Strain effect on the tilt angle in manganite thin films

Viewer
Transcript

Nuclear Instruments and Methods in Physics Research B 238 (2005) 259–263 www.elsevier.com/locate/nimb

Strain eﬀect on the tilt angle in manganite thin ﬁlms Narcizo M. Souza-Neto a

a,b

, Aline Y. Ramos

a,c,*

, He´lio C.N. Tolentino

a

Laborato´rio Nacional de Luz Sı´ncrotron – LNLS, P.O. Box 6192, 13084-971 Campinas, Sa˜o Paulo, Brazil b Dept. de Fı´sica dos Materiais e Mecaˆnica, DFMT-IF-USP, Sa˜o Paulo, SP, Brazil c Laboratoire de Mine´ralogie-Cristallographie de Paris, LMCP, UMR 7590, CNRS, Paris, France Available online 22 August 2005

Abstract We report on an angular resolved EXAFS study in the plane of La0.7Sr0.3MnO3 thin ﬁlms epitaxially grown by pulsed laser deposition on slightly mismatched substrates. EXAFS data collected in plane for tensile substrate show a change in the MnAO average bond distance and the increase of MnAMn length matching with the enlargement of the cell parameter. From these results, we conclude that there is no signiﬁcant change in the tilt angle MnAOAMn. Our observations suggest that the distortion within the octahedra is a key parameter in the strain-dependence of the transport and magnetic properties in manganite ﬁlms. 2005 Elsevier B.V. All rights reserved. PACS: 75.47.Lx; 68.55. a; 78.70.Dm; 31.15.Ar Keywords: EXAFS; Manganites; Films; Multiple scattering

Discovered about 10 years ago, the colossal magnetoresistance has already found technological applications including computer hard disk drives. Manganite thin ﬁlms are materials with potential application in magnetoresistive sensors and magnetic random access memory. In LaxA1 xMnO3 manganites (where A is a divalent alkaline earth *

Corresponding author. Address: Laborato´rio Nacional de Luz Sı´ncrotron – LNLS, P.O. Box 6192, 13084-971 Campinas, Sa˜o Paulo, Brazil. Tel.: +55 1932874520; fax: +55 1932874632. E-mail address: [email protected] (A.Y. Ramos).

ion), the largest magnetoresistance is found near the charge ordering temperature TC, whose value is primarily dependent, in bulk material on the A-site substitution, acting as a hydrostatic pressure and modifying the tilt angle MnAOAMn between connected octahedra. In epitaxial thin ﬁlms, biaxial strain has strong eﬀects on the change in TC and in the metal–insulator transition temperature TMI. Several publications have reported that both TC and TMI can vary with the ﬁlm thickness, as well as with the ﬁlm mismatch with the underlying substrate. This is generally understood as a consequence of the modiﬁcation of the electron hoping

0168-583X/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.06.059

260

N.M. Souza-Neto et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 259–263

by change in local organization at the magnetically active site, in the tilt angle MnAOAMn and/or in the MnAO bond length. However, no full consensus is obtained about the relative importance of each local scale eﬀect. We investigated the strain-induced distortions around the Mn ions by X-ray absorption spectroscopy in La0.7Sr0.3MnO3 thin ﬁlms. A quantitative analysis of the XANES spectra, using the semiempirical NatoliÕs rule, has shown that the distance variations are directly correlated to the macroscopic strain. This result is conﬁrmed by EXAFS analysis of the contribution of the coordination shell. In order to make clear whether or not the tilt angle is also modiﬁed, the analysis of the next nearest neighbors contribution should also be performed from X-ray absorption data in the EXAFS range. As multiple scattering has to be considered in the data analysis, some modelizations should be performed to disentangle the various high frequency contributions to the EXAFS signal. We present here this EXAFS study and the interpretation of the results. La0.7Sr0.3MnO3 crystallize in the rhombohedral R 3c variant of the cubic pervoskite, with La and Sr randomly distributed on the A-site (6A positions) [1]. A 60 nm thin ﬁlm (STO60) has been epitaxially grown in the [0 0 1] direction by pulsed laser deposition under tensile SrTiO3 [0 0 1] substrate with cubic structures. The small lattice mismatch between LSMO and SrTiO3 allows a pseudomorphic growth and the ﬁlms are fully constrained for thickness below the critical thickness of 100 nm. A thick sample (LAO300) was also used as relaxed reference ﬁlm. The Mn K edge X-ray absorption experiments were performed at the D04B-XAS beamline of the Brazilian synchrotron light laboratory (Laborato´rio Nacional de Luz Sı´ncrotron, LNLS) [6], with a Si(1 1 1) channel-cut monochromator. The data were collected in the ﬂuorescence mode using a Ge 15-elements solid state detector. Further details about the samples and the experiments setup are given elsewhere [2,5,4]. EXAFS signal for the La0.7Sr0.3MnO3 polycrystal and ﬁlms are shown in Fig. 1. The k2-weighted Fourier transform of the EXAFS signal is shown in the Fig. 2 for the La0.7Sr0.3MnO3 bulk reference ˚ compounds. The main peak (A) at around 1.4 A

a

b

c

Fig. 1. Mn K edge EXAFS measurements: (a) La0.7Sr0.3MnO3 ﬁlm under tensile misﬁt strain, (b) relaxed La0.7Sr0.3MnO3 ﬁlm, (c) polycrystalline bulk La0.7Sr0.3MnO3 sample.

Fig. 2. Fourier transform of the EXAFS signal in bulk La0.7Sr0.3MnO3 sample: modulus (solid line) and imaginary part (points). The peak labelled A corresponds to coordination shell (MnAO distance). The contribution of the single scattering of the metal shell (MnAMn) is included in the peak C.

corresponds to the oxygen coordination shell. This peak was back-transformed over the R-range 0.9– ˚ . MnAO amplitude and phases we extracted 1.9 A from spectrum of the polycrystalline sample and used to ﬁt the back-transformed signal in the ﬁlm

N.M. Souza-Neto et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 259–263

261

Table 1 Structural parameters obtained from the EXAFS analysis, for relaxed ﬁlm (LA300) and the ﬁlm under in-plane expansion (STO60) ˚) ˚ 2) N R (A r2 · 104 (A DE0 (eV) Coordination shell (MnAO) STO60 1.3 ± 0.5 2.7 ± 0.5 LAO300 4 (ﬁxed)

1.94 ± 0.02 2.05 ± 0.02 1.95 ± 0.01

17 ± 5 17 ± 5 25 ± 5

1.4 ± 0.2 1.4 ± 0.2 1.4 ± 0.2

Mn next neighbors (MnAMn) STO60 7±1 LAO300 6±1

3.890 ± 0.005 3.875 ± 0.005

25 ± 5 13 ± 5

0.4 ± 0.2 0.4 ± 0.2

samples. For the relaxed ﬁlm the coordination shell can be described as a single shell of six O at ˚ – as in the bulk compound – around 1.95 A whereas in the constrained ﬁlm an additional con˚) tribution of a long MnAO distance (2.05 A should be introduced (Table 1). The average MnAO in-plane distance is larger in the strained ﬁlm as compared to the relaxed sample. This accounts for the enlargement of the basal MnO4 square established by XANES spectroscopy. We should note, however, that in the EXAFS analysis, the variation in MnAO distance is calculated from the diﬀerence between values obtained with large uncertainties, while in the XANES study it is directly obtained from the energy shift at the edge. The best accuracy is then obtained from the XANES measurements, giving an increase of 1% in the MnAO distance between strained and relaxed ﬁlm [5,4]. At higher distance in the FT, the peaks (B and C) correspond to larger eﬀective distances, including the A-site and nearest-Mn neighbors single scattering (SS) contributions, as well as a large number of the multiple scattering (MS) contributions. In crystalline perovskite compounds all these contributions must a priori be taken into account together. We use ab initio simulations in the MS expansion using FEFF 7.00 code [3] on the basis of the crystallographic structure to evaluate how the analysis of the Mn next nearest neighbors can be simpliﬁed. In the La0.7Sr0.3MnO3 compound, the La and Sr atoms are considered as randomly distributed on the A-site. A structural model for random distribution is not easy to build and a simulation with a partial but deﬁned substitution of La by Sr would overestimate the actual weight of some multiple scattering paths. The sim-

ulations were then performed using the structure from [1], and considering only La on the site A. We analyzed the diﬀerent contributions path by path to disentangle the more signiﬁcant ones in each part of the Fourier transform. We checked then the validity of our conclusions for Sr on this site. Finally, we evaluated if the contributions involving Sr and La atoms would cancel or reinforced each other. The MnAMn SS contribution is included in the ˚ (Fig. 3). For further analpeak C at around 3.5 A ysis of the spectral content of this peak, we neglect all contributions with weight lower than 5% as well as those with eﬀective paths length lower than 3.0 ˚ . In addition we discard the and higher than 4.4 A

Fig. 3. Calculated main contributions to the peak C: MnAMn SS (plain line), MnAOAMn MS (dashed line) and La in A-site (dashed + open dot line). The dotted line corresponds to the A-site neighbors when considering a weighted average of the contribution of Sr and La atoms at this site. Inset: EXAFS signal of these two contributions, out of phase over the experimental k-range.

N.M. Souza-Neto et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 259–263 Mn Mn

173.2deg 13.7deg

O

-2

low-k part of the EXAFS signal, restricting the FT ˚ 1. In these conditions, beover the range 4–13 A sides the MnAMn single scattering, only two ˚ and 3.5 A ˚ , give signiﬁcant paths, centered at 3.2 A contributions to the C peak (Fig. 3). The contribu˚ is due to SS scattering of the La tion at 3.2 A atoms in the A-site. This contribution has been compared to that one given for a cluster with Sr on the A-site. The two signals are found almost perfectly out of phase (Fig. 3, inset) giving opposite contributions to the Fourier transform. In the mixed compound the total contribution can be approximated by a weighted average of these contributions resulting in lowering of the total contribution (Fig. 3). If we restrict our analysis ˚ , the signal can be considered as above R = 3.3 A almost free of A-site contribution. The peak cen˚ , results from three-core MS involvtered at 3.5 A ing two Mn and their common oxygen. It is closely related to the MnAMn SS path: this path is one of its legs and the other two legs are MnAO legs with a focusing angle close to 180. The eﬀective length of this paths is then very close the MnAMn distance and the variations of the lengths of the MS and SS paths are closely related. An increase in the MnAMn bond lengths will result, in a good approximation, to an identical increase in the eﬀective length of the MS path. This is an important statement because, in our conditions ˚ and of EXAFS analysis – back FT over 3.3–4 A 1 ˚ k-range around 4–13 A , the number of independent data points is limited around 4. It would be impossible to introduce two sets of independent parameters. We used the SS path and the MS path, with the associated parameters equalized, to ﬁt the EXAFS signal selected by back FT in the range ˚ , in the bulk La0.7Sr0.3MnO3 reference 3.3–4 A sample. The ﬁt are performed as well in k and R. As far as the phase of the signal is the main concern, the ﬁts are nice (Fig. 4). The value found ˚ ) is in a reasonfor the distance MnAMn (3.880 A able agreement with the crystallographic data and the value of DE0 3 eV is satisfactory. It was also veriﬁed, using simulated spectra, a good sensitivity of the ﬁt to a variation by a few percents of the MnAMn distance. Finally, this ﬁtting procedure has been applied to the ﬁlms (Table 1). The MnAMn distances are

k2 χ (k) / Å

262

0.0

4

6

8

10 k/ Å-1

12

14

Fig. 4. Fit of the EXAFS signal corresponding to the peak C ˚ ), using SS and MS with identical ﬁtting A ˚ (in the range 3.3–4 A parameters. Inset: MS path geometry (from crystallographic data).

determined with small relative uncertainties ˚ ). In the relaxed ﬁlms, the MnAMn dis(0.005 A ˚ , as it is in the bulk compound. In tance is 3.875 A ˚ , i.e. inthe tensile ﬁlm this distance is 3.890 A ˚ creased by about 0.020 A in comparison to bulk compound and the relaxed ﬁlms (Table 1). This corresponds to a relative increase of about 1%, matching the cell parameters, and also of the order of the increase of the MnAO distance. Consequently the tilt angle MnAOAMn is not or very weakly-modiﬁed. Due to the accuracy on the distance determination and the large angles involved (around 165), we cannot conclude about possible small variations of tilt angle. A variation in this angle by about 2–3 is indeed compatible with the accuracy of the results. However, this limited variation suggest that it is neither the only nor the main driving factor of the strain-dependence of the magnetic and transport in manganite ﬁlms.

Acknowledgements This work is partially supported by LNLS/ ABTLuS/MCT and FAPESP (1999/12330-6). N.M.S.N. and A.Y.R. acknowledge, respectively, CAPES and CNPq ﬁnancial support.

N.M. Souza-Neto et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 259–263

References [1] P.G. Radaelli, G. Iannone, M. Marezio, H.Y. Hwang, S.-W. Cheong, J.D. Jorgensen, D.N. Argyriou, Phys. Rev. B 56 (1997) 8265. [2] L. Ranno, A. Llobet, R. Tiron, E. Favre-Nicolin, Appl. Surf. Sci. 188 (2002) 170. [3] J.J. Rehr, R.C. Albers, S.I. Zabinsky, Phys. Rev. Lett. 69 (1992) 3397.

263

[4] N.M. Souza-Neto, A.Y. Ramos, H.C.N. Tolentino, E. Favre-Nicolin, L. Ranno, Appl. Phys. Lett. 83 (2003) 3587. [5] N.M. Souza-Neto, A.Y. Ramos, H.C.N. Tolentino, E. Favre-Nicolin, L. Ranno, Phys. Rev. B 70 (2004) 174451. [6] H.C.N. Tolentino, A.Y. Ramos, M.C.M. Alves, R.A. Barrea, E. Tamura, J.C. Cezar, N. Watanabe, J. Synchrotron Radiat. 8 (2001) 1040.

Local anisotropy in strained manganite thin films