Sibener, Surf. Sci. 442 (1999) L983. A. Eichler, F. Mittendorfer, J. Hafner, Phys. Rev. B 62 (2000) 4744. F. Besenbacher, Prog. Surf. Sci. 44 (1993) 5. H. Winter, Phys. Rep. 367 (2002) 387. H.D. Hagstrum, Phys. Rev. 96 (1954) 336. R. Souda, M. Aono, C. Oshima, S. Otani, Y. Ishizawa, Surf. Sci. 150 (1985) L59. T.M. Thomas, H. Neumann, A.W. Czanderna, J.R. Pitts, Surf. Sci. 175 (1986) L737. D. Primetzhofer, PhD thesis, JKU Linz. M. de Ridder, H.H. Brongersma, Private Communication, 2006 D. Primetzhofer, M. Spitz, E. Taglauer, P. Bauer, Surf. Sci. 605 (2011) 1913. R.J.M. Elfrink, Master thesis, Eindhoven University of Technology (1994). G.E. Becker, H.D. Hagstrum, Surf. Sci. 30 (1972) 505. S. Evans, J. Pielaszek, J.M. Thomas, Surf. Sci. 56 (1976) 644. L.C. Bartel, B. Morosin, Phys. Rev. B 3 (1971) 1039. S.N. Markin, D. Primetzhofer, P. Bauer, Nucl. Instrum. Methods B267 (2009) 634. S. Rund, D. Primetzhofer, S.N. Markin, D. Goebl, P. Bauer, Nucl. Instrum. Methods B269 (2011) 1171. H. Ibach, Surfaces and Interfaces of Solid Materials, 3rd edition Springer, 1995. J.T. Stuckless, C.E. Wartnaby, N. Al-Sarraf, J.B. Dixon-Warren, M. Kovar, D.A. King, J. Chem. Phys. 106 (1997) 2012. D.J. O'Connor, J.P. Biersack, Nucl. Instrum. Methods B 15 (1986) 14. D.T. Vu Grimsby, Y.K. Wu, K.A.R. Mitchell, Surf. Sci. 232 (1990) 51. D. Goebl, D. Valdés, E. Abad, R.C. Monreal, D. Primetzhofer, P. Bauer, Phys. Rev. B 84 (2011) 165428. T.M. Schuler, D.L. Ederer, S. Itza-Ortiz, G.T. Woods, T.A. Callcott, J.C. Woicik, Phys. Rev. B71 (2005) 115113.