Strain effects in charge-ordered Pr0.5Ca0.5MnO3 manganite thin films

Journal of Physics and Chemistry of Solids 64 (2003) 1665–1669 www.elsevier.com/locate/jpcs

Strain effects in charge-ordered Pr0.5Ca0.5MnO3 manganite thin films W. Prellier*, E. Rauwel Buzin, B. Mercey, Ch. Simon, M. Hervieu, B. Raveau Laboratoire Crismat, CNRS UMR 6508, 6 Bd du Mare´chal Juin, Caen Cedex F-14050, France

Abstract Thin films of the charge-ordered (CO) compound Pr0.5Ca0.5MnO3 have been grown by utilizing the Pulsed Laser Deposition technique. Films are deposited onto LaAlO3 and SrTiO3 substrates in order to check the effect of strains (compression and tensile). Using various techniques of characterization, we show that the strains of substrate influence the lattice parameters, the orientation of the orthorhombic structure, the transport properties and the stability of the CO state. q 2003 Elsevier Ltd. All rights reserved. Keywords: C. X-ray diffraction

Recently, perovskite type manganites such as RE12xAxMnO3 (RE ¼ rare-earth and A ¼ alkaline-earth ion), have renewed interest due to their properties of colossal magnetoresistance (CMR), a huge decrease in resistance when applying a magnetic field [1– 5]. Since most of the technological applications require thin films, it is essential to understand the effects of the substrate-induced strains on the properties of these manganites. These materials are very sensitive to the strains and even very small perturbations may result in observable effects on the properties. These effects have indeed been studied experimentally on various compounds [6– 11]. Moreover, among all the properties of these manganite materials, the phenomenon of chargeordering (CO) is probably one of the most remarkable effects. It appears for a certain value of x and particular average A-site cation radius. It corresponds to an ordering of the charges in two different Mn sublattices. In fact, the metallic state becomes unstable, below a certain temperature ðTCO Þ and the material goes to an insulating state. TCO decreases with increasing field and the insulating CO state can be totally suppressed by the application of an external magnetic field [12– 14]. As an example, a 25 T magnetic field is required to melt the CO state in the bulk Pr0.5Ca0.5MnO3 [15]. In the present work, we have studied the effects of the strains upon the structural and physical properties of Pr0.5Ca0.5MnO3 (PCMO) thin films grown on LaAlO3 * Corresponding author.

(LAO) and SrTiO3 (STO) using the Pulsed Laser Deposition (PLD) technique. We have particularly considered the changes in the lattice parameters, the orientation and the transport measurements of the film. On the basis of these results, we determined a temperature-field phase diagram for CO thin films and compared it to the bulk material. Thin films of PCMO were grown in situ using the PLD technique on (100)-SrTiOr (cubic with a ¼ 0:3905 nm) and (100)-LaAlO3 (pseudocubic with a ¼ 0:3789 nm) substrates. Detailed optimization of the growth procedure was completed and described previously [16,17]. The structural study was carried out by X-ray diffraction (XRD) using a Seifert XRD 3000P for the Q – 2Q scans and a Philips MRD X’pert for the in-plane measurements (Cu Ka, l ¼ 0:15406 nm). The in-plane parameters were obtained from the ð103ÞC reflection (where c refers to the ideal cubic perovskite cell). An electron microscope JEOL 2010 was used for the electron diffraction (ED) study. Resistivity ðrÞ was measured by a four-probe method with a Quantum Design PPMS and magnetization (M) was recorded using a Quantum Design MPMS SQUID magnetometer as a function of the temperature ðTÞ and the magnetic field ðHÞ: The composition of the films was checked by energydispersive scattering analyses. It is homogenous and corresponds exactly to the composition of the target (i.e. Pr0.5Ca0.5MnO) in the limit of the accuracy. The structure of bulk PCMO is orthorhombic (Pnma) with a ¼ 0:5395 nm; b ¼ 0:7612 nm and c ¼ 0:5403 nm [18]. Fig. 1 shows a typical Q – 2Q scan recorded for a film

0022-3697/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0022-3697(03)00253-1

1666

W. Prellier et al. / Journal of Physics and Chemistry of Solids 64 (2003) 1665–1669

Fig. 1. Room temperature Q – 2Q XRD pattern of a 75 nm film: (a) on STO, (b) on LAO. Note the high intensity and the sharpness of the peaks.

˚ ), respectively, on STO (a) and LAO (b). Note that of (750 A the same type of pattern is obtained for all of the films. This shows that the films are single phases and highly crystallized. From the XRD pattern, the out-of-plane lattice parameter is, respectively, calculated to be 0.378 nm on LAO and 0.385 nm on STO indicating that the film is under compression on LAO and under tension on STO as already reported. Fig. 2 shows the ED pattern recorded along a direction perpendicular to the substrate plane. It is clear that the orientation of the films, with respect to the substrate are different. In fact, the film is (010)-oriented, i.e. with the (010) axis perpendicular to the substrate plane [16] when grown on STO and (101)-oriented on LAO. This surprising orientation results from the lattice mismatch. Indeed, the mismatch (s) between the film and the substrate can be evaluated using the formula s ¼ 100 £ ðaS 2 aF Þ=aS (where aS and aF ; respectively, refer to the lattice parameter of the substrate and the film). The smaller mismatch on LAO is obtained for a (010)-axis in the plane ðsLAO ¼ 20:4%Þ; i.e. (101)-axis perpendicular to the substrate plane ðsLAO ¼ 20:7% for a (101) axis in the plane). In contrast, the smaller mismatch on SrTiO3 ðsSTO ¼ 2:2%Þ is found for a (101)axis in the plane (sSTO ¼ 2:5% for a (010) axis in the plane). Thus, a (010)-axis normal to the surface of the substrate as found experimentally [16,17]. The samples were cooled down to 92 K. The diffraction patterns recorded at low temperature in the strained film systematically exhibit a system of extra reflections characteristic of an incommensurate CO parallel to the ap direction on both substrates (not shown). The q values of

Fig. 2. Room temperature electron diffraction of PCMO on STO (a), on LAO (b).

W. Prellier et al. / Journal of Physics and Chemistry of Solids 64 (2003) 1665–1669

the modulation vector are close to 0.45 and STO and 0.38 on LAO and do not vary with the temperature below TCO contrary to what observed is in the bulk samples. This modulation disappears at TCO : Small film fragments separated from the substrate were also observed. In this case, a modulation vector close to 0.5 was observed. This value is similar to that of the bulk samples [19]. Since the composition is the same, this reinforces the assumption that the deviation from the ideal q value originates from the strain-induced by the substrate. This demonstrates that the orthorhombic distortion induced by the CO ordering is strongly reduced by the presence of the substrate, preventing the locking of the q value of the CO on the 1/ 2 value contrary to the bulk sample of the same composition, leading to q ¼ 0:45 for the film grown on STO and q ¼ 0:38 for the one deposited on LAO. The rðTÞ curves are presented in Fig. 3. Different magnetic fields (0, 5, 6 and 7 T) were applied perpendicular to the substrate. The behavior of both films is very different. Clearly, the film grown on STO (Fig. 3(b)) shows an insulating behavior in the whole temperature range when the magnetic field is below 4 T. A small anomaly in the resistivity (arrows around 240 K scarcely visible for this scale) is associated to the CO. But the most interesting feature is observed when applying a magnetic field of

Fig. 3. rðTÞ under different magnetic field (0, 5, 7 and 9 T) applied in the plane of the substrate for different films: (a) on LAO, (b) on STO.

1667

7 T. The film is becoming metallic a low temperature and shows a insulator-to-metal transition ðTIM Þ at about 120 K. The TIM occurs at a higher temperature (140 K) indicating a clear hysteresis. This melting field of 7 T is much lower than the 20 T one which is necessary to induce the metallic phase in the bulk sample. On the contrary, this effect is not observed for (101)-PCMO films grown on LAO (Fig. 3(a)) with field up to 7 T. Fig. 4 shows the temperature dependence with different applied magnetic fields (0, 5 and 9 T) for three film thicknesses deposited on STO. In the absence of applied field, one always observes a semiconducting behavior with an anomaly around 225 K corresponding to the TCO (this value is close to the one found in the bulk material [15]). On the contrary, the results are drastically different under application of external magnetic fields. A field up to our maximum value (9 T) has almost no effect on the thinner film (15 nm): the film is still semiconducting and the magnetoresistance is very small (Fig. 4(a)). For the intermediate thickness (75 nm), while a 5 T magnetic field keeps the film semiconducting, a 9 T renders it metallic with a TMI close to 125 K (Fig. 4(b)). This feature is a typical characteristic of the melting of the CO state. On the thicker film (110 nm), the melting magnetic field is reducing to 5 T (Fig. 4(c)). The thickness dependence of the properties suggests that the strains play an important role in determining the melting magnetic field. When the thickness of the film is increasing, there is a reduction of the strain on the film at room temperature. However, the properties of the film do not approach those of the bulk as the substrateinduced strain reduces. In PCMO single crystal, with a 6 T magnetic field, the material remains insulating [20], whereas the 110 nm film becomes metallic. The magnetic field dependence of the resistivity shows an important decrease on a logarithmic scale at a critical field ðHC Þ indicating the field-induced melting of the CO state. This field-induced insulator-to-metal transition, which accompanies the collapsing of the CO state takes place below TCO : The critical fields in the field-increasing scan and in the field decreasing scan will, respectively, be represented by HCþ and HC2 : This decrease of the resistivity vs magnetic field can be viewed as CMR of about five orders of magnitude at 75 K. A clear hysteresis is seen at these temperatures as previously reported on several CO compounds [13,20]. This hysteretic region is more pronounced when the temperature is decreasing. The temperature dependence of the large hysteresis region is a feature of a first order transition and has been extensively studied for the composition Nd0.5Sr0.5MnO3 in Ref. [13]. From the resistivity measurements, we can deduce a phase diagram for the 110 nm film (Fig. 5) and for the 75 nm film (inset of Fig. 5) Two remarks should be pointed out. First, as already mentioned, the transition of the metastable state (i.e. a fieldinduced metallic state) is easier or requires a lower field in case of a thin film than in the bulk [20] (5 T for a 110 nm film and 9 T for the 75 nm one). Second, the shape of

1668

W. Prellier et al. / Journal of Physics and Chemistry of Solids 64 (2003) 1665–1669

Fig. 4. rðTÞ under different magnetic fields (0, 5 and 9 T) applied in the plane of the substrate for different PCMO films on STO: (a) 15 nm, (b) 75 nm and (c) 110 nm. Arrows indicate the direction of the temperature variation. Experimental points were taken with a stabilized temperature and a control of the sample film itself. The crossing of the curves ramping up and down in temperature is real. The insets of Fig. 4(b) and (c), respectively, depict the rðHÞ for a 75 and 110 nm film. Runs in field-increasing and field-decreasing are denoted by arrows.

the H – T phase diagram is totally different as compared to the bulk single crystal Pr0.5Ca0.5MnO3. The hysteretic region grows when the temperature is decreasing and rapidly vanishes under 8 T. In others words, HC is smaller than 9 T at low temperature, then HC is decreasing when the T is increasing. Above 50 K, HC begins to increase again to 8 T. In the mean time, the hysteretic region is shrinking.

Note that this region is larger in the case of the 75 nm film indicating that the CO state is more stable as previously seen from the rðTÞ curves (Fig. 3). In fact, the phase diagram of the 110 nm film has almost the same profile as Pr12xCaxMnO3 single crystal with 0:35 , x , 0:45: This suggests that the whole electronic structure of Pr0.5Ca0.5MnO3 thin films grown on STO is similar, though the charge transfer is

W. Prellier et al. / Journal of Physics and Chemistry of Solids 64 (2003) 1665–1669

1669

almost no effect on the resistivity of a very thin film, a 5 T magnetic field can completely melt the CO state of a 110 nm thick film. When the film is under compression (on LaAlO3, the films are insulating under a 7 T magnetic field. By growing thin films, the properties of the materials can be modified compared to the bulk compounds. We have studied of the case of Pr0.5Ca0.5MnO3 on two different substrates that induced two different types of strains (tensile on SrTiO3 and compressive on LaAlO3) and their effects on various parameters. We observed that the lattice parameters of the film, the orientation of the film with respect to the substrate, the transport properties of the film and the stability of the charge/orbital ordered state are influenced by the strains suggesting some similarities between strain and pressure effects in oxides [21].

References [1] [2] [3] [4]

Fig. 5. Electronic phase diagram of a 110 nm PCMO/STO thin film determined by the critical fields. HCþ and HC2 are, respectively, taken as the inflexion points in rðHÞ for up and down sweeps. In the dashed region, both CO insulating and metallic state are coexisting. The inset shows the phase diagram for a 75 nm PCMO film.

different (x ¼ 0:5 instead of 0.4 in Pr12xCaxMnO3). The interesting point is that the modulation vector q is equal to 0.45 in PCMO on STO [16] while q is always equal to 0.5 in bulk Pr Pr12xCaxMnO3 compounds ð0:35 , x , 0:45Þ [19]. In summary, we have shown that the melting magnetic field of the insulating charge-ordered state, in Pr0.5Ca0.5MnO3 films, is strongly sensitive to the substrate-induced strain and thickness of the film. On SrTiO3 (tensile stress), while the application of a magnetic field up to 9 T has

[5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]

R. Von Helmolt, et al., Phys. Rev. Lett. 71 (1993) 2331. K. Chahara, et al., Appl. Phys. Lett. 63 (1993) 1990. M. McCormack, et al., Appl. Phys. Lett. 64 (1994) 3045. T. Venkatesan, et al., Phil. Trans. R. Soc. Lond. 356 (1998) 1661. W. Prellier, et al., J. Phys. Cond. Matter. 13 (2001) R915. S. Jin, et al., Science 75 (1996) 3336. A. Biswas, et al., Phys. Rev. B 61 (2000) 9665. W. Prellier, et al., Appl. Phys. Lett. 75 (1999) 397. H.S. Wang, et al., Appl. Phys. Lett. 74 (1999) 2212. J. O’Donnell, et al., Phys. Rev. B 54 (1996) 6841. J.Z. Sun, et al., Appl. Phys. Lett. 73 (1999) 3294. Y. Konishi, et al., J. Phys. Soc. Jpn 68 (1999) 3790. E.O. Wollan, W.C. Koehler, Phys. Rev. 100 (1950) 545. H. Kuwahara, et al., Science 270 (1995) 961. C.N.R. Rao, et al., J. Phys. Cond. Matter 12 (2000) R83. M. Tokunaga, et al., Phys. Rev. B 57 (1998) 5259. W. Prellier, et al., Appl. Phys. Lett. 77 (2000) 1023. A.M. Haghiri-Gosnet, et al., J. Appl. Phys. 88 (2000) 3545. Z. Jirak, et al., J. Magn. Magn. Mater. 53 (1985) 153. A. Barnabe´, et al., J. App. Phys. 84 (1998) 5506. Y. Tomioka, et al., Phys. Rev. B 53 (1996) R1689. H.Y. Wang, et al., Phys. Rev. B 52 (1995) 15046. P.G. Radaelli, et al., Phys. Rev. B 56 (1997) 8265.