Growth and electronic properties of ultra-thin Ag films on Ni(1 1 1)

Surface Science 603 (2009) 125–130

Contents lists available at ScienceDirect

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

Growth and electronic properties of ultra-thin Ag films on Ni(1 1 1) Vesna Mikšic´ Trontl a,b, Petar Pervan a,*, Milorad Milun a a b

Institute of Physics, P.O. Box 304, Bijenicka 46, HR-10000 Zagreb, Croatia Faculty of Electrical Engineering and Computing, Unska 3, HR-10000 Zagreb, Croatia

a r t i c l e

i n f o

Article history: Received 6 May 2008 Accepted for publication 28 October 2008 Available online 13 November 2008 Keywords: Ultra-thin films Epitaxy Low index single crystal surfaces Silver Nickel Angle resolved photoemission Scanning tunnelling microscopy Quantum wells

a b s t r a c t We studied the growth mode and electronic properties of ultra-thin silver films deposited on Ni(1 1 1) surface by means of scanning tunnelling microscopy (STM) and angle resolved photoemission spectroscopy (ARPES). The formation of the 4d-quantum well states (QWS) was analysed within the phase accumulation model (PAM). The electronic structure of the 1 ML film is consistent with the silver layer which very weakly interacts with the supporting surface. The line-shape analysis of Ag-4dxz,yz QWS spectrum support the notion of strong localization of these states within the silver layer. The asymmetry of the photoemission peaks implies that the decay of the photo-hole appears to be influenced by the dynamics of the electrons in the supporting surface. Ó 2008 Elsevier B.V. All rights reserved.

1. Introduction Detailed understanding of structural and electronic properties of ultra-thin metallic films is important from both fundamental and technological point of view. A vast number of metal-on-metal systems has been studied with the aim to learn how different combinations of constituents exhibit novel properties. Particularly attractive are low dimensional systems which exhibit a strong size dependence of their electronic structure. Such systems are ultrathin metallic films deposited on well-defined metallic or semiconductor surfaces where the electronic structure perpendicular to the surface, through the formation of stationary states is strongly quantized [1,2]. The discretization of the electronic structure is usually associated with discrete variation of a wide range of physical properties: density of states [3], work function [4], electron– phonon coupling [5–8], etc. The attractiveness of ultra-thin layers is in the fact that supported structures of reduced dimensionality (quantum wells) provide a broad ground for manipulation of geometrical and electronic properties. There is a large difference (16%) between lattice constants of Ag and Ni and therefore no pseudomorphic growth of Ag on Ni(1 1 1) may be expected. LEED studies have shown epitaxial silver growth at room temperature [9] with the lattice constant equivalent to the bulk silver value. In some studies a small rotation of the layer main axes away from the substrate main axes was found [10,11]. Within * Corresponding author. Tel.: +385 1 4698888; fax: +385 1 4698844. E-mail address: [email protected] (P. Pervan). 0039-6028/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.susc.2008.10.044

the first layer silver forms 2D islands. Shapiro et al. [12] have concluded that the second Ag layer starts to grow when the coverage of 0.7 ML is reached. Mroz and Jankowski [9] have found that the first layer was completed at a nominal coverage of 1.4 ML while the saturation of the first layer is followed with the growth of the islands in the third layer. The only published STM study of this system [13] was focused on the formation of the first monolayer and the formation of silver nanoclusters on top of it. This study shows that the first Ag atomic layer grows by coalescence of large 2D islands. These islands are attached to the step edges and have irregular shape. They do not cross the terrace edge and have a moiré structure that fits a model of a regular Ag(1 1 1) layer positioned over the Ni(1 1 1) layer. It has been shown that the Ni(1 1 1) surface supports formation of QW states of s-p and d-symmetry in epitaxially grown ultra-thin silver films [14]. An early study of Miller et al. [15] was focused on interface effect on formation of sp-derived quantum well (QW) resonances in the 14-ML-thick film. More recent work of Senkovskiy et al. [16] explored a range of silver films proving the existence of the s-p QW states. The states were successfully modelled using an extended phase accumulation model. Quantum well states of s-p and d-symmetry in silver films were used by Varykhalov et al. [14] to probe the ground state electronic structure of the Ni(1 1 1) surface. The effective energy band gap of Ni(1 1 1) was determined in the energy range between 2.6 and 4.8 eV. In this work we use STM, and high-energy and high-angle resolution ultraviolet photoemission spectroscopy to study ultra-thin silver films formed on the Ni(1 1 1) surface, in particular the quan-

126

V.M. Trontl et al. / Surface Science 603 (2009) 125–130

tization of the Ag-4d electron system. We confirm the STM results for the first monolayer from Ref. [13] and present new photoemission data from ultra-thin Ag films. We have studied in detail electronic properties of the 1 ML silver film and have shown that 4d electrons within silver film very weakly interact with the supporting surface as if they are decoupled from the nickel substrate, behaving as a non-interacting 2D system. 2. Experimental The experiments were carried out in an ultra high vacuum chamber with base pressure in the range of low 108 Pa. The chamber was equipped with the scanning tunnelling microscope (STM), Specs He discharge lamp, and Scienta SES-100 hemispherical analyzer which simultaneously collects photoelectrons at a range of energies and angles (12° or less). The ultimate instrumental resolution of the analyzer is around 5 meV while in these experiments the total energy resolution was 25 meV. The angular resolution was 0.2°. The photon excitation energy used in these experiments was 21.2 eV. The STM experiments were performed in the constant-current mode. A monocrystalline Ni disc, (1 1 1) oriented and cut, was mechanically and electrochemically cleaned and polished prior to insertion into the vacuum chamber. The initial cycles of 3 keV Ar+ sputtering and annealing at 925 K removed most of contaminations with exception of small amount of carbon, oxygen and sulphur. Further cycles removed carbon and oxygen. In order to remove sulphur the annealing was performed at 850 K and the cycles of sputtering and annealing were performed several days. Eventually a clean and well-ordered Ni(1 1 1) surface (as checked by LEED and STM) was formed. Silver films were prepared by resistive heating of tungsten basket filled with pure silver while the substrate was at room temperature. Prior to the silver deposition the cleanness and order of the Ni(1 1 1) substrate was checked by photoemission spectroscopy and the quality of the LEED pattern.

3. Experimental results 3.1. STM In Fig. 1 an overview of the room temperature growth of Ag on Ni(1 1 1) substrate is given by typical constant-current STM images recorded for different coverages. The STM image of the clean nickel surface shows large terraces divided by monoatomic steps (see Fig. 1a). The obtained results nicely agree with the previous measurements [17]. The scan taken along the line indicated in (a) reveals the step height of 1.9 ± 0.1 Å (Fig. 1b) which is the value reported also in Refs. [17,18]. We have used the reported interatomic distance on the Ni(1 1 1) surface [17] of 2.49 Å to calibrate lateral distances in STM topographs throughout this work. Silver grows in the form of a number of islands starting at terraces and step edges. With the increasing coverage the growth of silver film goes dominantly through the increase of the size of islands at the step edges. The vacancies within the islands can be found up to the saturation of the first silver layer. An example is given in Fig. 1c for a submonolayer coverage. The height of monoatomic Ag islands, measured from the line scans across the island edge, is found to be 2.9 ± 0.1 Å (Fig. 1d). The Ag islands appear 1 Å higher than Ni(1 1 1) monoatomic steps and can easily be distinguished. At the nearly saturation coverage most of the surface is covered with silver, forming a homogenous Ag monolayer with the fuzzy step edges (Fig. 1e). Occasionally, a small vacancy islands in a monolayer are observed, as well as the small fraction of a second

Ag layer that emerges as an extension of the first layer across a step edge. The atomically resolved STM image in Fig. 1f shows hexagonal moiré superstructure that is a result of 16% mismatch between the Ag(1 1 1)-film and Ni(1 1 1)-crystal lattices [13]. The interatomic distance of 2.87 ± 0.05 Å is determined from the images. The linescan analysis of the patterns shows a buckling of 0.40 ± 0.05 Å. The buckling periodicity is 17.40 ± 0.05 Å which fits the periodicity of 6 Ag atoms along 7 Ni atoms. The same periodicity was confirmed by LEED measurements (not shown here). At coverage above 1 ML the system shows more complex growth pattern. At temperatures around 200 K silver was reported to grow in a layer-by layer mode [9]. At the room temperature we find mixed thicknesses, as shown in Fig. 1g which presents STM image of a nominally 2 ML coverage silver film. The second layer grows in a similar way as the first one starting at the step edge but continuing to grow over the terrace edge and over the underlying silver layer. Such a growth mode results in a film where the third layer grows simultaneously with the second layer as indicated by STM image in Fig. 1g. The measured height of the second Ag layer with the respect to the first one is 2.3 ± 0.1 Å (Fig. 1h). The same value is found for the height of the third layer with respect to second one. 3.2. ARPES 3.2.1. Multilayer films Fig. 2 shows the normal emission spectra of the Ni(1 1 1) surface covered with different silver films taken with HeI (21.2 eV) photon excitation. The spectrum at the bottom of Fig. 2a is obtained from the clean Ni(1 1 1) surface. The spectrum is characterised by strong Ni 3d maxima just below the Fermi level and a weak satellite at 6 eV. The photoemission spectrum of the 1 ML Ag film shows additional two peak structure between 4 and 5 eV. In principle, each silver monolayer should contribute five peaks, associated with five spin-independent Ag-4d states containing 10 electrons, to the photoemission spectrum (see Fig. 1 in Ref. [19]). However, at the excitation photon energy of 21.2 eV, used in this work, we observe only two peaks while the emission from other states is suppressed. The absence of the emission from other states can be attributed to the variation of the excitation-cross-section with photon energy [20,21] and light polarization [22] which significantly affect the photoemission intensity from the Ag-4d bands. Fig. 2b shows spectra of submonolayer and monolayer silver coverage. The constant binding energy for both states was determined for all submonolayer coverages. Additional atomic layer brings in additional 5 (spin-independent) d-states which further complicates the corresponding photoemission spectrum. The 2-ML normal emission spectrum is characterised by the set of new states labelled C–I. Besides two distinctive peaks at high binding energy side (H and I at energies 5.5 eV and 6.0 eV, respectively) there is also a characteristic shift of the leading peak towards the Fermi level (in Fig. 2a peak C with respect to peak A). The leading peak (J) is even more pronounced for the 3-ML spectrum. This contrasts with the situation of QW states of s-p symmetry which can generally be easily associated with a particular film thickness [1,2]. The 4d bands produce, due to their weak dispersion, a large number of states within a narrow energy window prohibiting thus the possibility to distinguish them already for films thicker than 2 ML. The leading peak energy could be used as a fingerprint of a specific film thickness when dealing with layer-by-layer-grown uniform films [20]. Although in the case of the Ag/Ni(1 1 1) system we find multilayer growth at coverages exceeding 1 ML the d-QW states can be effectively used to detect when the first silver layer is saturated and when the second starts

V.M. Trontl et al. / Surface Science 603 (2009) 125–130

127

Fig. 1. (a) The STM image of a 80 nm  80 nm large patch of the of clean, well-ordered Ni(1 1 1) surface showing two monoatomic steps (UB = 0.4 V, IT = 1 nA). (b) The scan profile along the line indicated in (a). (c) The STM image of submonolayer silver coverage (UB = 0.45 V, IT = 0.8 nA). (d) The scan profile along the line indicated in (c). (e) The STM image of well-ordered 1 ML silver film showing four monoatomic steps (UB = 0.4 V, IT = 0.9 nA). (f) Atomically resolved STM image of monolayer coverage showing moiré structure. (g) STM image of 2 ML nominal coverage (UB = 0.4 V, IT = 1 nA). The numbers indicate real film thickness. (h) The scan profile along the line indicated in (g).

Fig. 2. (a) The normal emission photoelectron spectra of silver films of the nominal coverage 0–7 ML on the Ni(1 1 1) surface. The vertical lines denote the energies at which direct transition take place from bulk Ag(1 1 1) for 21.2 eV photon energy. (b) The normal emission photoelectron spectra of silver films in the submonolayer coverage range.

to grow (see the appearance of peak C in Fig 2a). With increasing film thickness the photoemission spectrum starts to be dominated by peak K. There is no significant spectral change between films of 4 ML and 7 ML nominal coverage; they show four distinctive spectral maxima L, K, M and N at 4.4 eV, 4.9 eV, 5.6 eV and 6 eV, respectively. At a nominal coverage of 7 ML we could detect first traces of the Ag(1 1 1) surface state just below the Fermi level. The appearance of the surface state does not depend on the annealing temperature of the silver film. 3.2.2. Monolayer film As it has been shown in Fig. 2 the valence band spectrum of the 1 ML silver film is dominated by the two peaks associated with silver film and the nickel valence band two peak feature, just below

the Fermi level. The 4d-derived QW states were found at 4.42 eV and 4.67 eV for all submonolayer and 1 ML coverage. The same binding energies were determined using NeI (16.46 eV) excitation (not shown here) suggesting that silver features are truly twodimensional. Fig. 3a shows photoemission intensity map of the valence bands of 1 ML Ag on Ni(1 1 1) measured in a wide polar angle range along the C  M high symmetry line of the surface Brillouin zone. Fig. 3b shows photoemission spectra taken for selected polar angles in Fig. 3a. All spectral features in the energy range between 0 and 2 eV are associated with nickel 3d band. Around the BZ centre (H = 0°) the dominant intensity comes from the silver bands a, b (A and B quantum well states in Fig. 2) that disperse negatively. The spectral intensity of state b decays with increasing polar angle. On the other hand, band a can be followed from the BZ centre

128

V.M. Trontl et al. / Surface Science 603 (2009) 125–130

Fig. 3. (a) The photoemission intensity map (polar angle vs. binding energy) of 1 ML of Ag on Ni(1 1 1) measured along C  M high symmetry line. (b) The energy distribution curves taken from the map in (a) for the polar angles indicated by vertical bars.

(4.42 eV) up to the BZ edge at H = 40° (4.78 eV). However, at polar angles between H = 15–20° there is a clear discontinuity in the dispersion and for that reason the band is marked with a0 . Just in the same range of polar angles band c emerges. This band disperses positively up to the BZ edge reaching the binding energy of 3.83 eV. Fig. 3a shows rather faint contribution from bands labelled with d and e. These bands are clearly resolved in Fig. 3b. The binding energies of bands d and e at the BZ edge are 5.15 and 6.22 eV, respectively. 4. Discussion Ours as well as some previous measurements [9] have shown that silver grows on Ni(1 1 1) surface following Stranski–Krastanov growth mode: on top of saturated silver monolayer a multilayer growth takes place. A silver film grown on Ni(1 1 1) surface has the structure of the Ag(1 1 1) surface. However, due to the large lattice mismatch (16%) between these two surfaces, the Ag film shows buckled moiré superstructure [13]. The STM measurements show the buckling amplitude of 0.40 ± 0.05 Å that is somewhat bigger than is obtained from the hard-sphere model. This difference might arise from the silver layer relaxation, yet it could be a pure electronic effect. The interatomic distance within the silver layer is 2.87 ± 0.05 Å which is the bulk value. It appears that the large lattice mismatch between silver and nickel unite cells does not induce significant distortion of the silver lattice compared to the bulk values. STM of submonolayer coverages shows that the saturation of the first layer goes through the enlargement of the existing silver islands. The photoemission results (Fig 2b) show constant binding energy of the 4d QW states for all submonolayer coverages which is consistent with the picture of silver islands which change the size but not the structure. The electronic structure of Ni(1 1 1) is favourable for the formation of quantum well states in the silver films of both s-p and d symmetry [14]. At the SBZ centre the projection of the valence band states to (1 1 1) surface shows that the energy range between the Fermi level and 2.6 eV is dominated by the nickel bands of dsymmetry. This electronic structure can support the formation of s-p QW states in silver films. In the film thickness range examined in this work the s-p QW states have indeed been observed but with significantly higher photon energies than used in this experiment [14]. The projected energy band gap on Ni(1 1 1) depends on the spin [23]. However, we assume that the effective band gap is between 2.6–4.8 eV as experimentally determined by Varykhalov et al.

[14]. For higher binding energies up to 9 eV only electronic states of s-p symmetry exist at the BZ centre. The whole energy range between 2.6 and 9 eV is supposed to be efficient localizing barrier for the silver 4d electrons. It is known that electronic structure of the s-p symmetry of the supporting surface can act as a localizing potential (symmetry gap), as efficiently as the energy gap for all delectrons in the overlayer film [20] except the ones having dz2 symmetry. These may couple to the substrate bands and form QW resonances [19]. The evolution of the Ag 4d-QW states as a function of the film thickness may be modelled by the phase accumulation model [24] using the phase equation

UB þ UC þ 2kðEÞNa ¼ 2pn

ð1Þ

where UB and UC are the energy dependent phase shifts of the wave function at the vacuum side and substrate-film interface, respectively [24]. k(E) describes the energy dependence of the wave vector of electron propagating in silver overlayer film. Eq. (1) is numerically solved (Fig. 4) for the films of five thicknesses and four wave vectors (k1–k4, out of five) associated with four Ag-4d bands dispersing in the energy range between 4.2 and 6 eV [25] (the fifth

Fig. 4. Graphical solution of the phase equation for d-band QW states in silver films on Ni(1 1 1) surface. Each symbol denotes QW states formed in a film of a particular thickness: (d) 1 ML, (s) 2 ML, (j) 3 ML, (h) 4 ML, (N) 5 ML.

V.M. Trontl et al. / Surface Science 603 (2009) 125–130

band at 7 eV does not contribute to the QW states). The filled circles show four possible solutions of the equation for N = 1 (1 ML film). As pointed out earlier, only two QW states (A and B) are observed in the photoemission spectra (Fig. 2b). There are several general conclusions that can be drawn from the solution of Eq. (1). Fig. 4 clearly shows the initial shift of the leading QW state to lower binding energy with increasing film thickness (QW states A, C). From the figure one can see that no further shift is expected for thicker films. The major reason for the absence of the more pronounced shift to lower binding energies as in the case of Ag films on V(1 0 0) [20] or Fe(1 0 0) [33] surface, is relatively weak dispersion of the K4,5 bulk band which generates QW states at the low energy side of the spectrum. Due to the dispersionless nature of k3 and k4 it is obvious from Fig. 4 why QW states H and I virtually do not shift with increasing film thickness. The analysis shows that the solutions of Eq. (1) for 2 ML film (open circles) can not account for all photoemission maxima seen in the 2 ML spectrum (Fig. 2 a). Indeed, the QW states from third layer are needed to explain peaks F and G. This is a spectroscopic indication of what we know from the STM measurements – above 1 ML silver coverage the film develops following the Stranski–Krastanov growth mode. Of course, the reason for this comparison being a problem is not only the large number of possible states, but also the fact that the growth mode does not allow isolation of the 2 ML spectrum. From Fig. 2a one can see that 4 and 7 ML spectra show maxima close or at energies corresponding to the direct transitions from the bulk Ag(1 1 1) surface, when excited with 21.2 eV (indicated by vertical lines) [25,26]. Similar bulk like spectra are observed from Ag multilayers on some other metal surfaces [27,28]. This indicates a possible influence of the periodic potential on the photoemission spectrum which is seen as the QWS intensity enhancement at energies that correspond to the bulk transitions. The influence of the periodic potential and the localization potential edges on the photoemission intensities of QW states of s-p symmetry has been a subject of some controversies [29,30]. A systematic study of the dependence of the photoemission intensity of QW states of d-symmetry on the photon energy and film thickness is still missing. Now we focus on the 1-ML film which is structurally well-defined. The normal emission spectrum, that probes states in the centre of BZ, is characterized by d–QW states formed from K4,5 and K6 bands which have 4dxz,yz symmetry [28]. The DFT calculations show that these two states are consequence of the spin-orbit interaction [31] of the, otherwise degenerate, 4dxz,yz state [28]. The energy of the states was determined to be 4.42 eV and 4.67 eV which is very close to the values (4.37 eV and 4.60 eV respectively) reported in an early work of Shapiro et al. [32]. Notice that the energy positions of the QW states are rather well reproduced by the PA model (4.46 and 4.69 eV). For some quantum well systems the spectral width of the d-QW states had been used to estimate the localization of electrons within the overlayer film [33]. We have analyzed the normal emission spectrum taken at 80 K substrate temperature as shown in Fig. 5. The low temperature reduces relatively small phonon-induced broadening that appears at room temperature. The two peaks were fitted with pure lorentzians. The fitting peaks are very narrow, as judged from the lorentzian width (LW) of the synthetic peaks: LW(a)=89 meV and LW(b)=185 meV. The width of these peaks is even smaller than the widths of the corresponding states in silver monolayer films on Fe(1 0 0) [33], V(1 0 0) [20] and Mo(1 1 0) [20]. This corroborates the notion that electrons within the silver films on Ni(1 1 1) surface are well localized and that 4dxz,yz forms true quantum well states. We should point out that the two-lorentzian synthetic envelope nicely reproduces the experimental spectrum except for the asymmetry at the low binding energy side. As the 4d bands in silver show a great deal of localization the decay of

129

Fig. 5. Normal emission photoemission spectrum of 4d-derived quantum well states of 1 ML Ag film on Ni(1 1 1) surface fitted with: two-lorentzians (sum of two peaks is shown as a dashed line) and Doniach–Šunjic´ curve (thick line).

the valence band photo-hole might be affected by similar effects as core-holes, at least partly. Therefore, we applied Doniach–Šunjic´ [34] line-shape analysis to account for possible contribution of the electron-hole excitations at the Fermi level of the Ni substrate. By applying the asymmetry factor a = 0.05 we have obtained better agreement with the experimental data. It has been shown that in some cases the asymmetry of the bulk d-bands of silver and copper line-shape requires unphysical values of a indicating that other mechanisms for the photo-hole decay have to be applied [35]. Although, the overall asymmetry effect on the spectrum may look weak it clearly points out that the substrate may have an influence on the dynamics of the photo-hole in the overlayer film. The 4dxz,yz QW states show negative dispersion around the BZ centre (see Fig. 3a) which is characteristic for self-standing 1 ML silver film [28,31] as well as for the 1 ML film supported by some other metal substrates [36]. Generally, the in-plane dispersions of the silver 4d bands along the C  M high symmetry direction are almost the same as those observed in the monolayer films on Pd(1 1 1) (see for the comparison Fig. 5a in Ref. [28]). Most of the features shown in Fig. 3a can be reproduced by rigidly shifting silver bands in the film on palladium surface by 220 meV to higher binding energies, which is the energy difference between the 4dxz,yz QW states in the BZ centre at Ni and Pd surface. The energy difference between bands a0 and c at the zone edge (polar angle equal to 40°) is 0.95 eV, slightly smaller than for the Pd(1 1 1) substrate (1.1 eV) [28] which is consistent with the lateral compression of the silver lattice on palladium with respect to the relaxed layer on Ni(1 1 1). Having in mind that DFT calculation showed that silver monolayer 4d bands are virtually unaffected by the interaction with Pd substrate [28] we deduce that same conclusion can be extend to the 1 ML Ag/Ni(1 1 1) system. An additional difference between the bands in 1 ML films on Ni and Pd substrates is the discontinuity of band a (a0 ) in Fig. 3a. The structural differences between silver monolayers on Ni(1 1 1) and Pd(1 1 1), that may produce some differences in the electronic structure, are: (a) the existence of the moiré periodic structure on nickel surface with fully relaxed silver layer and (b) slightly compressed silver layer on palladium surface. While the expansion or compression can induce only slight shifts of the bands and affect band width [36] it is known that moiré periodic superstructure can open band gaps at the edge of the mini-zone boundary [37]. However, the ‘‘gap opening” observed in this experiment cannot be associated with the moiré induced effects. The reciprocal lattice

130

V.M. Trontl et al. / Surface Science 603 (2009) 125–130

vectors GNi and GAg of the Ni(1 1 1) surface and Ag monolayer, respectively, create moiré reciprocal vectors Gm = GNi  GAg . These vectors define positions in the BZ where the band gaps due to the superstructure open. The discontinuity of band a appears close to the half way between the BZ centre and M corresponding to kII which is much bigger than Gm. The DFT calculations of the electronic structure for unsupported silver monolayer with and without spin-orbit interaction indicate that the discontinuity of band a (a0 ) may be associated with the avoided crossing induced by the spin-orbit interaction [31]. In other words, it appears as if the spin-orbit interaction in monolayer silver film on Ni(1 1 1) is stronger than on Pd(1 1 1) substrate. A more systematic study has yet to prove this assumption. 5. Conclusions We have characterized the growth of ultra-thin silver films on Ni(1 1 1) surface by means of scanning tunnelling microscopy. We have confirmed the Stranski–Krastanov growth mode. By means of the high resolution photoemission spectroscopy we have studied the development of quantum well states of d-symmetry with increasing Ag coverage. The phase accumulation model has shown that the obtained spectra are consistent with the discretization of the weakly dispersing silver bands and SK growth mode. The line-shape analysis of the 4dxz,yz QW states corroborates strong localization within the silver overlayer despite the fact that the decay of the corresponding photo-hole might be under influence of dynamic effects in the Ni substrate. In-plane dispersion of the 1 ML QW states has most of the characteristics of the electronic structure of the silver monolayer on Pd(1 1 1) surface which is proved to be unperturbed by the interaction with supporting surface. This implies that the 4d electron states in monolayer silver film on Ni(1 1 1) are, to large extent, decoupled from the supporting surface and can be regarded as 2D electronic system. Finally, the comparison of the photoemission spectra of the 4 and 7 ML films with the spectrum of Ag(1 1 1) surface of the bulk specimen indicates a significant level of similarity suggesting considerable influence of the periodic potential on the photoemission intensities from these QW states. Acknowledgements The authors thank R. Brako for fruitful discussion. The financial support of the Ministry of Science, Education and Sports of the Republic of Croatia through Project No. 035-0352828-2840 is gratefully acknowledged.

References [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]

M. Milun, P. Pervan, D.P. Woodruff, Rep. Prog. Phys. 65 (2002) 99. T.-C. Chiang, Surf. Sci. Rep. 39 (2000) 181. J.E. Ortega, F.J. Himpsel, G.J. Mankey, R.F. Willis, Phys. Rev. B 47 (1993) 1549. J.J. Paggel, C.M. Wei, M.Y. Chou, D.-A. Luh, T. Miller, T.-C. Chiang, Phys. Rev. B 66 (2002) 233403. T. Valla, M. Kralj, A. Šiber, M. Milun, P. Pervan, P.D. Johnson, D.P. Woodruff, J. Phys. Condens. Matter 12 (2000) L477. M. Kralj, A. Šiber, P. Pervan, M. Milun, T. Valla, P.D. Johnson, D.P. Woodruff, Phys. Rev. B 64 (2001) 085411. D.-A. Luh, T. Miller, J.J. Paggel, T.-C. Chiang, Phys. Rev. Lett. 88 (2002) 256802. S. Mathias, M. Wiesenmayer, M. Aeschlimann, M. Bauer, Phys. Rev. Lett. 97 (2006) 236809. S. Mróz, Z. Jankowski, Surf. Sci. 322 (1995) 133. L.G. Feinstein, A. Blanc, D. Dufayard, Surf. Sci. 19 (1970) 269. K. Meinel, O. Lichtenberger, M. Klaua, Phys. Status Solidi (a) 116 (1989) 47. A.P. Shapiro, T. Miller, T.-C. Chiang, Phys. Rev. B 37 (1988) 3996. S. Nakanishi, K. Umezawa, M. Yoshimura, K. Ueda, Phys. Rev. B 62 (2000) 13136. A. Varykhalov, A.M. Shikin, W. Gudat, P. Moras, C. Grayoli, C. Carbone, O. Rader, Phys. Rev. Lett. 95 (2005) 247601. T. Miller, A. Samsavar, T. Chiang, Phys. Rev. B 50 (1994) 17686. B.V. Senkovskiy, A. Yu, A.M. Varykhalov, V.K. Shikin, Adamchuk O. Rader, Phys. Solid State 48 (2006) 1974. J. Jacobsen, L. Pleth Nielsen, F. Besenbacher, I. Stensgaard, E. Lgsgaard, T. Rasmussen, K.W. Jacobsen, J.K. Nørskov, Phys. Rev. Lett. 75 (1995) 489. S. Terada, T. Yokoyama, N. Saito, Y. Okamoto, T. Ohta, Surf. Sci. 433 (1999) 657. P. Lazic´, Zˇ. Crljen, R. Brako, Phys. Rev. B 71 (2005) 155402. M. Kralj, P. Pervan, M. Milun, T. Valla, P.D. Johnson, D.P. Woodruff, Phys. Rev. B 68 (2003) 245413. V. Yu. Aristov, M. Bertolo, K. Jacobi, F. Maca, M. Scheffler, Phys. Rev. B 48 (1993) 5555. C. Deisl, E. Bertel, M. Bürgener, G. Meister, A. Goldmann, Phys. Rev. B 72 (2005) 155433. F. Mittendorfer, A. Eichler, J. Hafner, Surf. Sci. 423 (1999) 1. N.V. Smith, N.B. Brookes, Y. Chang, P.D. Johnson, Phys. Rev. B 49 (1994) 332. P.S. Wehner, R.S. Williams, S.D. Kevan, D. Denley, D.A. Shirly, Phys. Rev. B 19 (1979) 6164. J.G. Nelson, S. Kim, W.J. Gignac, R.S. Williams, J.G. Tobin, S.W. Robez, D.A. Shirly, Phys. Rev. B 32 (1985) 3465. A. Elbe, G. Meister, A. Goldman, Surf. Sci. 397 (1998) 346. V. Mikšic´ Trontl, I. Pletikosic´, M. Milun, P. Pervan, P. Lazic´, D. Šokcˇevic´, R. Brako, Phys. Rev. B 72 (2005) 235418. M. Milun, P. Pervan, B. Gumhalter, D.P. Woodruff, Phys. Rev. B 59 (1999) 5170. A. Mugarza, J.E. Ortega, A. Mascaraque, E.G. Michel, K.N. Altmann, F.J. Himpsel, Phys. Rev. B 62 (2000) 12672. R. Brako, Private Communication. A.P. Shapiro, A.L. Wachs, T.C. Hseih, T. Miller, P. John, T.-C. Chiang, Phys. Rev. B 34 (1986) 7425. D.-A. Luh, J.J. Paggel, T. Miller, T.-C. Chiang, Phys. Rev. Lett. 84 (2000) 3410. S. Hüfner, Photoelectron spectroscopy principles and applications, Springer Series in Solid State Science, vol. 82, Springer, Berlin, 1995. A. Gerlach, K. Berge, T. Michalke, A. Goldmann, R. Müller, C. Janowitz, Surf. Sci. 497 (2002) 311. I. Pletikosic´, V. Mikšic´ Trontl, M. Milun, D. Šokcˇevic´, R. Brako, P. Pervan, J. Phys.: condens. Matter 20 (2008) 355004. M. Grioni, Ch.R. Ast, D. Pacile, M. Papagno, H. Berger, L. Perfetti, New J. Phys. 7 (2005) 106.