Growth of silver on ZnO and SnO2 thin films intended for low emissivity applications

1 Growth of silver on ZnO and SnO2 thin films intended for low emissivity applications Rafael Alvareza, Juan C. Gonzáleza, Juan P. Espinósa, Agustin R. González- Elipea, Ana Cuevab, Francisco Villuendasb aGroup of Nanotechnology of Surfaces. Instituto de Ciencia de Materiales de Sevilla (CSICUSE). Avda. Américo Vespucio 49. 41092 Sevilla (Spain) bGroup of Photonic Technologies. Departamento de Física Aplicada, Universidad de Zaragoza, c/ Pedro Cebuna 12, 50009, Zaragoza (Spain) Corresponding author: Rafael Alvarez Tlf.: +34 954 48 95 76 Fax: +34 954 46 06 65 [email protected] Abstract In the present work we have investigated the relationships existing between the optical properties and the growth mechanism, microstructure and surface roughness of SnO2 and ZnO oxide films prepared by magnetron sputtering under conditions resembling those utilized in industry. Thin films of these oxides with different thicknesses were characterized by atomic force microscopy, glancing incidence X-ray diffraction (GIXRD), X-ray reflectometry and spectroscopic Ellipsometry. The roughness evolution of the film properties (density, surface roughness and refraction index) as a function of their thickness has been evaluated within the concepts of the Dynamic Scaling Theory of thin film growth. Zinc oxide films were rougher than tin oxide films of similar *Manuscript Click here to view linked References 2 thickness, indicating a different growing mechanism for the two materials. Silver was evaporated onto the surface of the two oxide thin films and its earlier stages of nucleation studied by background analysis of the X-ray photoemission spectra. A different nucleation mechanism was found depending on the nature of the oxide acting as substrate. The superior performance of the zinc oxide based low emissive coatings is related with a better wetting of silver on the surface of this oxide despite the comparatively lower roughness of the tin oxide layers. Keywords: ZnO and SnO2 thin films, silver nucleation and wetting, low emissivity coatings, roughness evolution, growth mechanism. 1. Introduction Low emissivity glasses (low-e glass), used in homes, offices, automobiles, and other applications incorporate an oxide-metal-oxide composite coating that provides a high IR reflectivity.1 To achieve this functionality, these glasses incorporate a stacked structure consisting typically of a thin metal sandwiched between two oxides thin films. 2,3 This type of structures has been modeled and fabricated for several decades and nowadays represent a commercial product manufactured and sold by different companies all over the world.4 Low-e glasses must depict two main features: a high optical transparency in the visible region and a high reflectance in the near-infrared region. Although different metal/dielectric stacking structures have been proposed to achieve this goal, the most common coating structure consists of a silver layer sandwiched 3 between two transparent oxide layers, usually SnO2 and ZnO (i.e., MOx/Ag/barrier/MOx structures, where a barrier layer is introduced to avoid the oxidation of the silver layer during the second deposition process of the oxide). 4-6 Important characteristics of glasses incorporating these coatings are a low surface resistivity, a high optical transmittance in the visible and a high energy gap of about 3.6 eV and 3.3 eV, depending on the oxide. Despite the numerous theoretical modelling7.8 and experimental investigations9 dealing with these structures, there is a clear lack of fundamental knowledge relating their performance with critical microstructural characteristics of the system (e.g., roughness of the oxide layers, homogeneity of the metal film, etc.). Thus, only recently there have been some studies addressing systematically the relationship existing between these topographic properties of the individual layers, fundamentally the roughness of the oxide layers, and the final performance of the complete structure.10 In this regard, flat oxide surfaces seem to improve the transparency of the whole structure and avoid the prevalence of anomalous transmission losses caused by scattering of the light at the metal-oxide interfaces. It is also believed that flat oxide surfaces might contribute to decrease the agglomeration degree of the silver layers and, consequently, reduce its electrical resistance and the emissivity of the whole structure.11 Magnetron-sputtering (MS) is likely the most common technique for the largescale processing of low-e coatings.12 In principle, one of its recognized advantages is that the ion bombardment effects associated with this technique contribute to smooth the surface of the oxide films and hence to improve the quality of the silver overlayers. However, despite the application of similar 4 multilayer architectures, it is generally found that ZnO/Ag/barrier/ZnO systems are superior to equivalent SnO2/Ag/barrier/SnO2 multilayers in providing an enhanced emissivity and higher transmission in the visible.13 Trying to unravel the causes contributing to this different behavior, in the present paper we carry out first a systematic study of the evolution with thickness (from 20nm to 800nm) of the roughness and other properties such as density and refraction index of ZnO and SnO2 thin films prepared by MS on silicon and glass. To approach as much as possible the industrial conditions of preparation of these thin film oxides, deposition conditions have been selected by following the industrial criterion of maximizing the power density of the magnetrons for each oxide. Under these conditions, the growth mechanisms of the two oxides have been critically analyzed by using the concepts of the Dynamic Scaling Theory (DST) of surface growth.14,15 Then, we have studied the growth processes and wetting efficiency of silver deposited by evaporation on thin film surfaces of these two thin film oxides prepared by MS. For the first part of our study we have used X-ray reflectometry (XRF), X-ray diffraction (XRD), Scanning electron microscopy (SEM), atomic force microscopy (AFM) and spectroscopic ellipsometry. To assess the efficiency of silver to wet these two substrates we have employed the Tougaard´s principles16,17 to analyze the backgrounds of the X-ray photoemission spectra for successive evaporations of silver on ZnO and SnO2 thin film surfaces. This procedure permits to ascertain the growing morphology of a given material on rough substrates and, more specifically, if it forms a continuous overlayer or agglomerates in the form of islands.18,19 A critical evaluation of all these results in comparison with the actual emissivity values determined in real multilayer structures indicates that a better spreading 5 of silver on the surface of ZnO is the main cause of the lower emissivity of industrial ZnO based multilayers, even if tin oxide films present a lower roughness. 2. Experimental and Methods ZnO and SnO2 thin film preparation and characterization ZnO and SnO2 layers have been deposited by reactive magnetron sputtering on silicon and soda lime glass substrates under conditions resembling as much as possible those utilized in industry. The used experimental facility consists of a Leybold L-560 vacuum system equipped with 3 inches diameter magnetrons powered by pulsed DC power supplies. The distance between magnetrons and substrates is 250mm and the operating conditions by the reactive sputtering are summarized in Table 1. The discharge power was selected as 150W and 300 W for SnO2 and ZnO respectively, because these values are close to the maxima compatible with the magnetron size. These values yield power densities of 6.5 W/cm2 for ZnO and 3.3 W/cm2 for SnO2, quite similar to those used in industry by the manufacturing processes of low-emissivity coating where the dielectric layers are deposited to the maximum possible power. The O2 partial pressure values were set as the minimum pressures assuring that the deposition is made in the reactive mode yielding stoichiometric films. The selected values were determined from reactive deposition curves determined for each compound and chamber condition. In 6 our case the optimum values were 5x10-4 mbar and 8x10-4 mbar for SnO2 and ZnO respectively. The residual vacuum pressure was 5x10-6 mbar and layers of the two materials were deposited with thicknesses ranging from 20 to 800nm. The morphology of the oxide thin films was examined by cross section scanning electron microscopy (SEM) for thin films deposited on silicon and then cleaved. Images were taken in a field emission scanning electron Phillips FEG-SEM microscope. Atomic force microscopy (AFM) images were taken with a Dulcinea microscope from Nanotec (Madrid, Spain) working in tapping mode and using high frequency cantilevers with silicon tip. The root-mean-square (rms) roughness and the height-height correlation function were computed from 10x10m scans of films with increasing thickness, in order to obtain the roughness (  ) and growth (  ) exponents used for the dynamic scaling evaluation of thin film growth mechanisms. Grazing incidence x-ray diffraction (GIXRD) measurements taken for thin films thinner than 100 nm were carried out in a PANalitycal X’Pert Pro MPD diffractometer by using the Cu-Kα radiation at a fixed glazing incident angle of 0.5º and a 2 ranging from 10º to 80º, with a step size of 0,05º. The Bragg- Brentano X ray diffraction (XRD) analysis of films thicker than 100 nm was performed in a PANalitycal X’Pert Pro MPD diffractometer provided with X'Cellerator detector and graphite monochromator and using the Cu-Kα radiation, 40 kV, 40 mA, in a 2 range from 10º to 80º and a step of 2=0,016º. 7 Specular x-ray reflectivity analysis for films with a thickness d<100 nm was carried out in the PANalitycal X’Pert Pro diffractometer working in reflection mode (-2 scan), using the Cu-Kα radiation and a W/Si crystal as parabolic xray mirror. The scan axis was -2 and the scan mode was continuous with a scan range from 0.0º to 5.0º and a step size 0.005º. The X’Pert reflectivity program (PANalitycal B.V.) was used to fit the reflectivity raw data. The intensity reflections for ZnO and SnO2 thin films were modeled by means of a uniform homogeneous layered media with a sharply defined boundary to account for the Si-substrate. Spectroscopic Ellipsometry (SE) was carried out in a J.A. Woollam VASE (variable angle spectroscopic ellipsometry) spectroscopic ellipsometer. Values of  and  were obtained over the spectral range of 300 to 1400 nm, at 2 nm resolution. As a consistency check of the data, the incidence angle was varied at three angles of incidence: 65º, 70º, and 75º degree. Optical modeling and parameter fitting were done with the WASE32© program (J.A. Woollam Co.,Inc.). To model the ellipsometric spectra for a fixed film thickness (thickness values were obtained by transversal SEM images) non-uniformity in thickness and angular spread of the beam entering the detector were taken into account through fitting parameters. Partial polarization or monochromator bandwidth effects were not found. Quality assessment of the fit data was done by imposing that the mean-squared error (MSE) value was below 5 units for all studied samples. The ellipsometric optical model consisted of three layers deposited onto a silicon substrate that accounted for 1) roughness-, 2) Bruggeman effective medium approximation-, and 3) Cauchylayers. The rough surface layer was assumed to consist of 50% film material and 50% voids. The 8 Bruggeman effective medium approximation layer was coupled to the Cauchy layer. The Cauchy dispersion equation was used for the transparent region (> 500 nm). No dispersion of refractive index and extinction coefficient was considered for 500 <  < 1400 nm, while for  < 500 nm one classic Lorentz oscillator was added to the Urbach absorption of the Cauchy dispersion. Preparation and characterization of MOx/Ag/barrier/MOx low-E structures Low-E structures were prepared in the same installation Leybold 560, with 3 inch magnetrons, in which ZnO and SnO2 have been deposited. Deposition conditions of oxide layers are the same as for individual layers. Titanium was used as barrier layer and the silver low emissivity layer was deposited in 10-3 mbar Ar atmosphere. Reflection and transmission coefficients have been obtained in a homemade spectrophotometer, while emissivity has been obtained in a TIR100-2 (Inglass) emissometer as the reflectivity of radiation emitted at 100ºC, and derived from the sheet resistivity measured with an inductive sheet resistivity meter, Nagy SRM-12.20 Emissivity at 100ºC was measured in an emissometer (TIR100-2, Inglass) and derived from sheets resistivity measurements (Sheet Resistivity Meter SRM-12, Nagy. 9 XPS and QUASES analysis of the growth of silver To study the wetting behavior of silver on the two oxide surfaces, increasing amounts of metallic silver, from a fraction of a monolayer to several monolayers, have been deposited simultaneously on two clean surfaces of ZnO and SnO2 films of similar thickness (around 20 and 200 nm) for the two materials. Since the results were quite similar irrespective of the thickness of the oxide thin films we only describe in the text the results obtained for the thinnest ones. The simultaneous evaporation on the two substrates ensures that each set of samples contain the same amount of deposited silver. The metal was deposited, step by step, by thermal evaporation from a home-made resistive source, consisting of three wrapped W wires (0.3 mm diameter) with an Ag wire (0.2 mm diameter) wound around. Evaporation was carried out in the preparation chamber of the X-ray photoemission spectrometer, under ultra-high vacuum (P<10-8 mbar), at a deposition rate of ~0.1 nm/min, with the substrates kept at room temperature situated at 30 cm from the evaporation source. After each deposition, the samples were transferred to the analysis chamber of the spectrometer under ultra-high vacuum conditions (base pressure, 10-10 mbar). Prior to the first deposition stage, the substrates were surface cleaned in situ by treatment with a plasma of O2/Ar (5% of O2) obtained with a microwave downstream plasma source, to remove the adventitious carbon contaminating their surface. XPS spectra were recorded after each Ag deposition step with a VG ESCALAB 210 spectrometer, with the hemispherical energy electron analyzer working in the pass energy constant mode at a value of 30 eV. Non 16 deposition conditions we can estimate (Zn) ~ 0.22 and (Sn) ~ 0.15. These  values imply that Zn atoms arrive more obliquely at the film surface than Sn atoms and that shadowing will be more important in ZnO growth, in agreement with the higher  value of these films. Other factors affecting the roughness of the films might also cause the observed differences in the growth exponent. Several experimental studies suggest that oriented grain growth can enhance the surface roughening during film growth33,34 which, in turn, could further enhance the shadowing effects and lead to higher  values. This difference in growth behaviour might explain the different growth exponent values found in SnO2 and ZnO films, in agreement with the fact that XRD measurements performed on our films show that ZnO films present a preferential (002) orientation. Optical properties of thin films The determined optical parameter functions of the SnO2 and ZnO thin films measured by spectroscopic ellipsometry are reported in Figure 6. These parameter functions have been obtained after modeling the ellipsometric raw data as explained in the experimental section. As a justification of the model employed for the simulations, it is important to stress that both the specular X- ray reflectivity and AFM techniques have shown that the films are characterized by a nanoscopic roughness, where both the mean height and correlation length of the irregularities are much smaller than the wavelength of light in the visible and infrared part of the spectrum. For these rather flat surfaces it is expected that multiple scattering depolarizations are not important and therefore the 17 contribution of the field induced polarization of the surface to the far-field radiation pattern can be approximated by layers of a polarizable Bruggeman EMA model sandwiched between a perfect substrate and air, both taken as continuous media.35 Then the Cauchy model was employed to determine the film thickness, roughness and real refractive index in the region where the films were transparent (500<<1400 nm). According to the plots in Figure 6, the real part of the refraction index presents a smooth dispersion behavior for >450 nm. In the spectral region 300<<450 nm, this real part displays a variable dispersion due to the presence of the semiconductor band gap represented in our case by a Lorentz oscillator. From these results, it is important to stress that the real refraction index increases with the film thickness as noted in figures 6(a) and 6(c). Moreover, the extinction coefficient shows some absorption features for <450 nm, but becomes negligible in the region 450>>1400 nm (c.f., figures 6(b) and 6(d)), in agreement with the high transparency expected for these films. After fitting the experimental spectra, the obtained thickness values were close to nominal values, i.e., 21.9nm, 49.3nm, 94.9nm, 193.6nm, 343.4nm and 655.6nm for the ZnO samples and 24.5nm, 59.1nm, 109.3nm, 227.4nm, 427.4nm and 838.3nm for the SnO2 films. These values are in relatively good agreement with those obtained by specular X-ray reflectivities for d≤100 nm and SEM analysis (see Figure 1). A similar agreement is found with the roughness results provided by AFM: the calculated roughness parameters of the films were 7.5nm, 27.5nm and 19.9nm for ZnO, and 2.9nm, 1.8nm and 5.3nm for SnO2, for the films of nominal thicknesses of 200nm, 400nm and 800nm, respectively. 18 Spreading of silver on SnO2 and ZnO thin film surfaces A way of assessing the spreading of metals deposited onto real surfaces presenting a specific roughness is by following the deposition process by XPS and studying the evolution of the background of the spectra with the deposited amount. We have previously applied this methodology for a large variety of systems where the growth mechanism consisted on the formation of islands.18,19,36,37 Herein, we have applied a similar methodology to ascertain the nucleation and growth behavior of silver on the surface of both SnO2 and ZnO thin films. The idea of this methodology is to analyze the evolution of the backgrounds behind the elastic photoelectron peaks as a function of the amount of deposited material. As an example of the observed changes, Figure 7 shows four Ag3d photoemission peaks recorded for increased amounts of Ag deposited on a ZnO film with a thickness of 20 nm. For the sake of comparison, the spectra have been normalized to the height of the Ag3d5/2 elastic peak. Similar results were obtained on a ZnO thin film with a thickness of 200 nm and therefore higher roughness. This equivalence in spectral results for the two thin film thicknesses was also found when comparing the series of experiments performed on SnO2 thin films (spectra not shown). It is apparent in this figure, that the heights of the backgrounds behind the elastic peaks increase with the amount of deposited silver. The analysis of the shape of these backgrounds provides information about the growing mechanism of silver. Such an analysis has been carried out here for a series of experiments where the metal has been evaporated simultaneously on the surfaces of ZnO and SnO2 films. This ensures that the two substrates receive the same amount of silver. Consequently, any difference in the growing mode of 19 silver on each particular substrate will be unequivocally evidenced by differences in both the intensity of the elastic Ag3d photoemission peaks and the height and shape of their inelastic backgrounds. The spectra have been simulated with the QUASES software.21 This analysis has shown that silver condenses on the surface of these two oxides in the form of three dimensional clusters, whose size and height increase with the amount of deposited metal. The average height of these islands (Ih) and the surface coverage (Sc) of the surface of the oxides determined with the QUASES calculations are represented in Figure 8 for the whole series of experiments carried out here on oxide thin films with a thickness of 20 nm. Equivalent curves were obtained by analyzing the results of the deposition experiments carried out on the SnO2 and ZnO thin films with 200 nm thickness. For a better guidance, in this plot we have also included a series of the dotted lines representing theoretical curves obtained by assuming that the silver islands have a cubic form and that the product IhxSc is constant and equivalent to the amount of deposited silver as indicated in the right axis of the plot. Figure 8 reveals that for equivalent amounts of deposited silver, the coverage degree of the oxide surface is always greater (or alternatively, the islands´ heights smaller) on ZnO than on SnO2. This behavior implies that under our experimental conditions silver wets and spreads better on the former oxide than on the latter. This is further sustained for the results corresponding to very low amounts of deposited silver on ZnO, clearly indicating that this metal completely spreads on the surface of this oxide and that it only starts to nucleate on its surface after the fourth evaporation experiment. Besides these trends, another interesting feature of the reported experiment are that total coverage implying 20 the complete coalescence of the silver islands only occurs after depositing an equivalent amount of silver of aproximately14 nm. It must be stressed that the similar results obtained when analyzing the deposition of silver on SnO2 and ZnO of 200 nm thickness sustain that silver spreading mainly depends on the chemical nature of the sputtered deposited surfaces and that their roughness is a negligible factor for the control of silver wetting on these two oxide thin films. Optical behavior of MOx/Ag/barrier/MOx structures It is likely that the different spreading behaviour of silver on the two oxides may have an influence on the optical behaviour of the final stacked MOx/Ag/barrier/MOx structures. To test the final optical and low-emissivity properties of low-emissivity coatings, we have prepared structures of the type SnO2/Ag/barrier/SnO2, and ZnO/Ag/barrier/ZnO where both the oxide and the silver layers were deposited by MS according to experimental protocols resembling those utilized in the industry. Thickness of the oxides was kept around 40 nm, while the thickness of the Ag layer was 9 nm in the two cases to get identical contribution of Ag layer in both structures. Thicknesses of both dielectric layers were determined to get coatings with a green-bluish neutral color aspect both in transmission and in reflection modes. Optical properties of both structures are shown in Figure 9. These plots represent the transmission and the reflectivity of the two examined structures. Thicknesses of the two dielectric layers were determined to get coatings with a green-bluish neutral color aspect both in transmission and in reflection modes. 21 Figure 9 shows the transmission and the reflectivity curves obtained for the two examined structures. Besides, the emissivity at 100ºC of both structures yielded values of 0.09 for the SnO2-based structure and 0.07 for ZnO-based structure, results that agree with the commonly accepted trend of lower emissivity values for ZnO-based structures. A similar conclusion can be gained by observing the plots in Figure 9 where small, though not negligible differences can be noticed between the SnO2 and ZnO structures, with a slightly higher transmission in the IR range for the former. The different behavior found for the two structures can be accounted for by assuming a different distribution of silver on the two oxides and/or that the oxide themselves present different roughness. Referring to this latter point, the morphological analysis of the oxide thin films reported previously has clearly evidenced that for all range of investigated thicknesses ZnO prepared by MS is rougher than SnO2. Hence, if this were the sole factor contributing to the different emissivity of the two structures, SnO2-based structures should have a superior performance than the ZnO ones. Therefore, we have to admit that the more favorable tendency of silver to spread onto the surface of ZnO surface than on that of SnO2 evidenced by our silver evaporation experiment should be the main factor contributing to the superior performance of the ZnO based stacked structure. As it is shown in Figure 8, for a thickness of 9 nm in our experiment (note that the experimental conditions are different than those used to prepare the silver film in the stacked structures which implies the use of magnetron sputtering), the coverage degree of the substrate is higher than 80% on ZnO, while it is only around 60% for SnO2, Assuming a similar behavior for 22 MS layers, a higher coverage degree (and therefore a higher percolation and continuity of the metal sheet) in the stacked structure should reduce the sheet resistivity of the silver deposited onto ZnO and thus generate a lower emissivity value for the whole coating. 4. Summary and Conclusions In this work we have primarily studied the evolution of roughness of ZnO and SnO2 sputtered layers prepared by simulating industrial conditions of deposition. Then, we have analyzed the wetting and coverage degree of Ag deposited on these thin films in order to clarify the different emissivity behavior of MeOx/Ag/barrier/MeOx low-emissivity coatings. The analysis of the roughness evolution of the films by using the principles of the DST shows that the higher roughness presented by the ZnO thin films must be mainly attributed to an enhancement of the shadowing effects due to the preferential growth of these crystalline films according to a given crystallographic plane. Despite the higher roughness of the ZnO thin films, silver appears to spread more efficiently on the surface of this material than on that of SnO2 thin films. Coverage degree by the silver layer, instead of roughness of the oxide layers, has been revealed as the main factor explaining the lower emissivity values of ZnO/Ag/barrier/ZnO structures. Future works to improve emissivity of this type of structures should therefore concentrate in getting higher coverage degrees of the oxide (i.e., improve the wetting by the deposited metal), rather than on reducing the roughness of the oxide layers. 23 Acknowledgements We thank the Junta de Andalucía (Projects P09-CTS- 5189, TEP5283 and FQM-6900) and the Ministry of Economy and Competitiveness (Projects CONSOLIDER CSD2008-00023, MAT2010-21228, and MAT2010-18447) for financial support. JC González thanks the Spanish Scientific Research Council (CSIC) for a JAE-Doc contract (2009-2012) at the ICMSE-CSIC-US. References 1 C.G. Granqvist, Adv. Mater. 15, 1789 (2003) 2 C. Guillén, J. Herrero, Thin Sol. Films 520 1 (2011) 3 J. Kulczyk-Malecka, P.J. Kelly , G. West, G.C.B. Clarke, J.A. Ridealgh, Thin Sol. Films 520, 1368 (2011) 4 B. P.Jelle, A. Hynd, A. Gustavsen, D. Arasteh, H. Goudey, R. Hart, Solar Ener. Mater. Solar Cells 96, 1 (2012) 5 P. Grosse, R. Hertling, T. Miiggenburg, J. Non-Crystal. Solids 218 38 (1997) 6 E. Ando, M. Miyazaki, Thin Sol. Films 516, 4574 (2008) 7 M. Rubin, D. Arasteh and J. Hartmann, Int. Comm. Heat Mass Transfer 14, 561 (1987) 8 J. Mohelnikova, Const. Build. Mater. 23, 1993 (2009) 9 G.T. West, P.J. Kelly, J.W. Bradley, IEEE Transact. Plasma Sci. 38, 3057 (2010) 24 10 K. Kato, H. Omoto, T.Tomioka, A. Takamatsu, Solar Ener. Mater. Solar Cells, 95,2352 (2011) 11 Z.B. Huang, W.C. Zhou, X.F. Tang, D.M. Zhu, F. Luo, Thin Sol. Films 519, 3100 (2010) 12 J. Szczyrbowski, G. Brauer, M. Ruske, H. Schilling, A. Zmelty, Thin Sol. Films 351, 254 (1999) 13 R.J. Martin-Palma, L. Vazquez, J.M. Martinez-Duart, Malats-Riera, Solar Ener. Mater. Solar Cells 53, 55 (1998) 14 A.-L. Barabasi, H.E. Stanley, Fractal Concepts in Surface Growth, CambridgeUniversity Press, Cambridge, England, 1994. 15 M. Pelliccione, T.-M. Lu, Evolution of Thin Film Morphology: Modeling and Simulations, Springer Series in Materials Science Vol. 108, Springer, New York, 2008 16 S.Tougaard, J.Vac.Sci. & Technol.A, 14, 1415, (1996) 17 S. Tougaard, Surf. Interf. Anal. 26, 249 (1998). 18 A.I. Martín-Concepción, F. Yubero, J.P. Espinós, A.R. González-Elipe, S. Tougaard, J. Vac. Sci. & Technol. A 21 1393 (2003) 19 F.Gracia, F.Yubero, J.P.Espinós and A.R.González-Elipe, Appl. Surf. Sci. 252, 189 (2005) 20 H. J. Glässer. Large Area Glass Coatings. Von Ardenne Anlagentechnik GMBH 2000, p. 188 21 S.Tougaard, QUASES, Software package for quantitative XPS/AES of surface nanostructures by inelastic peak shape analysis, www.QUASES.com 25 22 S.Tanuma, C.J.Powell, D.R.Penn, Surf.Interf. Anal., 21, 165 (1993). 23 J. Krim, J.O. Indekeu, Phys. Rev. E 48, 1576 (1993) 24 S. Das Sarma, P. Tamborenea, Phys. Rev. Lett. 66, 325 (1991) 25 Z.J. Liu, N. Jiang, Y.G. Shen, Y.W. Mai, J. Appl. Phys. 92, 3559 (2002) 26 M.A. Auger, L. Vazquez, R. Cuerno, M. Castro, M. Jergel, O. Sanchez, Phys. Rev. B 73, 045436 (2006) 27 A. Yanguas-Gil, J. Cotrino, A. Barranco, A.R. Gonzalez-Elipe, Phys. Rev. Lett. 96, 236101 (2006) 28 S. Yim, T.S. Jones, Phys. Rev. B 73, 161305 (2006) 29 J.G. Yu, J.G. Amar, Phys. Rev. E 66, 021603 (2002) 30 W. Schwarzacher, J. Phys. Condens. Matter 16, 859 (2004) 31 M. Pelliccione, T. Karabacak, T.-M. Lu, Phys. Rev. Lett. 96, 146105 (2006) 32 J.M. Garcia-Matin, R. Alvarez, P. Romero-Gomez, A. Cebollada, A. Palmero, Appl. Phys. Lett. 97, 173103 (2010) 33 J.H. Xu, L.H. Yu, I. Kojima, J. Appl. Phys. 94, 6827 (2003) 34 X. Jiang, B. Hao, Journal of Crystal Growth 312, 1844 (2010) 35 D.E. Aspnes, J. B. Theeten and F. Hottier. Phys. Rev. B. 20, 3292 (1979) 36 C.Mansilla, F. Gracia, A.I. Martin-Concepción, J.P. Espinós, J.P. Holgado, F. Yubero, A.R. González-Elipe, Surf. Interf. Anal., 39, 331 (2007). 37 J.P.Espinós, A.I.Martín-Concepción, C.Mansilla, F.Yubero, A.R.González- Elipe, J. Vac. Sci. Technol. A 24, 919 (2006) 10 100 1000 10 -1 10 0 10 1 10 -1 10 0 10 1 10 -1 10 0 = 0.88 Roughness (nm) T hickness (nm) = 0.44 SnO2 ZnO a) Height - height correlation (normalized) r (normalized) SnO2 - 2 00 nm SnO2 - 400 nm SnO2 - 800 nm ZnO - 2 00 nm ZnO - 400 nm ZnO - 800 nm = 0.77 b) Figure 5 Click here to download Figure(s): Fig 05.eps 1.8 2.0 2.2 2.4 20 nm 5 0 nm 100 nm 200 nm 400 nm 800 nm R e fr a ct io n I nde x S nO2 R e fr a ct io n I nde x E x ti c tio n C oe ffic ie n t 1.8 2.0 2.2 2.4 20 nm 5 0 nm 100 nm 200 nm 400 nm 800 nm ZnO 400 600 800 1000 1200 1400 0.00 0.06 0.12 (c) (d) (a) E x ti c tio n C oe ffic ie n t Wavelength (nm) (b) 400 600 800 1000 1200 1400 0.0 0.4 0.8 Wavelength (nm) Figure 6 Click here to download Figure(s): Fig 06.eps 440 420 400 380 360 340 Ag 3d Intensity (a.u) B in d in g Ener g y (eV) 1. 6 A 6 .7 A 30 . 4 A 1 36 . 0 A Ag/ZnO Nom inal thickness Figure 7 Click here to download Figure(s): Fig 07.eps 0.0 0.2 0.4 0.6 0.8 1.0 0 20 40 60 80 100 120 140 Ag/ZnO Coverage Isl and Height (A) Ag/SnO 2 Nom ina l Thickne ss Ag (A ) 0 20 40 60 80 100 120 140 Figure 8 Click here to download Figure(s): Fig 08.eps 0 20 40 60 80 500 1000 1500 2000 2500 0 20 40 60 80 Reflectivity (%) SnO 2 Glass SnO 2 Layer ZnO Glass ZnO Layer a) Transmission (%) (nm) SnO 2 ZnO b) Figure 9 Click here to download Figure(s): Fig 09.eps 1 Table 1.- Experimental conditions of sputtering deposition Material Target Total pressure (mbar) O2 pressure (mbar) Power (W) Power density W/cm-2 SnO2 Tin 1 x 10-3 5 x 10-4 150 3.3 ZnO Zinc 1 x 10-3 8 x 10-4 300 6.5 Table 2.- Structural and microstructural parameters of the films deduced from their XRD and X-ray reflectivity analysis Sample Size of crystalline domains (nm) Average Density (g/cm3) Surface Roughness (nm) Substrateinterface Roughness (nm) SnO2 20 nm 50 nm 100 nm 200 nm 400 nm 800 nm 3.4 3.3 3.1 7.3 6.6 6.5 5.95 6.60 6.83 0.5 0.9 1.1 0.5 0.6 0.4 ZnO 20 nm 50 nm 100 nm 200 nm 400 nm 800 nm 4.0 4.7 4.6 8.2 6.6 6.4 4.64 4.97 4.84 1.0 1.0 1.0 1.1 1.0 1.0 Table(s)