Stability and Structural Characterization of Epitaxial NdNiO 3 Films Grown by Pulsed Laser Deposition

Transcription

1 Mat. Res. Soc. Symp. Proc. Vol Materials Research Society Stability and Structural Characterization of Epitaxial NdNiO 3 Films Grown by Pulsed Laser Deposition Trong-Duc Doan, Cobey Abramowski, and Paul A. Salvador Carnegie Mellon University, Department of Materials Science and Engineering, Pittsburgh, PA, ABSTRACT Thin films of NdNiO 3 were grown using pulsed laser deposition on single crystal substrates of [100]-oriented LaAlO 3 and SrTiO 3. X-ray diffraction and reflectivity, scanning electron microscopy, and atomic force microscopy were used to characterize the chemical, morphological and structural traits of the thin films. Single-phase epitaxial films are grown on LaAlO 3 and SrTiO 3 at 625 C in an oxygen pressure of 200 mtorr. At higher temperatures, the films partially decompose to Nd 2 NiO 4 and NiO. The films are epitaxial with the (101) planes (orthorhombic Pnma notation) parallel to the substrate surface. Four in-plane orientational variants exist that correspond to the four 90 degenerate orientations of the film's [010] with respect to the in-plane substrate directions. Films are observed to be strained in accordance with the structural mismatch to the underlying substrate, and this leads, in the thinnest films on LaAlO 3, to an apparent monoclinic distortion to the unit cell. INTRODUCTION Thin film deposition of complex oxides having interesting properties, such as ferroelectricity, magnetoresistivity, superconductivity, optical activity, has attracted great attention for the development of advanced and novel devices (see for example [1]). Moreover, thin film deposition has also proven to be an important tool for synthesizing new and metastable phases and for allowing certain strain states to be accessed, which can lead to a variation in the film's structure and properties. Because advanced multilayer structures and complex device architectures require deposition of thin films with controlled structural and physical properties at very small length scales, it is imperative to understand the nature of materials as epitaxial thin films and their difference to their bulk counterparts if many of the potential applications are to be realized. In this communication we describe how the structure of the metastable perovskite oxide NdNiO 3 is affected by the growth conditions, underlying substrate, and overall film thickness. The perovskite nickelates, of the stoichiometry ANiO 3 (A = Lanthanide, Y), are of interest because of their physical properties, which include a sharp, thermally driven metal-insulator transition. The temperature at which this transition occurs, T MI, is known to be a strong function of the A cation and of external pressure[2-4]. Such electronic properties render these materials useful as metallic electrodes [5, 6] in ferroelectric devices, as optical switches, as bolometers, and as actuators. From a fundamental scientific perspective, the first order metal-insulator transition that occurs is of great interest because it is not a common feature of complex oxides and it is strongly coupled to the structural parameters (bond angles and distances) and weakly to the magnetic properties. For a review of these materials see Ref. [7] One of the drawbacks to the applicability of the RNiO 3 compounds (not including R = La) is that the materials are unstable at elevated temperatures (> C) at 1 atm of pressure. The most common approach to their synthesis involves the application of hydrostatic pressure at high temperatures to react the oxides in the presence of an oxidizing agent [2, 3, 7]. GG3.27.1

2 The entire series can be synthesized in this manner, although the pressure must be increased for the smaller lanthanides. Thin film deposition has been used as an alternate method of synthesis for R = La, Pr, Nd, and Sm. Extensive work has been carried out on the stable LaNiO 3 material as a metallic electrode. The high-pressure phases, R = Pr, Nd, Sm, were first synthesized by sputtering [8, 9], and they were more recently grown by pulsed laser deposition (PLD) [10, 11] and metal organic chemical-vapor deposition (MOCVD) [12]. Because the structural parameters play an important role in the determination of the physical properties, the T MI can be drastically varied by changing these parameters. For NdNiO 3, T MI has been decreased from 200 K to 100 K by application of hydrostatic pressure [13, 14]. In all cases where thin films have not been subjected to a post-deposition high-pressure anneal, the T MI has been suppressed from the bulk value of 200 K. It is clear that the biaxial stress that the films experience from the rigid substrate leads to a variation in the structure and properties of the films. This communication reports the evolution of the stability and structure of NdNiO 3 films on [100]-oriented SrTiO 3 (STO) and LaAlO 3 (LAO) substrates. The physical properties of these films will be reported elsewhere. EXPERIMENTAL The NdNiO 3 PLD-target was synthesized using a standard powder synthesis route [15]. The target was black, dense, and consisted of a mixture of NiO and Nd 2 NiO 4 according to the X-ray diffraction pattern. The overall stoichiometry was reconfirmed with energy dispersive spectroscopy to be 1:1 for Nd:Ni. The target was then mounted inside of the deposition chamber and used for film growth by pulsed laser ablation in a system described previously [15]. Polished, cleaned [15] single crystal substrates (Crystal GMBH, Berlin, Germany) of approximate dimensions mm were glued to the commercial heater support with silver paste. Samples were heated to the deposition temperature, which ranged from 575 to 800 C, in a background pressure of 10-5 mtorr. A dynamic deposition atmosphere was established [15] with an oxygen pressure of 200 mtorr. The KrF laser (λ = 248 nm, pulse duration 20 ns) was operated at 3 Hz and was focused to an energy density of 2.25 J/cm 2 at the surface of the target. The target to substrate distance was 60 mm. After cleaning the target surface [15], films were deposited for an amount of time between 30 min and 2 hr. At the end of each deposition, the oxygen pressure was increased to 300 Torr and films were cooled at a rate of 20 C/min in this static oxygen atmosphere. X-ray diffraction (XRD) was carried out in Θ 2Θ, ω, and φ scan modes using Rigaku diffractometers and a Philips MRD diffractometer, all equipped with a CuKα radiation. The substrate peaks were used as internal standards to correct for potential alignment inaccuracies and to determine instrumental resolution. The film thickness of thin samples was determined using the Philips MRD in reflectivity mode, and this was converted to a deposition rate. Scanning electron microscopy (SEM) was carried out using a Philips FEG-SEM equipped with an Oxford energy dispersive spectroscopy (EDS) analyzer. Ex-situ AFM measurements were carried out in air using an AutoProbe CP (Park Scientific Instruments) in contact mode with ultralever tips. To minimize surface contamination resulting from long exposure to air, AFM experiments were performed immediately after deposition. GG3.27.2

3 RESULTS AND DISCUSSION The Θ 2Θ XRD scans for NdNiO 3 films deposited on [100]-oriented STO as a function of deposition temperature are shown in Figure 1. These samples were deposited for 30 minutes and are 670 Å thick. The region between 2Θ = 27 and 45 has been magnified for each diffractogram (the magnification is listed above it) to highlight the presence or absence of impurity phases. All peaks observed in the diffractograms correspond to either the (00l) peaks from the substrate (S's), the (h0h) peaks from the perovskite NdNiO 3 (stars), NiO (triangles), and Nd 2 NiO 4 (circles). At low temperatures, only the substrate and NdNiO 3 peaks are observed. At temperatures above 700 C, the films begin to develop second phases according to the decompositions reaction, 2 NdNiO 3 NiO + Nd 2 NiO 4 +1/2 O 2 (g). Further work on the NdNiO 3 films were, therefore, carried out on those samples grown at 625 C, where the relative intensities of the (101) and (202) are in best agreement with the expected bulk values. S 800 C x 20 S Intensity (Arb. Units) 750 C 700 C 650 C 625 C x 60 x 60 x 80 x C x Two Theta (deg) Figure 1. X-ray diffractograms of NdNiO 3 films deposited upon [100]-SrTiO 3 at various temperatures. See text for explanation of symbols and of the central regions. Figure 2a-c shows the effect of the substrate on the film's structure for films deposited at 625 C for 30 minutes. Figure 2b shows the calculated pattern for a [101]-oriented crystal of NdNiO 3. For both substrates, the film's (h0h) peaks are shifted from the expected position. These shifts are in accordance with the type of strain imparted by the substrate, as depicted by the cartoons in the figure. The average mismatch [16], f avg, for NdNiO 3 is 2.59 % (tensile) and 0.43 % (compressive) on STO and LAO, respectively. Hence, the biaxial substrate strain should lead to a decrease in the (h0h) spacing on STO and an increase on LAO, as observed. It appears that the peaks are shifted more on LAO than on STO, in spite of the larger mismatch on STO. This means that other relaxation mechanisms are important for films of this thickness on STO. Film thickness was determined by x-ray reflectivity measurements in a manner similar to that reported elsewhere [17]. Films deposited for 30 minutes were 670 Å thick. This corresponds to a deposition rate of 22.3 Å/ min, or 0.12 Å / laser pulse. This is a relatively slow deposition rate, which allows for high-quality growth of these materials. SEM results indicated GG3.27.3

4 Intensity (Arb. Units) (101) N (100) S (100) L (101) N (c) NdNiO 3 on LaAlO 3 (b) Calculated (h0h) NdNiO 3 (a) NdNiO 3 on SrTiO 3 (202) N (200) S (200) L (202) N Intensity (Arb. Units) g S f e d Two Theta (deg) Two Theta (deg) Figure 2. a-f: XRD scans of NdNiO 3. Figs a-c highlight the effect of substrates on film structure. Cartoons indicate the strain state of films. Figs d-g highlight the effect of thickness for films grown on SrTiO 3. The films are 670 Å (d), 1270 Å (e), and 1970 Å (f). The pattern in g is a calculated pattern. that the films had very flat smooth surfaces with some standard defects associated with the PLD method [1]. A typical SEM is shown in Figure 3a for a 670 Å thick film on STO. Increased magnifications yielded no useful contrast as to the grain structure or surface morphology. Atomic Force Microscopy was therefore used to probe the surface morphology of these films. An image is shown in Figure 3b for a film of 670 Å thick on STO. The surface is characterized by a granular structure with lateral feature sizes on the order of a couple of hundred nanometers. Over the area of a square micron, however, the surface roughness is in the order of only a few nanometers. EDS indicated that the films had the appropriate nominal cationic stoichiometry, within the experimental error. In order to ascertain the true epitaxial state of the films, we conducted rocking curves around the (202) peak and φ scans for various peaks in the films. The rocking curves for the 670 Å thick films were 0.17 and 0.22 on LAO and STO, respectively. This implies good out-of plane alignment of the (101) planes for the NdNiO 3 films. The out-of-plane orientation was confirmed using φ scans. Initially, both [101] and [010] variants were presumed to be present, but no intensity was found at any expected location corresponding to the [010] variants. On the 20 µm a b Å µm Figure 3. Scanning electron micrograph (a) and atomic force micrograph (b) of NdNiO 3 on SrTiO 3. GG3.27.4

5 (a) Variant A A A B (111) B (d) Variant B [010] F [111] F [001] F [101] F Intensity (Arb. Units) (b) (311) A 1 A A 2 B 2 2 A 1 B1 B2 B 1 52 [001]F [101] F [111] F [010] F [100] F substrate (c) * * * * substrate phi (deg) Figure 4. Schematics of variant formation for a 90 rotation in plane, (a) and (d).the phi-scans demonstrate that four 90 variants exist for NdNiO 3 on SrTiO 3. The two variants A and B are observed in (b), a phi scan of the (111) for the film. The existence of all four potential variants, A and B and their 180 rotated versions, A' and B', are confirmed in (c), a phi-scan of the (311) for the film. Peaks marked by a star arise from residual substrate intensity. other hand, strong intensity was observed for peaks related to [101] variants. Phi-scans are given in Figure 4b and 4c for the {111} and {311} family of planes of the [101]-oriented films. Sharp, narrow diffraction peaks indicate that the films are highly epitaxial. Only two peaks are expected in these phi scans from any single variant; the existence of more than two peaks implies that there are in-plane orientational variants in the film. These variants are related to each other by 90 rotations around the substrate normal. Two of the four possible variants are shown in Figure 4a and 4d. They are denoted as variant A and B, and peaks arising from these variants are marked accordingly in the (111) phi-scan. The (311) phi scan is more complex, because it discerns all four of the possible 90 variants, as denoted. The 52 separation between peaks of the A 1 and A 2 or B 1 and B 2 variants is expected in this experiment for the 180 rotations that relates them. Similar results were observed on the other films, including those deposited upon LaAlO 3. Finally, using the observed two-theta and phi values of the distinct planes observed (202), (111), (311), (210), (230), and (430) it is possible to crudely refine the unit cell parameters of the thin films. A simple refinement program was written to determine the unit-cell parameters from the 11 observables recorded (there is no observable phi value for the (202) plane because it is perpendicular to the phi-rotation axis). It was observed that the thinnest films (670 Å) on LaAlO 3 were pseudomorphic in nature; this means that their in-plane periodicities matched that of the substrate. Because of this, it was difficult to match an orthorhombic unit cell to the observed diffraction values. On the other hand, a monoclinic cell fit the data quite well, with the following parameters: a = Å, b = Å, c = Å, β = This large monoclinic distortion needs to be confirmed by electron diffraction, but it makes sense that the biaxial compressive stress leads to a closing of the angle between the [100] and [001] in the strained [101] oriented films (see Figures 4a and 4d). As the films grew thicker the lattice parameters relaxed towards the bulk structure, although the films were not completely relaxed even at 2000 Å (see Figures 2f and 2g). A monoclinic cell fit the data well for all films, with the monoclinic angle relaxing rather quickly to 90. This monoclinic distortion could have important ramifications on the physical properties. GG3.27.5

6 CONCLUSIONS High-quality, epitaxial films of [101]-oriented NdNiO 3 were be grown under certain conditions using pulsed laser deposition on LAO and STO substrates. Although the films are epitiaxial, four degenerate variants exist corresponding to 90 rotations of the films crystalline axes. The built in strain decreases as the film thickness increases. The thinnest films on LAO experience a monoclinic distortion and are pseudomorphic with the substrate. Further work is being done to characterize better the unit cell parameters of these films, as well as their physical properties. ACKNOWLEDGMENTS This work was supported in part by the MRSEC program of the National Science Foundation under Award Number DMR Partial funding for CA was provided by Carnegie Mellon's Undergraduate Research Initiative. TDD would like to thank the Lavoisier Program of the French Ministry of Foreign Affairs for Partial Support. The authors would like to thank J. Wolf and M. De Graef for their insightful comments and help with the diffraction analysis. REFERENCES 1. D. B. Chrisey and G. K. Hubler. Pulsed Laser Deposition of Thin Films (Wiley, New York, 1994). 2. P. Lacorre, J. B. Torrance, J. Pannetier, A. I. Nazzal, P. W. Wang and T. C. Huang, J. Solid State Chem. 91, p. 225 (1991). 3. J. A. Alonso, M. J. Martinez-Lope, M. T. Casais, M. A. G. Aranda and M. T. Fernández-Díaz, J. Am. Chem. Soc. 121, p (1999). 4. J. B. Torrance, P. Lacorre, A. I. Nazzal, E. J. Ansaldo and C. Niedermayer, Phys. Rev. B 45, p (1992). 5. K. M. Satyalakshmi, R. M. Mallya, K. V. Ramanatham, X. D. Wu, B. Brainard, D. C. Gautier, N. Y. Vasanthacharya and M. S. Hegde, Appl. Phys. Lett. 62, p (1993). 6. K. M. Satyalakshmi and K. B. R. V. S. Hegde, J. Appl. Phys. 78, p (1995). 7. M. L. Medarde, J. Phys. Condens. Matter 9, p (1997). 8. J. F. DeNatale and P. H. Kobrin,, edited (Materials Research Society, Warrendale, 1997), p P. Laffez, M. Zaghrioui, I. Monot, T. Brousse and P. Lacorre, Thin Solid Films 354, p. 50 (1999). 10. G. Catalan, R. M. Bowman and J. M. Gregg, J. Appl. Phys. 87, p. 606 (2000). 11. G. Catalan, R. M. Bowman and J. M. Gregg, Phys. Rev. B 62, p (2000). 12. M. A. Novojilov, O. Y. Gorbenko, I. E. Graboy, A. R. Kaul, H. W. Zandbergen, N. A. Babushkina and L. M. Belova, Appl. Phys. Lett. 76, p (2000). 13. P. C. Canfield, J. D. Thompson, S.-W. Cheong and L. W. Rupp, Phys. Rev. B 47, p (1993). 14. J. S. Zhou, J. B. goodenough, B. Dabrowski, P. W. Klamut and Z. Bukowski, Phys. Rev. B 61, p (2000). 15. A. J. Francis, A. Bagal and P. A. Salvador, in Innovative Processing and Synthesis of Ceramics, Glasses and Composites, edited by N. Bansal (The American Ceramic Society, Inc., Westerville, OH, 2000), p M. Ohring, The Materials Science of Thin Films (Academic Press, San Diego, 1992). 17. P. A. Salvador, T.-D. Doan, B. Mercey and B. Raveau, Chem. Mater. 10, p (1998). GG3.27.6