Structural and optical characterization of nanostructured ZnO grown on alumina templates

Structural and optical characterization of nanostructured ZnO grown on alumina templates M Yuste 1 , R Escobar Galindo 1 , O Martínez Sacristán 3 , I Mínguez-Bacho 1 , Sonia Rodriguez 2 , M Hernández-Vélez 2 and O Sánchez 1,4 1 Instituto de Ciencia de Materiales de Madrid (CSIC), Spain 2 Universidad Autónoma de Madrid, Spain 3 UGdS-Optronlab Group, Dpto. Física Materia Condensada, Universidad de Valladolid, Paseo de Belén 11, 47011—Valladolid, Spain E-mail: [email protected] Received 29 July 2014, revised 3 October 2014 Accepted for publication 24 October 2014 Published 12 November 2014 Materials Research Express 1(2014) 045028 doi:10.1088/2053-1591/1/4/045028 Abstract The growth of nanostructured ZnO thin films on nanoporous anodic alumina substrates (NAAF) by dc reactive magnetron sputtering using a pure Zn metal target is reported. ZnO nanostructures reproduce the pore arrays of the NAAF substrates used as templates mimicking their hexagonal long range order. Stoichiometric nanostructured ZnO samples were grown with wurtzite type structure, highly textured and oriented in the (002) direction. The study of the NAAF pore size effect in the final morphology and optical properties of the nanostructured ZnO is presented for different ZnO thicknesses. The pore size of the nanostructured ZnO films was controlled in the range of 15–65nm by choosing appropriate NAAF and the sputtering deposition conditions. The broad emission band observed in PL spectra of the samples should be associated with color centers transitions (F and H centers) appearing in the alumina templates. Keywords: nanostructured ZnO, nanoporous anodic alumina, magnetron sputtering 1. Introduction Zinc oxide (ZnO), with band gap of 3.37 eV at room temperature (RT) and large excitonic binding energy, around 60 meV, has attracted increasingly worldwide attention in the past few decades due to its unique properties and a large number of potential applications such as transparent conductive contacts [1], solar cells [2], laser diodes [3], ultraviolet lasers [4], thin 4 Author to whom any correspondence should be addressed. Materials Research Express 1(2014) 045028 2053-1591/14/045028+16$33.00 © 2014 IOP Publishing Ltd film transistors [5], thin-film gas sensors [6] and optoelectronic and piezoelectric applications to surface acoustic wave devices [7–9]. ZnO films grown on Si substrates have been widely reported, however, ZnO nanostructured thin films show very interesting additional properties in comparison to continuous thin films [10,11]. Those properties are likely to be altered and controlled by the ZnO growth processes [12] such that many research groups have reported the synthesis of ZnO nanostructures by means of both bottom-up and top-down techniques, among them, chemical vapour deposition (CVD) [13], molecular beam epitaxy (MBE) [14], electrochemical deposition [15], evaporation and RF sputtering [16]. Magnetron sputtering techniques have been successfully used in the deposition of functional materials such as, TiN [17] and TiO 2 [18] on nanostructured templates achieving a conformal reproduction of the main features of the substrate surfaces. This technique gives directly the required nanostructured material without additional treatment after the deposition process. In addition, this technique allows developing deposition processes with high stability and purity control of the final products. Moreover, the nanostructures are fabricated using a vacuum-based technique offering a higher level of process stability and purity control than in the case of solution based methods. In this work, we report the growth of nanostructured ZnO thin films on nanoporous anodic alumina films (NAAFs) used as templates, by dc magnetron sputtering. The surface morphology and structural properties of the ZnO nanostructures were characterized by means of field-emission scanning electron microscopy (FESEM), atomic force microscopy (AFM) and x-ray diffraction (XRD). The chemical composition was determined by RBS (Rutherford backscattering spectrometry). The optical properties were determined by UV-visible spectrophotometry and photoluminescence (PL) measurements. The use of dc magnetron sputtering technique allows us to produce material deposits with grain sizes smaller than 30–40 nm which is an advantage with respect to other physical deposition techniques. In the early stages of the deposition, the complex structures of the relief pattern are perfectly covered by small particles of material to be deposited, growing a layer of material on ‘conformal contact’with the pattern whose surface is reproduced with high efficiency [17]. The combination of dc magnetron sputtering technique and anodization techniques to produce ZnO nanostructures provides an economical strategy to obtain ZnO nanomaterials. 2. Experimental details ZnO films were grown by using a home-made high vacuum dc magnetron sputtering system on NAAF templates. The sputtering cathode consists of a standard circular planar magnetron with a high purity commercial Zinc target (99.99%), 3-in. diameter and 0.125-in. thickness. All experiments were performed introducing a mixture of Ar (99.999%) and O 2 (99.992%) in the vacuum chamber at a constant total gas flow (Ar + O 2 ) of 30 sccm. The chamber was pumped out to a base pressure of 2 · 10 −6 mbar and the working pressure was approximately of 3·10 −3 mbar. During the deposition process the relative O 2 gas concentration in the reactive atmosphere (p O 2 ) was maintained at 50%, the target power at 50 W and the target-substrate distance at 20 cm. NAAF templates were prepared by following the two-step anodization process described elsewhere [19]. Sulfuric and oxalic acid solutions were used as electrolytes in the anodization 2 Mater. Res. Express 1(2014) 045028 M Yuste et al processes. Table 1summarizes the pore sizes, Φ, and inter-pore distances, D int . of the NAAF used as substrates. In order to perform the optical characterization of the samples (in transmittance mode) the aluminum back substrates were removed by an etching process using an acid solution (CuCl 2 /HCl). These structures exhibit high optical transparency in the UVvisible ranges and high chemical and thermal stability which makes them very suitable for use as substrates-patterns, as will be shown later. RBS experiments were carried out to elucidate the chemical composition of the deposited ZnO films. RBS were performed using the 5MeV HVEE Tandetron accelerator located at the ‘Centro de Micro-Análisis de Materiales of Universidad Autónoma de Madrid’. The samples were irradiated with 3.035 MeV He + ions to make use of the cross section resonance 16 O(α,α) at that particular energy and, therefore, to improve the sensitivity to oxygen nucleus. The chemical composition of the ZnO films has been extracted using the RBX software [20]. Crystalline structure of the films was determined by XRD using a Cu anode (Cu K α = 1.54 Å) registered under grazing incidence conditions with an incidence angle of 0.7° at room temperature. The XRD measurements were performed on a Siemens D-5000 diffractometer. The surface and fracture cross-section morphologies of the thin films were examined by a Nova NanoSEM 230 FEG scanning electron microscope (HRSEM). An AFM system from Nanotec Electrónica S. L. working in non-contact mode at ambient conditions was also used to characterize the surface morphology of the samples. Optical characterization of NAAFs is not abundant in the current literature and their influence on the optical properties of functional thin films deposited on them has been scarcely reported. Certainly it is not an easy task. Therefore, as a first approach, we made a study of the optical behaviour of the NAAF templates used in this work for completeness after the optical characterization of the ZnO/NAAF samples. Transmittance measurements were acquired in a SolidSpec-3700 near UV-visible and near infrared range (190–1500 nm) spectrophotometer. PL spectra were carried out by means of a LabRAM system from Horiba Jobin-Yvon, model HR800 UV, with an He–Cd laser (325 nm) as the excitation light source. 3. Results and discussion 3.1. ZnO nanostructures Prior to the growth of ZnO on NAAF templates we performed a study to optimize the experimental conditions for growing this compound forming well defined nanostructures on the selected templates displayed in table 1. Table 1. Pore size and inter-pore distance of the NAAF substrates. NAAF Φ(nm) D int (nm) M0 15 44 M1 15.5 45 M2 18 51 M3 21.5 57 M4 25 65 M5 35 105 M6 50 105 3 Mater. Res. Express 1(2014) 045028 M Yuste et al In agreement with our results previously reported [21], the experimental sputtering conditions chosen in this work led to the growth of nearly stoichiometric ZnO films at a low deposition rate (8 nm min −1 ) which enhanced the conformal character of the sputtering technique. Hence, under these conditions, 150 nm of ZnO were grown on M4 and M5 samples with order parameters values (Φand D int ) in the intermediate range of the previously fabricated NAAFs (see table 1). Figure 1shows typical sectional HRSEM images of M4 (a) and M5 (b) NAAFs after sputtering of ZnO films. The growth of nanostructured ZnO films is clearly shown reproducing the pore arrangement of NAAFs perpendicularly to the Al substrates. The x-ray diffractograms corresponding to M4 and M5 samples coated with ZnO are displayed in figure 2. In both spectra, signals of cubic aluminum phase [22] (JCPDS card No. 40–787) coming from the aluminum disk used to grow the NAAF templates are present. For comparative purposes figure 2also shows the spectrum of a ZnO film 150nm thick grown in the same conditions on Silicon 〈100〉. According to our previous results [21], ZnO film on silicon presents wurtzite type structure, highly textured and oriented in the 〈002〉direction. The mean crystal size value in that preferred direction was estimated in 22 ± 5 nm by using Scherrer’s formula [23]. When ZnO is deposited on the NAAF templates, the 〈002〉preferred orientation remains for both coatings but with a greater FWHM than those exhibited by the thin film on silicon. This fact results in lower crystal sizes achieving values of 16 ± 5 nm in both samples. Other diffraction peaks corresponding to ZnO hexagonal crystalline structure could also be found in these Figure 1. Sectional view of samples M4 (a) and M5 (b) (25 and 35 nm pore size, respectively), after depositing ZnO on its surface. 4 Mater. Res. Express 1(2014) 045028 M Yuste et al diffractograms. The dispersion in the preferred orientations is attributed to the surface roughness of the grown ZnO nanostructures mimicking the surface features of the NAAF templates. It is known that substrates with small roughness promote better crystalline quality of ZnO films [24]. Figure 3(a) shows the experimental and simulated RBS spectra of the NAAFs M4 (a) and M5 (b) before and after being coated with ZnO. From RBS fittings we derived Zn/O ratios of 49/51 and 48/52 for the films deposited on M4 and M5, respectively i.e. a ZnO compound almost stoichiometric). In the case of M4 a signal has been detected located at 1850 keV, corresponding to the sulfur contamination coming from the sulfuric acid used in the NAAF growth. Elemental concentrations obtained from the NAAF fits were Al/O/S = 38/59/3 for M4 and Al/O = 39/61 for M5, these values being very close to the stoichiometric composition of alumina (Al 2 O 3 ). In the case of M4 template, a signal at 1850 keV has been detected, corresponding to the sulfur contamination coming from the sulfuric acid used in the NAAF growth. The main differences between both RBS analyses arise when studying the ZnO/NAAF interface (see dotted lines in figure 3(a)). For the case of the ZnO film deposited on M4, the ZnO/NAAF interface is abrupt and the rise of the aluminum signal was found to resemble the one of the uncoated M4 substrate. By contrast, for the sample deposited on M5 there are evidences of diffusion of the ZnO into the substrate. In order obtain a more accurate determination of the ZnO/NAAF interface we have calculated the in-depth concentration profiles. Figure 3(b) shows the profiles for samples M4 (a) and M5 (b) after the deposition of ZnO. The RBX fit made in figure 2(a) was obtained assuming a single ZnO layer on Al 2 O 3 substrate. This profile presented an abrupt interface down to the resolution limit of RBS (approximately 1 nm [25]) with the composition of both Al 2 O 3 and ZnO, being homogeneous throughout the entire thickness. On the other hand, for M5 sample, in order to obtaining a good RBX fit of the spectrum (figure 2(b)) it was necessary to consider the presence of intermediate layers with a variable composition Zn x Al y O z (x+y+z= 1) between the ZnO and the NAAF substrate. Therefore, in this sample the indepth concentration profile reveals that the ZnO/ NAAF interface is not completely abrupt having approximately 50 nm in thickness. These results suggest that there is some growth of ZnO into the alumina but only in the NAAFs with Figure 2. XRD patterns for samples M4 and M5 coated with ZnO. 5 Mater. Res. Express 1(2014) 045028 M Yuste et al higher pore size, i.e. M5 template. This diffusive effect in the growth of functional materials on NAAFs has been reported elsewhere [18]. On the other hand, one of the crucial parameters to obtain well defined nanostructures on porous surfaces is the thickness of the functional material to be sputtered on them, in order to avoid the completed coverage of the porous in the NAAF template. In order to study the coverage of the NAAF surfaces, MO templates (with the smallest pore size i.e. 15 nm) were coated with three different ZnO thicknesses: 29, 43 and 108 nm. HRSEM images of these samples (not shown) revealed that only for the thinnest coating (29 nm) was a porous ZnO surface developed. For a ZnO thickness around 40nm the pore surface structure collapse and only a small percentage (∼20%) of pores could be distinguished. Besides, most observed porous surfaces lose the long range order with hexagonal symmetry and the typical grain growth of ZnO thin films on flat substrates [21] was observed. Finally, in the case of coatings thicker than 108 nm, no evidence of pore growth was detected. Taking into account these results, ZnO films with constant thickness of 30 nm were grown on a NAAF series (samples Figure 3. (a) RBS spectra (experimental and simulated) of samples M4 and M5 before and after the deposition of ZnO, (b) concentration depth profiles extracted from RBS measurements for M4 (a) and M5 (b) after the deposition of ZnO. 6 Mater. Res. Express 1(2014) 045028 M Yuste et al Figure 4. SEM images of NAAF surfaces before and after depositing 30 nm of ZnO. 7 Mater. Res. Express 1(2014) 045028 M Yuste et al M1–M6) with increasing pore diameter sizes, from 15 to 50 nm, to study the pore diameter effects in the morphology and optical properties of nanostructured ZnO films. Figure 4shows HRSEM images of the starting NAAF and after the deposition of 30nm of ZnO on them. There is a change in the contrast of the HRSEM images after the ZnO deposition. This effect is related to the partial reduction of pore diameters being more pronounced for samples with smaller pore diameter sizes, i.e., samples M1 and M2. In order to obtain a quantitative analysis of the pore diameter reduction after 30nm of ZnO deposition the HRSEM images were processed by using the Image J program [26]. In table 2, the nominal and final pore diameters, as well as the estimated pore diameter reduction are shown. In this way it was possible to modulate the pore diameter size and also their lattice parameter of the nanostructured ZnO films by choosing the appropriate NAAF template or tailoring others not reported in this work. Figure 5(a) shows a typical AFM image of the surface of the deposited samples along with its Fourier fast transform (FFT) pattern acquired from image processing software [27]. The FFT image provides the high hexagonal ordering degree of nanostructured ZnO layer, revealing the existence of a self-ordered pores structure (up to 3rd neighbours) in a single domain. This fact allows us to confirm the guided growth of ZnO nanostructure on the NAAF template surfaces. In figure 5(b) the surface profile along the path reflects an accurate measurement of the pore size and interpore distance after the ZnO deposition. The width and height of the peaks in the surface profile gave place to an average pore diameter (Φ) and interpore distance (D int ) values of 45 ± 4 nm and 112 ± 5 nm, respectively according to the expected data indicated in table 1. 3.2. Optical characterization of ZnO/NAAF nanostructures The characterization of samples formed for more than one phase is not an easy task. In our case we have fabricated nanostructures compound by two different solid phases, i.e., on one side the nanoporous anodic alumina with insulator properties and on the other hand, zinc oxide, which is a well-known II–VI semiconductor. Therefore, we have studied the optical properties of these nanostructures in two steps, firstly those of the NAAF and later on the joint effect on such properties of the nanostructured ZnO/NAAF. Table 2. Pore diameters obtained processing HRSEM images for samples with and without ZnO deposit. Sample theoretical Φ(nm) calculated Φ(nm) % Reduction M1 15.5 15 31 M1+ZnO —10.5 — M2 18 19 16 M2+ZnO —16 — M3 21.5 21 11 M3+ZnO —19 — M4 25 27 8 M4+ZnO —25 — M5 35 39 2.5 M5+ZnO —38 — M6 50 57 7 M6+ZnO —53 — 8 Mater. Res. Express 1(2014) 045028 M Yuste et al Figures 6and 7show the transmittance spectra of the uncoated NAAF substrates, samples M1–M6, and the ones corresponding to 30nm of ZnO sputtered on those NAAFs, respectively. Figure 7has shifted vertically for clear purposes. Both spectra groups reveal the high transparency (>80%) of all samples in the UV-Visible range. The observed Fabri–Perot interference fringes for the longest wavelengths are associated with interferences among multiple reflections coming from the air–film interfaces, and also probably from internal interfaces existing in the bulk NAAFs due to their complex compositional structure, as RBS results suggest. 3.2.1. NAAF absorption coefficients and band gaps. In this work, the absorption coefficients of the all samples were derived from the transmittance spectra and by using the Beer–Lambert Law [28] given by: Figure 5. (a) AFM image of sample M6 covered with 30 nm of ZnO with the corresponding FFT, (b) topographic surface profile along the line shown in (a). 9 Mater. Res. Express 1(2014) 045028 M Yuste et al [33] Miyazaki S 2001 J. Vac. Sci. Technol. B19 2212 [34] Ealet B, ElyakhloufiM H, Gillet E and Ricci M 1994 Thin Solid Films 250 92 [35] Ciraci S and Batra I P 1983 Phys. Rev. B28 982 [36] Hong S, Kim E, Kim D W, Sung T H and No K J 1997 J. Non-Cryst. Solids 221 245 [37] Li G H, Zhang Y, Wu Y C and Zhang L D 2003 J. Phys.: Condens. Matter 15 8663 [38] Momida H, Nigo S, Kido G and Ohno T 2011 App. Phys. Lett. 98 042102 [39] Xu W L, Chen H, Zheng M J, Ding G Q and Shen W Z 2006 Opt. Mater. 28 1160 [40] Minseok C, Anderson J and van de Walle C G 2013 J. Appl. Phys. 113 044501 [41] Cardona M 1993 Springer Series in Solid State Sciences vol 105 (Berlin: Springer) [42] Lushchik A, Kirm M, Lushchik C, Martinson I and Zimmerer G 2000 J. Luminesc. 87-89 232 16 Mater. Res. Express 1(2014) 045028 M Yuste et al