Electrically conducting oxide thin films of (Sr,Ca)RuO3 and structural compatibility with (Ba,Sr)TiO3

Electrically conducting oxide thin films of (Sr,Ca)RuO3 and structural compatibility with (Ba,Sr)TiO3

Materials Research Bulletin, Vol. 34, No. 6, pp. 933–942, 1999 Copyright © 1999 Elsevier Science Ltd Printed in the USA. All rights reserved 0025-5408...

447KB Sizes 3 Downloads 91 Views

Materials Research Bulletin, Vol. 34, No. 6, pp. 933–942, 1999 Copyright © 1999 Elsevier Science Ltd Printed in the USA. All rights reserved 0025-5408/99/$–see front matter

PII S0025-5408(99)00091-4

ELECTRICALLY CONDUCTING OXIDE THIN FILMS OF (Sr,Ca)RuO3 AND STRUCTURAL COMPATIBILITY WITH (Ba,Sr)TiO3 Z.R. Dai1, S.Y. Son2, B.S. Kim2, D.K. Choi2, and F.S. Ohuchi1,2* Department of Materials Science and Engineering, University of Washington, Seattle, WA 98195, USA 2 Department of Inorganic Materials Engineering, Hanyang University, Seoul 133-791, Korea

1

(Refereed) (Received August 5, 1998; Accepted August 6, 1998)

ABSTRACT Electrically conducting oxide thin films, CaxSr1⫺xRuO3, where x varies from 1 to 0, were prepared by radio frequency (RF) magnetron sputtering, and their structural compatibility with (Ba,Sr)TiO3 thin films was investigated. It was found that both materials crystallize into the perovskite structure, and the lattice parameter for CaxSr1⫺xRuO3 can be tuned to that of (Ba,Sr)TiO3 by adjusting the Ca/Sr ratio, so that the compatibility between the two materials at the interface is increased. Structural, chemical, and electrical properties of the thin films and the heterostructures were characterized. Cross-sectional high-resolution electron microscopy (HREM) revealed that the (Ba,Sr)TiO3 thin film was grown epitaxially on the (Ca,Sr)RuO3. The potential utility of CaxSr1⫺xRuO3 as a bottom electrode for (Ba,Sr)TiO3 is suggested. © 1999 Elsevier Science Ltd KEYWORDS: A. oxides, B. sputtering, C. electron microscopy, D. crystal structure, D. electrical properties INTRODUCTION Alkaline-earth ruthenates with perovskite-based crystal structure have been known for a long time [1,2], yet it is only recently that their electronic structures and properties have been investigated extensively [3– 6]. Of these materials, SrRuO3 and CaRuO3 (hereafter referred to as SRO and

*To whom correspondence should be addressed. 933

934

Z.R. DAI et al.

Vol. 34, No. 6

CRO, respectively) are of particular interest, since they are electrically conductive and have potential applications for underlying bottom electrodes of ferroelectric thin film materials, such as barium-strontium titanate, (Ba,Sr)TiO3 (referred to as BST). Both SRO and CRO crystallize into the perovskite structure; however, CRO has greater orthorhombic distortion than SRO, and SRO is very close to being cubic [1,7]. It has been demonstrated that an alloy of these oxides with composition SrxCa1⫺xRuO3 (referred to as SCR) forms a pseudo-cubic structure with lattice parameters varying from 0.383 to 0.393 nm when x is changed from 0 to 1 [8]. Since the lattice parameter for Ba0.5Sr0.5TiO3 is 0.392 nm, and the symmetry of the atomic arrangement for SCR is similar to that for BST [9], tunability of the lattice parameter of SCR would enhance the structural compatibility of the thin film BST/SCR heterostructure systems [10]. The purpose of our study was to investigate (Sr,Ca)RuO3 as a potential electrode material for BST thin films. In this paper, we report the preparation of SCR thin films by magnetron sputtering and their structural, chemical, and electrical properties as well as compatibility with BST thin films. EXPERIMENTAL SCR thin films were deposited on p-Si(100) substrates by means of RF magnetron sputtering from composite targets that were specifically designed and fabricated for this experiment. SCR powder was synthesized by calcining a powder mixture consisting of RuO2, SrO, and CaO at 1000°C in air. The SCR powder was then compacted by unidirectional pressing into a 3-in. diameter target. Four targets with different chemical compositions were prepared by adjusting Sr/Ca to 5/5, 7/3, 9/1, and 10/0. Thin films fabricated from these targets with compositions of Sr/Ca ⫽ 5/5, 7/3, 9/1, and 10/0 are referred to hereafter as 5/5-, 7/3-, 9/1-, and 10/0-SCR thin films, respectively. During the deposition, process parameters such as target-to-substrate distance, deposition pressure, and RF power were fixed, but the deposition gas mixture ratio Ar/O2 was varied from 9/1 to 6/4, and the substrate temperatures were varied between 300 and 550°C. In some cases, post annealing was done in a tube furnace at temperatures ranging from 500 to 700°C in O2 ambient pressure for 270 min. For BST/SCR hetero-thin films, BST thin films were subsequently deposited onto the SCR thin films in a separate RF magnetron sputtering apparatus. Deposition of the BST thin film was made from a BST composite target with the composition adjusted to (Ba ⫹ Sr)/Ti ⫽ 1.025. Deposition conditions for the BST/SCR thin films and some of their properties are summarized in Table 1. Crystallographic phases were determined using a standard X-ray diffractometer (Rigaku, RAD-C with Cu K␣). Rutherford backscattering (RBS) was performed to obtain the chemical compositions of the SCR thin films. Microstructure of the SCR thin films and the BST/SCR heterointerfaces was characterized by a Phillips 430 transmission electron microscope and a JEOL JEM-4000EX high-resolution transmission electron microscope. Electrical resistivity of SCR thin films was measured using the Van der Pauw method in Hall measurement (BIO-RAD, HL5200). Dielectric and electrical properties of the BST/SCR systems were measured using a Hewlett-Packard HP-4145B. RESULTS AND DISCUSSION Several deposition parameters influence the physical properties of SCR thin films. Degree of crystallization, preferred orientation, and electrical resistivity are closely related to the substrate temperature and the Ar/O2 ratio during the deposition.

Vol. 34, No. 6

RUTHENATE TITANATE FILMS

935

TABLE 1 Deposition Conditions for the BST/SCR Thin Films and Their Properties BST Target composition Substrate temperature Ar/O2 ratio Working pressure RF power Target-substrate spacing Background pressure Film thickness Film composition

Lattice parameter Electrical resistivity

(Ba ⫹ Sr)/Ti ⫽ 1.025 550°C 1/1

250 nm Ba0.45Sr0.55TiO3

0.39 nm NC

SCR Sr/Ca 5/5

Sr/Ca 7/3 Sr/Ca 9/1 500°C 9/1 1.33 Pa 130 W 3.5 cm ⬍4 ⫻ 10⫺4 Pa 300 nm 300 nm 300 nm Sr0.43 Sr0.61 NA Ca0.57 Ca0.39 RuO3 RuO3 0.387 nm 0.391 nm 0.392 nm 340 ⍀cm 300 ⍀cm 275 ⍀cm

Sr/Ca 10/0

300 nm NA

0.393 nm 240 ⍀cm

Shown in Figure 1 is a series of X-ray diffraction patterns taken from SCR thin films that were deposited from a Sr/Ca ⫽ 7/3 target at various substrate temperatures. The Ar/O2 ratio was fixed at 9/1. Two peaks appear at 2␪ ⫽ ⬃32 and ⬃57° which correspond to {110} and {211} reflections of the SCR thin films, indicating that the thin films were preferentially grown in two directions. While the relative peak intensity between the {110} and {211} reflections and their widths varied with deposition temperature, thin films with a high degree of crystallinity were obtained when deposited above 500°C. Below 500°C, thin films were basically oriented with the {110} planes parallel to the substrates. Electrical resistivity of the

FIG. 1 A series of XRD patterns of SCR thin films deposited on p-Si(100) substrates from the Sr/Ca ⫽ 7/3 target as a function of substrate temperature.

936

Z.R. DAI et al.

Vol. 34, No. 6

FIG. 2 Effect of substrate temperatures on resistivity of the 7/3-SCR thin films during deposition.

SCR thin films varied considerably with deposition temperature; however, the resistivity values became minimal when the deposition was performed above 500°C (Fig. 2). Similar trends in terms of the degree of crystallinity, preferred orientation, and electrical resistivity were observed in other SCR thin films. The appearance of the {211} peak together with the {110} peak in X-ray diffraction patterns indicates that the SCR thin films were not grown with a single orientation. Structurally, SCR thin films were favorably grown with their {110} planes parallel to the substrate because of the closed-pack atom configuration. A cross-sectional HREM image was obtained from a 5/5-SCR thin film specimen grown at 500°C (Fig. 3). These observations revealed that the {110} planes of SCR were several degrees off from the (100) plane of Si, and there was an amorphous layer between the substrate and SCR film. In some areas of the SCR film, it was found that the {100} planes were off by ⬃50° from the substrate. This orientation was equivalent to the {211} planes of the SCR thin films parallel to the (100) of Si, since the angle between the {110} and {211} planes should be 50.8°. The effect of the Ar/O2 ratio and temperature on the electrical resistivity during annealing was investigated next. The data collected from 7/3-SCR thin films (Fig. 4) shows that the electrical resistivity increased noticeably as the Ar/O2 ratio changed from 9/1 to 6/4 during the annealing, the lowest resistivity was observed at Ar/O2 ⫽ 9/1. Although not shown here, thin films annealed in pure Ar, or Ar/O2 ⫽ 1/0, showed considerably higher resistivity. As for the annealing temperatures, only a slight change was observed for temperatures higher than 500°C. X-ray diffraction indicated no significant improvement in the crystallinity of SCR thin films with annealing. The SCR thin films reported hereafter were deposited at a substrate temperature of 500°C with the Ar/O2 ratio of 9/1. No further annealing was made.

Vol. 34, No. 6

RUTHENATE TITANATE FILMS

937

FIG. 3 Cross-sectional HREM showing microstructure of the 5/5-SCR thin film deposited on Si(100) substrate at 550°C.

Chemical compositions of the SCR films were analyzed by Rutherford backscattering spectroscopy. Due to peak overlap of Sr (atomic number Z ⫽ 38) and Ru (Z ⫽ 44) in the RBS spectra, a simulation program, RUMP [11] was used to deduce the chemical compositions. Analyses yielded Sr0.43Ca0.57RuO3 and Sr0.61Ca0.39RuO3 for the 5/5- and 7/3-SCR thin films, respectively. Although the chemical compositions of thin film and target were reasonably close to each other, there are differences due to different sputtering yields of the elements. No RBS data were collected from the 9/1-SCR thin film. A pseudocubic lattice constant is an important parameter when considering lattice matching of SCR with (Ba,Sr)TiO3 in hetero-thin film systems. Lattice constants, a, were deduced from 2␪ values for the (211) peaks of the X-ray diffraction patterns; they were 0.387, 0.391, 0.392 m and 0.393 nm for the 5/5-, 7/3-, 9/1-, and 10/0-SCR thin films, respectively, and 0.392 nm for BST. These values are plotted in Figure 5 as a function of stoichiometry, x, of SrxCa1⫺xRuO3. While the exact stoichiometry value for x ⫽ 0.9 is not known, a data point corresponding to x ⫽ 0.9 was nevertheless plotted. Data for single crystals of CaRuO3 (x ⫽ 0) and SrRuO3 (x ⫽ 1) [6], sintered polycrystalline specimens of SrxCa1⫺xRuO3 [12], and Sr0.5Ca0.5RuO3 thin films [8] are shown for

938

Z.R. DAI et al.

Vol. 34, No. 6

FIG. 4 Effect of Ar/O2 ratio and temperature on resistivity of the 7/3-SCR thin films during annealing.

comparison. It was found that the pseudocubic lattice constant, a, does not follow a linear relation with the stoichiometry of SCR, but exhibits an s-shaped curve. The reason for its not following a linear relation is suggested later. It has been reported [12] that the (110) peak from the similar systems, SrxCa1⫺xRuO3, splits into two peaks as x decreases below 0.4, implying that orthorhombic distortion occurs, whereas the structure becomes more pseudocubic-like for x ⬎ 0.4. Electrical resistivity data are also shown in Figure 5. Data from the single crystals of CRO and SRO [6] were plotted for comparison. One of the questions regarding the structure of SrxCa1⫺xRuO3 is the location of Ca and Sr cations in the pseudocubic unit cell. In other words, when the SrxCa1⫺xRuO3 structure is formed, are the Sr and Ca ions distributed uniformly within the unit cell, or do Sr and Ca cations separately form the SrRuO3 and CaRuO3 micro-domains, respectively? For an ideal ABO3-type cubic perovskite structure, cation B coordinates with oxygen anion forming BO6 octahedron and B is located at the center of the octahedron. The octahedra connects each other by corner sharing, leading to the formation of cuboctahedra cages, with cation A occupied in the center of the cuboctahedra. For SrxCa1⫺xRuO3 that crystallizes into the pseudocubic perovskite structure, the coordination polyhedra may be distorted. As Ru occupies the B site, Sr and Ca occupy the A sites, but their arrangement may be ordered or random. Figure 6(a) and (b) show two microdiffraction patterns collected from different regions of the 5/5-SCR film (or Sr0.43Ca0.57RuO3), where the zone axes are [001] and [011], respectively. The electron beam size used was about 40

Vol. 34, No. 6

RUTHENATE TITANATE FILMS

939

FIG. 5 A plot of the pseudocubic lattice parameters, a(110), and electrical resistivities as a function of stoichiometry, x, of SrxCa1⫺xRuO3.

nm in diameter. The intensity of the {100} reflections in the pattern shown in Figure 6(a) was very faint, whereas the identical reflections observed in Figure 6(b) were very strong. Since the diffraction intensity is determined by atomic species and symmetry of the crystal, i.e., atomic configuration in the crystal, therefore, the difference in the {100} reflection intensities implies that the chemical composition might not be uniform from area to area within the thin film. To understand the variation of the {100} reflection intensities for two different regions, a kinematical analysis was made. For a unit cell of the perovskite structure ABO3, where A is either Sr or Ca and B is Ru, if the A atom is located at the (000) position, the B atom is at the (1/2 1/2 1/2) position and O atoms are located at (1/2 1/2 0), (1/2 0 1/2) and (0 1/2 1/2) positions. The structure factor can be found from F(hkl) ⫽ f( A) ⫹ f(B)䡠exp{i(h ⫹ k ⫹ l)}⫹f(O)䡠{exp[i(h ⫹ k)] ⫹ exp[i(k ⫹ l)] ⫹ exp[i(l ⫹ h)]}

(1)

where f(A), f(B), and f(O) are the atomic scattering factors corresponding to A, B, and O, respectively, and h, k, and l are the Miller indices of the diffraction plane. The diffraction

940

Z.R. DAI et al.

Vol. 34, No. 6

FIG. 6 Microdiffraction patterns obtained from different regions of the 5/5-SCR film (a) zone axis [001] and (b) zone axis [011]. intensity is proportional to the value of ⱍF(hkl)ⱍ2. The structure factor corresponding to {100} reflection, therefore, is F(100) ⫽ f(A) ⫺ f(B) ⫺ f(O)

(2)

where B represents Ru and A is either Sr or Ca. Obviously, the structure factor F(100) changes depending on whether the A positions are occupied by Ca or Sr atoms, or both. If Sr and Ca randomly occupy the A sites and are uniformly distributed in the SCR film, the value of F(100) should be the same at any area of the SCR film. In other words, diffraction intensity of the {100} crystal plane should not vary from one area to another. For the case of sin␪/␭ ⫽ 0.5, where ␪ is the scattering angle and ␭ is the wavelength of electron, the atomic scattering factors are f(Ca) ⫽ 1.07, f(Sr) ⫽ 1.80, f(Ru) ⫽ 2.03, and f(O) ⫽ 0.54 [13]. The value of f(Ca) is nearly one-half of f(Ru), whereas f(Sr) is close to f(Ru). From eq. 2, a maximum value for F(100) occurs if the A sites are occupied by Ca, not by Sr, or if the A sites are occupied by Sr only, F(100) becomes minimal. From the pattern shown in Figure 6(a), therefore, the faint {100} reflection indicates that the A sites are predominantly occupied by Sr in the area from which the diffraction pattern was taken. In the pattern of Figure 6(b), however, the strong {100} reflection is attributed to the A sites being predominantly occupied by Ca in corresponding areas. Therefore, it is suggested that Sr and Ca are not uniformly distributed in the SCR film, and that Sr-rich and Ca-rich regions co-exist. Structurally, the SCR film probably consists of mixed SrRuO3 and CaRuO3 domains. This might be why the lattice constant does not follow a linear relation with the stoichiometry of SCR but exhibits the s-shaped curve. X-ray diffraction data obtained from the BST/SCR hetero-thin films showed that the Bragg peak positions corresponding to the (110) and (211) planes for SCR thin films move successively toward the positions corresponding to those for BST thin films as the SCR composition varies from 5/5 to 7/3 to 9/1. Lattice mismatching of the BST with the 5/5- and 7/3-SCR thin films are 1.29% and 0.26%, respectively, and nearly perfect lattice matching is attained for 9/1-SRC thin film with BST.

Vol. 34, No. 6

RUTHENATE TITANATE FILMS

941

FIG. 7 A cross-sectional HREM image obtained from the BST/SCR(Sr/Ca ⫽ 7/3) sample.

A cross-sectional HREM image obtained from the BST/SCR (Sr/Ca ⫽ 7/3) sample is shown in Figure 7. The microstructure of BST is found to be relatively uniform; no significant contrast in the micrograph is observed. The SCR thin film, however, exhibits small irregular domains with an average size of 2–3 nm, which are distributed in the matrix near the interface. The crystalline structure of these domains is severely disordered. Away from the interface, very small irregular domains are present in the film. The cause of these irregular domains near the interface is not known exactly at this point; however, it might be associated with compositional segregation and elemental diffusion, in particular that of Ca, at the interface. A close look at the HREM micrograph indicates that lattice fringe lines run across the interface between BST and SCR and are smoothly connected to each other. The spacing between the fringe lines observed in BST thin films is only slightly larger than that observed in SCR thin films. Analyses showed that the observed lattice fringes are from a set of the (110) planes for both BST and SCR, and the BST lattices are epitaxially grown over the SCR lattices.

942

Z.R. DAI et al.

Vol. 34, No. 6

Electrical properties of the BST/SCR thin film systems were characterized by fabricating an Al/BST/SCR structure with Al as an upper electrode and SCR as a bottom electrode. Current-voltage characteristics were measured using this configuration to look at the variations of the leakage current densities as a function of applied voltage. The leakage currents measured from the BST/SCR structures with reverse bias showed systematic decreases for improving lattice matching conditions, whereas the forward currents (positive bias applied to the top aluminum electrode) were in the order of low 10⫺7A/cm2 at 2.5 V (105 V/cm). The leakage current densities at 0.1 MV/cm were found to be most improved for the 9/1-SCR thin film, where a nearly perfect lattice matching condition was attained. CONCLUSIONS A series of thin films of SrxCa1⫺xRuO3 with pseudocubic structure was prepared by reactive magnetron sputtering in the mixture of Ar and O2. Electrical resistivity was minimal when the deposition was made above 500°C at Ar/O2 ratio of 9/1. A pseudocubic lattice parameter does not follow a linear relation with the stoichiometry but exhibits an s-shaped curve. Microdiffraction analysis suggested that Sr and Ca were not uniformly distributed in the SrxCa1⫺xRuO3 thin films, but the film consisted of mixture of SrRuO3 and CaRuO3 micro-domains; this could be a reason for its not following the linear relation. Tunability of the lattice parameters through judicious selection of the Sr/Ca ratios in SCR thin film compositions enhances the structural and electrical compatibility with (Ba,Sr)TiO3 dielectric thin films. ACKNOWLEDGMENTS This work was supported by the National Science Foundation “US-Korea International Program” and the Korean Science and Engineering Foundation. REFERENCES 1. J.J. Randall and R. Ward, J. Am. Chem. Soc. 81, 2629 (1959). 2. A. Callaghan, C.W. Moeller, and R. Ward, Inorg. Chem. 5, 1572 (1966). 3. Q. Gan, R.A. Rao, C.B. Eom, J.L. Garrett, and M. Lee, Appl. Phys. Lett. 72, 978 (1998). 4. Q.X. Jia, F. Chu, C.D. Adams, X.D. Wu, M. Hawley, J.H. Cho, and A.T. Findikoglu, J. Mater. Res. 11, 2263 (1996). 5. I.I. Mazin and D.J. Singh, Phys. Rev. B 56, 2556 (1997). 6. M. Shepard, S. McCall, G. Cao, and J.E. Crow, J. Appl. Phys. 81, 4978 (1997). 7. W. Bensch, H.W. Schmalle, and A. Reller, Solid State Ionics 43, 171 (1990). 8. C.B. Eom, R.J. Cava, R.M. Fleming, J.M. Phillips, R.B. van Dover, J.H. Marshall, J.W.P. Hsu, J.J. Krajewski, and W.F. Peck, Jr., Science 258, 1766 (1992). 9. W.D. Kingery, H.K. Bowen, and D.R. Uhlmann, Introduction to Ceramics, 2nd ed., p. 69, Wiley, New York (1976). 10. S.Y. Son, B.S. Kim, D.K. Choi, D.S. Lee, Z.R. Dai, and F.S. Ohuchi, J. Korean Phys. Soc. 32, S1517 (1998). 11. L.R. Doolittle, Nucl. Instrum. Methods B9, 344 (1985). 12. A. Kanbayashi, J. Phys. Soc. Jpn. 44, 108 (1978). 13. B.K. Vainshtein, Structure Analysis by Electron Diffraction, transl./ed. E. Feigl and J.A. Spink, pp. 401– 406, Macmillan, New York (1964).