Nanocolumnar Association and Domain Formation in Porous Thin Films Grown by Evaporation at Oblique Angles

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by IOP Publishing in: Nanotechnology (2016), vol. 27, no. 39 available at: https://doi.org10.1088/0957-4484/27/39/395702 ©IOP. En idUS Licencia Creative Commons CC BY NC ND 1 Nanocolumnar Association and Domain Formation in Porous Thin Films Grown by Evaporation at Oblique Angles C. Lopez-Santos 1,* , R. Alvarez 1 , A. Garcia-Valenzuela 1 , V. Rico 1 , M. Loeffler 2 , A. R. Gonzalez-Elipe 1 , A. Palmero 1,* 1 Instituto de Ciencia de Materiales de Sevilla (CSIC-Universidad de Sevilla). Americo Vespucio 49, 41092 Seville, Spain. 2 Dresden Center for Nanoanalysis (DCN), Center for Advancing Electronics Dresden (cfaed), Helmholtzstr. 18, 01187 Dresden Germany * Email: [email protected]; [email protected] Abstract.- Porous thin films grown at oblique angles by evaporation techniques are formed by tilted nanocolumnar structures which, depending on the material type and growth conditions, associate along certain preferential directions, giving rise to large domains. This arrangement, commonly denoted as bundling association, is investigated in the present work by performing fundamental experiments and growth simulations. It is proved that trapping processes of vapor species at the film surface, together with the shadowing mechanism, mediate the anisotropic widening of the nanocolumns and promote their preferential coalescence along certain directions, giving rise to domains with different shape and size. The role of these two processes is thoroughly studied in connection with the formation of these domains in materials as different as SiO 2 and TiO 2 . Page 1 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2 Keywords.- Porous thin films, Nanocolumns association, TiO 2 , SiO 2 , OAD evaporation Introduction Nanomaterials and porous materials showing direction-dependent properties are widely employed in numerous technological fields such as electronic and optoelectronic devices [1-3], memory elements [4], energy applications [5,6], photonic devices [7] or sensors [8-11], among others [12, 13]. In all these cases, the anisotropic response reflects the different geometrical arrangement of elemental structures at different scales, from nanometric patterns (as in nanorods or nanowires [14,15]) to mesoscopic structures [16] associated to the coalescence of nanometer-size elements along a preferential direction. A typical example of this anisotropic morphology corresponds to porous films grown by the electron beam-assisted deposition technique operated at oblique angles (e-beam assisted OAD) [1, 5, 7, 17], where the film microstructure is defined by large arrays of slanted nanocolumns with diameters in the order of few tens nanometers [18,19]. Besides the intrinsic morphological anisotropy associated to the existence of these tilted nanostructures, in some cases, they may associate and give rise to anisotropic domains with typical scale length of 1 micrometer [20-22]. This phenomenon has been commonly referred as “bundling association”, a concept that may encompass a rich variety of microstructural types. Our group has used these bundles for the development of advanced dichroic composite materials thanks to the incorporation of gold or silver nanoparticles in the open pores of these host structures [7,21]. We have also shown that optofluidic devices presenting bundling associations Page 2 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 and open mesoporous channels can be used for the analysis of liquids [17]. Yet, the nature of the large scale coalescence of columnar nanostructures is still unclear, and only few studies have roughly described the basic features of this type of association [23-27]. From these previous studies, it appears that while some materials and deposition geometries favor the development of these domains, in other cases, the nanocolumns may remain rather isolated [28, 29]. The need of a conceptual framework to explain how these associations arise is the main motivation of this paper, where we thoroughly analyze the association of nanocolumns in thin films with different chemical composition and under different geometrical configurations. This work relates to a previous publication by our group [30], where we studied the growth of nanocolumnar porous thin films by the e-beam assisted technique at oblique angles and, in particular, the relation between the incident angle of the deposition flux onto the substrate, α , and the tilt angle of the columns, β . In agreement with other studies in the literature, we proved that the heuristic tangent rule, tan α = 2 tan β , [31] only applies to very specific cases and that the chemical nature of the deposited material strongly mediates the columnar growth and tilt angle. We also demonstrated that the relation between both angles can be explained by considering two fundamental mechanisms: i) the angular broadening of the deposition flux from the evaporation source to the film, highly dependent on the characteristics of the deposition reactor, and ii) the so-called surface trapping mechanism, by which vapor species that pass few angstroms away from the material surface may bend its otherwise rectilinear trajectory and be deposited. A good match between theoretical calculations and experimental data was found for numerous materials such as SiO 2 , TiO 2 , Ti, Si, ITO or Ni, among others [32]. Within this same framework, in this work we study the association mechanisms of nanocolumns from experimental and simulation points of view, aiming at determining Page 3 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 4 the fundamental mechanisms responsible for the development of large anisotropic domains. For this purpose, we have chosen TiO 2 and SiO 2 thin films as case examples which, according to previous results in ref. [32], depict rather different nanocolumnar features (i.e., shape of nanocolumns and tilt angle) and, hence, are recommended for this analysis. Experimental Setup Amorphous TiO 2 and SiO 2 thin films were deposited on silicon substrates by electron beam-assisted evaporation at oblique angles at room temperature. Evaporation was carried out in an O 2 atmosphere at a pressure of 10 −2 Pa using TiO (Kurt J. Lesker, 99.9%) and SiO 2 (Sico Technologies GmBH) pellets as target material, respectively. For comparative purposes, metallic Ti thin films were also grown using metallic titanium (Kurt J. Lesker, 99.9%) as a target (the residual pressure during deposition was 10 −5 Pa). The films were prepared in the experimental system already described in ref. [30] by placing the substrates at zenithal angles of α =70º and α =85º with respect to the target, located 50 cm apart. From now forth, samples will be labelled using this deposition angle followed by its chemical composition (e.g., 70-TiO 2 or 85-TiO 2 and similarly for SiO 2 and Ti). Relevant directions within the samples for the present study are those corresponding to the evaporation line (∆ direction) and its perpendicular ( Π direction), as depicted in figure 1. The thickness of the layers were set at approximately 600 nm by fixing the evaporation rate at 0.1 nm s −1 by means of a calibrated quartz crystal monitor. The microstructure of the films was firstly examined by a field enhanced scanning electron microscope (FESEM) HITACHI S4800. To assess the existence of possible Page 4 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 5 associations between the tips of the nanocolumns at the surface, normal view FESEM micrographs were analyzed using the home-made MatLAB-based domain detection software IDUN (Image Domains UNcovered), which identifies all unconnected regions in a given image. It operates according to the following sequential protocol: i) the brightness and contrast of the image are locally enhanced to avoid inhomogeneities; ii) the image is converted into a black and white binary map; iii) features with size below certain threshold are considered noise and are removed, and iv) a MatLAB-based numerical routine is employed to differentiate each independent domain (two columns are considered to belong to the same domain if there is a continuum path from one to the other). A version of this software, incorporating a user-friendly interface, can be freely downloaded from our webpage [33]. Finally, based on this analysis, the value of the following morphological quantities have been estimated: mound size, taken as the size of the nanocolumnar tips at the surface in the ∆ and Π directions, shadow length, defined as the typical distance between nanocolumns in the ∆ direction and the domain length, taken as the average size of the domains. Focused Ion Beam (FIB-SEM) analysis was performed using a FEI Helios Nanolab 660 tool. First, SEM images were acquired normal to the surface at 2 kV using both secondary electrons (SE) for morphological and back scatter electrons (BSE) for compositional information. Thereafter, an optional strip of Pt was deposited using electron beam and ion beam induced deposition in order to prevent ion damage to the surface. Subsequently 2 trenches were milled at 30keV Ga+ ion energy at 90° angle with one another and oriented parallel and orthogonal to the preference direction seen in the SEM images from the surface (∆ and Π directions, respectively). Especially for TiO 2 , it has been observed that without the protective Pt layer, the top of the layers became rounded and lost their fine-grained morphology due to spill-over ions. After Page 5 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 6 coarse milling, the side surfaces were polished using a 30 keV, 80 pA Ga + ion beam. The cuts were then observed with the electron beam under 52°of incidence in the same modes (SE and BSE) as mentioned before. Theoretical Model The growth model is described in detail in ref. [30], where it has been successfully applied to describe the growth of TiO 2 thin films prepared by evaporation at oblique angles. Besides, it has also been successfully tested to describe the nanocolumnar growth of materials as different as Ti, Si, Ni, Ta 2 O 5 or SiO 2 , among others, in ref [32]. Here, we describe briefly its fundamentals although, for a more detailed description, readers are addressed to these previous works [30, 32]. The model considers the deposition of vapor species on a cubic three-dimensional grid with periodic boundary conditions, whose cells may take the values 0 (empty cell) or 1 (filled cell). Particles are assumed to arrive at the substrate following a narrow Gaussian angular distribution function with nominal angle α and variance σ . The model also assumes that vapor species follow straight trajectories until they hit the surface, where they stick. Yet, we consider that when a particle moves 3–4 Å over the surface, it may stick at that particular location with a probability, s t , the so-called surface trapping probability. As it was demonstrated in ref. [32], s t depends on the chemical nature of the evaporated species and the film surface, ranging from 0 for metals (e.g., Ti, Ag or Ni) to 1 for SiO 2 or Ta 2 O 5 . It is also worth mentioning that, in the case of oxides, we assume that once a vapor species is deposited, it becomes fully oxidized due to the oxygen-rich atmosphere in the deposition reactor. L L H N N N × × Page 6 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 7 The growth of the film is considered to take place in the so-called zone I of the Thornton Structure Zone Model (SZM) (i.e., when , where is the film temperature and the melting temperature of the deposited material), in which surface shadowing mechanisms are known to dominate over thermally activated mobility processes [34]. A MatLAB-based computer code, Surface TRapping in Oblique Nanostructured Growths (STRONG), is freely available in our webpage to solve the presented model [33]. Based on the results in ref. [32], the geometry of the deposition reactor induced an angular broadening of the vapor flux of σ =6º, whereas the trapping probability was s t =0.12 for TiO 2 , s t =1.00 for SiO 2 and s t =0.00 for Ti . These values have been used in the calculations, assuming a substrate with size 3000 3000 × particles with periodic boundary conditions and, unless it is stated otherwise in the text, a film thickness of 2000 monolayers (1 monolayer ~ 3-4 Angstroms). Simulation were run in parallel in a Dell Workstation (32 cores, 2.66 GHz, 256 Gb RAM), taking approximately 4 weeks to end. Results and Discussion Domain structures A first assessment on the existence of columnar associations appears in figures 2a-d, where the top view SEM micrographs of the TiO 2 and SiO 2 deposited films show that nanocolumns are not randomly distributed over the surface but associate into domains (note that, as mentioned in the Experimental Setup section, the thickness of the films is approximately 600 nm in all the cases). For instance, figure 2a (70-TiO 2 ) shows the existence of a large and fully coalescent domain, containing strips that extend along the / 0.3 f m T T < f T m T Page 7 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 8 Π direction. A similar situation is encountered for the 70-SiO 2 film (figure 2b), although column seems larger in this case. Meanwhile, in the 85-TiO 2 and 85-SiO 2 films (figures 2c and 2d), anisotropic domains of various sizes and shapes are apparent (2D Fast Fourier Transforms of figure 2a-d are given in the supplementary information). To clearly assess them, in figures 3a-d we have analyzed figures 2a-d with the IDUN software, employing colors to differentiate each independent domain. Figures 3a-b reveal the existence of a large single domain spreading along both ∆ and Π directions in the samples 70-TiO 2 and 70-SiO 2 mentioned above, even though columns seem to be ordered following strip-like patterns in the Π direction that also touch in the ∆ direction. Conversely, figures 3c-d show the existence of large domains for the two samples prepared at 85º: while the 85-SiO 2 film (figure 3d) depicts large associations of columns along the Π direction, the 85-TiO 2 film contains anisotropic domains that extent along both Π and ∆ directions. The values of the typical length of surface features and average domain size are gathered in Table I, and marked in figures 3a-d for a visual evaluation. Interestingly, domain lengths are larger for SiO 2 in comparison with TiO 2 (c.f., figure 3 and Table I). From the evaluation of these images it also appears that the average values of the mound size and the shadow length increase with the deposition angle and are higher for SiO 2 than for TiO 2 thin films (c.f., Table I). It is important to note that no shadow length exists along the Π direction, and that is why we have not included any value in Table I. To extend the analysis below the film surface, we have studied the possible association of nanocolumns in the bulk of the layers by means of FIB-SEM analyses. Figures 4a-b show cross-sectional FIB-SEM micrographs of the 70-TiO 2 and 70-SiO 2 samples taken from the ∆ and Π perspectives, respectively. Top left image in figure 4a (70-TiO 2 thin film) corroborates the existence of highly coalescent tilted nanostructures with an angle Page 8 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 homogenous distance between them provides a homogeneous intercolumnar void space very useful for the development of host structures for the fabrication of advanced composite films [7]. Conclusions In this work, we have studied the anisotropic growth of SiO 2 and TiO 2 thin films deposited at oblique angles and analyzed the results with a model relying on the competition between surface shadowing and surface trapping processes. We have demonstrated that this latter mechanism affects the columnar growth and its anisotropic widening over the film surface. These columnar associations have been explained by solving a deposition model that links the surface shadowing and trapping efficiencies to the anisotropic development of surface nanostructures and, ultimately, to the columnar association and formation of domains. Overall, it is concluded that the same fundamental principles that govern the growth of the nanocolumns, also promote the columnar associations along preferential directions. This is of relevance to tailor the columnar arrangement of the films for certain applications where anisotropic structures are required, or even for their use as a template to induce the anisotropic growth of other materials with different chemical compositions. Yet, we have constrained our study to typical evaporation conditions at low temperatures, where thermally-induced surface mobility processes, crystallization or potential barrier mobility limitations do not compete with surface shadowing or trapping mechanisms in the morphological development of the layers. This implies that the extrapolation of our results to other conditions (e.g., in typical magnetron sputtering depositions), although possible, must be critically analyzed. For instance, from a fundamental point of view, the concept of trapping efficiency is intimately connected to the interaction between a vapor species Page 15 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 16 and the film surface, which would strongly depend on their chemical nature and kinetic energy: in typical evaporation conditions, this latter quantity amounts ~0.05 eV, while in magnetron sputtering depositions, for instance, it may reach values above 10 eV, thus making surface trapping processes very unlikely. Nevertheless, and taking into account specific limitations in each particular case, our results can be considered of general application whenever the constrains of low temperatures, preferential oblique angle of incidence of deposition species onto the film and relatively low kinetic energy of deposition species are fulfilled. Acknowledgements The authors thank the Junta de Andalucía (P12-FQM-2265) and the Spanish Ministry of Economy and Competitiveness (Projects, MAT2013-40852-R, MINECO-CSIC 201560E055) for financial support. References [1] Hawkeye M M, Taschuk M T and Brett M J 2014 Glancing Angle Deposition of Thin Films: John Wiley & Sons, Ltd) pp i-xiv [2] Lu W and Lieber C M 2007 Nanoelectronics from the bottom up Nat Mater 6 841-50 [3] Yan H, Choe H S, Nam S, Hu Y, Das S, Klemic J F, Ellenbogen J C and Lieber C M 2011 Programmable nanowire circuits for nanoprocessors Nature 470 240-4 [4] Wuttig M and Yamada N 2007 Phase-change materials for rewriteable data storage Nat Mater 6 824-32 [5] Flaherty D W, Hahn N T, May R A, Berglund S P, Lin Y-M, Stevenson K J, Dohnalek Z, Kay B D and Mullins C B 2012 Reactive ballistic deposition of nanostructured model materials for electrochemical energy conversion and storage Acc Chem Res 45 434-43 [6] Li X, Dhanabalan A, Meng X, Gu L, Sun X and Wang C 2012 Nanoporous tree-like SiO2 films fabricated by sol–gel assisted electrostatic spray deposition Microporous and Mesoporous Materials 151 488-94 Page 16 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 17 [7] Gonzalez-García L, Parra-Barranco J, Sanchez-Valencia J R, Ferrer J, Garcia-Gutierrez M-C, Barranco A and Gonzalez-Elipe A R 2013 Plasmonic Nanostructures: Tuning Dichroic Plasmon Resonance Modes of Gold Nanoparticles in Optical Thin Films (Adv. Funct. Mater. 13/2013) Advanced Functional Materials 23 1605- [8] Cui Y, Wei Q, Park H and Lieber C M 2001 Nanowire Nanosensors for Highly Sensitive and Selective Detection of Biological and Chemical Species Science 293 1289-92 [9] Kesapragada S V, Victor P, Nalamasu O and Gall D 2006 Nanospring Pressure Sensors Grown by Glancing Angle Deposition Nano Letters 6 854-7 [10] Moon H G, Choi Y R, Shim Y-S, Choi K-I, Lee J-H, Kim J-S, Yoon S-J, Park H-H, Kang C-Y and Jang H W 2013 Extremely Sensitive and Selective NO Probe Based on Villi-like WO3 Nanostructures for Application to Exhaled Breath Analyzers ACS Applied Materials & Interfaces 5 10591-6 [11] Wu B, Wang L, Wu H, Kan K, Zhang G, Xie Y, Tian Y, Li L and Shi K 2016 Templated synthesis of 3D hierarchical porous Co3O4 materials and their NH3 sensor at room temperature Microporous and Mesoporous Materials 225 154-63 [12] Barranco A, Borras A, Gonzalez-Elipe A R and Palmero A 2016 Perspectives on oblique angle deposition of thin films: From fundamentals to devices Progress in Materials Science 76 59-153 [13] Islam S Z, Reed A, Kim D Y and Rankin S E 2016 N2/Ar plasma induced doping of ordered mesoporous TiO2 thin films for visible light active photocatalysis Microporous and Mesoporous Materials 220 120-8 [14] Duan X, Huang Y, Cui Y, Wang J and Lieber C M 2001 Indium phosphide nanowires as building blocks for nanoscale electronic and optoelectronic devices Nature 409 66-9 [15] Holmes J D, Johnston K P, Doty R C and Korgel B A 2000 Control of Thickness and Orientation of Solution-Grown Silicon Nanowires Science 287 1471-3 [16] Fodor B, Agocs E, Bardet B, Defforge T, Cayrel F, Alquier D, Fried M, Gautier G and Petrik P 2016 Porosity and thickness characterization of porous Si and oxidized porous Si layers – An ultraviolet– visible–mid infrared ellipsometry study Microporous and Mesoporous Materials 227 112-20 [17] Oliva-Ramirez M, González-García L, Parra-Barranco J, Yubero F, Barranco A and González-Elipe A R 2013 Liquids Analysis with Optofluidic Bragg Microcavities ACS Applied Materials & Interfaces 5 6743-50 [18] Karabacak T, Mallikarjunan A, Singh J P, Ye D, Wang G-C and Lu T-M 2003 β-phase tungsten nanorod formation by oblique-angle sputter deposition Applied Physics Letters 83 3096-8 [19] Lintymer J, Martin N, Chappe J-M and Takadoum J 2008 Glancing angle deposition to control microstructure and roughness of chromium thin films Wear 264 444-9 [20] Alouach H and Mankey G J 2004 Texture orientation of glancing angle deposited copper nanowire arrays Journal of Vacuum Science & Technology A 22 1379-82 [21] Parra-Barranco J, García-García F J, Rico V, Borrás A, López-Santos C, Frutos F, Barranco A and González-Elipe A R 2015 Anisotropic In-Plane Conductivity and Dichroic Gold Plasmon Resonance in Plasma-Assisted ITO Thin Films e-Beam-Evaporated at Oblique Angles ACS Applied Materials & Interfaces 7 10993-1001 [22] Stagon S P, Huang H, Kevin Baldwin J and Misra A 2012 Anomaly of film porosity dependence on deposition rate Applied Physics Letters 100 061601 [23] Abelmann L and Lodder C 1997 Oblique evaporation and surface diffusion Thin Solid Films 305 1- 21 [24] Mukherjee S and Gall D 2013 Structure zone model for extreme shadowing conditions Thin Solid Films 527 158-63 [25] Poxson D J, Mont F W, Schubert M F, Kim J K and Schubert E F 2008 Quantification of porosity and deposition rate of nanoporous films grown by oblique-angle deposition Applied Physics Letters 93 101914 [26] Tanto B, Ten Eyck G and Lu T-M 2010 A model for column angle evolution during oblique angle deposition Journal of Applied Physics 108 026107 Page 17 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 18 [27] van Kranenburg H and Lodder C 1994 Tailoring growth and local composition by obliqueincidence deposition: a review and new experimental data Materials Science and Engineering: R: Reports 11 295-354 [28] Izquierdo-Barba I, García-Martín J M, Álvarez R, Palmero A, Esteban J, Pérez-Jorge C, Arcos D and Vallet-Regí M 2015 Nanocolumnar coatings with selective behavior towards osteoblast and Staphylococcus aureus proliferation Acta Biomaterialia 15 20-8 [29] Krause K M, Thommes M and Brett M J 2011 Pore analysis of obliquely deposited nanostructures by krypton gas adsorption at 87 K Microporous and Mesoporous Materials 143 166-73 [30] Álvarez R, González-García L, Romero-Gómez P, Rico V, Cotrino J, González-Elipe A R and Palmero A 2011 Theoretical and experimental characterization of TiO2 thin films deposited at oblique angles Journal of Physics D: Applied Physics 44 385302 [31] Haanstra J M N a H B 1966 Microfractography of thin films Philips Tech. Rev. 27 [32] Alvarez R, Lopez-Santos C, Parra-Barranco J, Rico V, Barranco A, Cotrino J, Gonzalez-Elipe A R and Palmero A 2014 Nanocolumnar growth of thin films deposited at oblique angles: Beyond the tangent rule Journal of Vacuum Science & Technology B 32 041802 [33] Alvarez R and Palmero A 2016 Software IDUN for determination of domains in an image http://nanoscops.icmse.csic.es/software/idun/ [34] Petrov I, Barna P B, Hultman L and Greene J E 2003 Microstructural evolution during film growth Journal of Vacuum Science & Technology A 21 S117-S28 Page 18 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 19 Table Caption Table 1.- Estimated values of surface characteristic lengths in the TiO 2 and SiO 2 thin films appearing in figures 3a-d. The values obtained from the simulations are given between brackets. Sample Mound size (nm) Shadow length (nm) Coalescent domain size (nm) ∆ Π ∆ Minor axis Major axis 70-TiO 2 ~50 (50) ~65 (60) ~ 90 (90) ∞ (∞) ∞ (∞) 70-SiO 2 ~90 (80) ~130 (115) ~170 (185) ∞ (∞) ∞ (∞) 85-TiO 2 ~75 (75) ~95 (85) ~180 (205) ~155 (145) ~ 450 (410) 85-SiO 2 ~115 (100) ~140 (125) ~ 320 (305) ~ 210 (220) ~ 650 (570) Page 19 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 20 Figure Caption Figure 1.- Idealized scheme of the nanocolumn arrangement in the OAD thin films with indication of the ∆ and Π directions. Figure 2.- Normal view SEM micrographs of a) 70-TiO 2 , b) 70-SiO 2 , c) 85-TiO 2 , d) 85- SiO 2 . Insets in each image correspond to the simulations of these structures using the developed model. Figure 3.- Color domains resulting from the application of the border detection software to the images in figure 2. Different colors highlight domains of associated nanocolumns. White rectangles, parallels lines and ellipses highlight the geometrical features evaluated for the geometrical analysis of, respectively, mound sizes, shadowing lengths and coalescent domain sizes. Figure 4.- FIB-SEM image of films from the ∆ and Π perspectives (note that, as mentioned in the Experimental Setup section, there is a thin Pt coating on the patterns): a) 70-TiO 2 , b) 70-SiO 2 . Results of the simulations appear just below each image. Figure 5.- FIB-SEM image of films from the ∆ and Π perspectives (note that, as mentioned in the Experimental Setup section, there is a thin Pt coating on the patterns): a) 85-TiO 2 , b) 85-SiO 2 . Results of the simulations appear just below each image. Figure 6.- Schemes for the influence of the trapping mechanism on the columnar growth: a) the existence of a trapping region promotes a larger shadow length, b) the straightening up of the columns associated to higher trapping probabilities leads to longer shadow lengths. Figure 7.- a) Results of the growth simulations when a cylindrical seed is used as nucleation site for an incident angle of the vapor flux of 85º for s t =0.00, s t =0.12 and s t =1.00. b) Relation between the parameter / L L ε Π ∆ = and the surface trapping probability, s t , for different simulated thin film thickness. Figure 8.- Top view scheme of the columnar development when a) ε > 1, b) ε ∼ 1 and c) ε < 1. Figure 9.-Normal view SEM micrograph of a metallic Ti thin film grown at α =85º. Simulation of the growth with s t =0.00 is included as an inset. Page 20 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 Figure 1 Figure 2 Page 21 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 22 Figure 3 Figure 4 Page 22 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 23 Figure 5 Page 23 of 28 AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 24 Figure 6 Page 24 of 28AUTHOR SUBMITTED MANUSCRIPT - NANO-110700.R1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60