Ferromagnetism in epitaxial orthorhombic YMnO3 thin films

ARTICLE IN PRESS Journal of Magnetism and Magnetic Materials 321 (2009) 1719–1722 Contents lists available at ScienceDirect Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm Ferromagnetism in epitaxial orthorhombic YMnO3 thin films X. Marti a, V. Skumryev b,c, A. Cattoni d, R. Bertacco d, V. Laukhin a,b, C. Ferrater e, M.V. Garcı́a-Cuenca e, M. Varela e, F. Sánchez a, J. Fontcuberta a, a Institut de Ciència de Materials de Barcelona-CSIC, Campus UAB, Bellaterra 08193, Spain Institut Català de Recerca i Estudis Avanc- ats, Barcelona, Spain Departament de Fı́sica, Universitat Autònoma de Barcelona, Campus UAB, Bellaterra 08193, Spain d L-NESS, Dipartimento di Fisica, Politecnico di Milano, via Anzani 42, Como 22100, Italy e Departament de Fı́sica Aplicada i Òptica, Universitat de Barcelona, Diagonal 647, Barcelona 08028, Spain b c a r t i c l e in fo abstract Available online 11 February 2009 Epitaxial orthorhombic YMnO3 thin films, (0 0 1) oriented, have been grown by pulsed laser deposition on (0 0 1)SrTiO3 substrates. Their crystal structure and magnetic response have been studied in detail. Although bulk o-YMnO3 is antiferromagnetic, our magnetic measurements reveal intriguing thermal hysteresis between the zero-field-cooled and field-cooled curves below the onset of the antiferromagnetic ordering temperature, thus signaling a more complex magnetic structure with net ferromagnetic moments. We discuss on the possible origin of this net magnetization and we have found a correlation of the magnetic response with the strain state of the films. We propose that substrate-induced strain modifies the subtle competition of magnetic interactions and leads to a non-collinear magnetic state that can thus be tuned by strain engineering. & 2009 Elsevier B.V. All rights reserved. Keywords: Multiferroic films Induced ferromagnetism YMnO3 Strain effects 1. Introduction The investigation on oxides presenting coexistence of electric and magnetic orders is a very active field. When both functional properties are coupled, these materials would allow to control the magnetic (electric) state by means of electric (magnetic) inputs. In a first approach, biferroic materials, i.e. ferroelectric (FE) and ferromagnetic (FM), would be desirable. However, since FM and FE usually exclude each other [1], the list of suitable candidates is rather small and the focus of the research has been extended to antiferromagnetic (AF) and FE materials, which may be exchangecoupled to ferromagnetic layers [2,3] to achieve new functionalities [4,5]. Among materials displaying simultaneous AF and FE character, probably the hexagonal rare-earth manganites (RMnO3) are best known. When R is a rare-earth of small size, (Ho–Lu, Y) have hexagonal crystal structure and present both FE (Tc700–1000 K) [6] and AF orders (TN80–100 K) [7]. In thin film form, one of the most studied material is YMnO3 [8,9]. As the ionic radius of the rare-earth increases (Tb, Dy and Gd), the stable phase is orthorhombic [10]. In this phase, the AF nature is well established [11] and magnetoelectric effects and observations of net electrical polarization have been reported [12,13]. Although the stable phase of bulk, of YMnO3 is hexagonal, the orthorhombic phase o-YMnO3 can be obtained either by high  Corresponding author. Tel.: +34 935801853x228; fax: +34 935805729. E-mail address: fontcuberta@icmab.es (J. Fontcuberta). 0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2009.02.025 pressure techniques [14] or, in the case of thin films, by epitaxial stabilization by proper substrate selection [15–17]. The crystal structure, in the Pbnm setting, presents a ¼ 5.26 Å, b ¼ 5.85 Å and c ¼ 7.36 Å lattice parameters [18]. On the other hand, the AF magnetic structure has been studied by neutron diffraction experiments concluding that spins are arranged in a sinusoidal modulation along the [0 1 0] direction below a transition temperature TN40 K [19,20] and down to 1.7 K. However, the understanding of the magnetic measurements in powder samples is still uncompleted. All studies signal the order–disorder magnetic transition at about 40 K, although this transition is manifested in a very different way: while a sharp kink in the temperature dependence of dc-susceptibility was observed by Lorenz et al. [14] at the transition, there is only a very subtle bump on the data reported by Muñoz et al. [19]. While the inverse susceptibility approaches the transition in a way signaling ferrimagnetic order in the work of Brinks et al. [20], it does not show the characteristic ferrimagnetic downturn in the data reported in by Muñoz et al. [19] or Iliev et al. [21]. Regarding the existence of net magnetization, it is worth noting the small hysteresis that exists in the temperature range 20–30 K between the cooling and heating branches of dc-susceptibility reported by Lorenz et al. [14] indicating net magnetization. In contrast, field dependence curves, reported only by Muñoz et al. [19], display neither a remanent magnetization nor hysteresis at any temperature down to 2 K. Moreover, a sinusoidal magnetic structure suggests a subtle equilibrium among different magnetic interactions which are expected to be strongly dependent on the ARTICLE IN PRESS 1720 X. Marti et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 1719–1722 Mn–O–Mn bond topology (angles and distances). The epitaxial stress in thin film samples will modify the unit cell lattice parameters and, therefore, the magnetic ordering is expected to be severely affected. Indeed, we reported that, under certain deposition conditions, films presented FM behavior [22], but the microscopic origin of this behavior could not be determined. Here, we report a detailed study of the role of strain, growth conditions and thickness on the magnetic properties of o-YMnO3 thin films and discuss on the relation between crystal structure and magnetic properties. Results suggest that the epitaxial strain in the [0 1 0] direction, which confines the sinusoidal modulation of the spins, is the most relevant parameter in determining the appearance of a ferromagnetic response at temperatures below the onset of antiferromagnetic order. 2. Experimental YMnO3 (YMO) thin films were deposited on (0 0 1)-oriented SrTiO3 (STO) substrates by pulsed laser deposition. A KrF excimer laser (248 nm wavelength, 34 ns pulse duration) was used at a repetition rate of 5 Hz. The laser beam was focused on a stoichiometric YMnO3 target, being the substrate placed at a distance of 5 cm. The films were grown at a substrate temperature of 785 1C, with oxygen pressure in the 0.1–0.4 mbar range. Aiming to obtain films of different thicknesses at each pressure, series of films were grown using different number of laser pulses (4, 6 and 10 k) and fluences (1.5 and 2.2 J/cm2, with growth rates of 0.07 and 0.12 Å/pulse, respectively). Growth rate has been observed to decrease very slightly with oxygen pressure in the used range. After thickness determination of all films using wavelength dispersive X-ray spectroscopy (WDS), it turned out that the following films have been obtained: 119 and 34 nm at PO2 ¼ 0.1 mbar, 122 and 32 nm at PO2 ¼ 0.2 mbar, 111, 75 and 31 nm at PO2 ¼ 0.3 mbar and 49 nm at PO2 ¼ 0.4 mbar. At the end of the growth, the samples were cooled down and 1 atm of oxygen was introduced at 530 1C into the chamber. The crystal structure, epitaxial relationships and lattice strain (out-of-plane and in-plane) were investigated by 4-circle Philips MRD X-ray diffractometer. Magnetic properties were measured by using a superconducting quantum interference device (SQUID). The temperature dependence of the magnetic moment was measured after zero-field-cooling (ZFC) and field-cooling (FC) procedures at a rate of 2 K/min with a magnetic field of 500 Oe applied in the plane of the film. The surface morphology was studied by atomic force microscopy (AFM). Stoichiometry of the films has been checked by WDS showing constant Y/Mn ratio independent of the oxygen pressure in each of the thickness series. Chemical state of the Mn has been measured by X-ray photoemission spectroscopy (XPS) experiments performed in a PHI 5500 Multitechnique System (Physical Electronics) with a monochromatic X-ray source (Al–Ka line, 1486.6 eV, 350 W), calibrated using the 3d5/2 line of Ag with a full-width at half-maximum of 0.8 eV. Spectra have been taken at room temperature at 451 degrees of collection angle. 3. Results and discussion X-ray diffraction experiments show that YMO thin films are orthorhombic, single phase and (0 0 1) textured. As an example, Fig. 1a displays a y/2y scan of a film 120 nm thick grown at 0.1 mbar, where only peaks from STO(0 0 1) and YMO(0 0 1) families are observed. The same behavior is observed in all other films. To determine the in-plane lattice parameters and, thus, the strain state of the films, we performed reciprocal space maps. Fig. 1b shows the vicinity of the STO(11 4) reflection. Two more Fig. 1. X-ray diffraction data from a 120 nm film grown at 0.1 mbar: (a) y/2y scan showing orthorhombic YMO(0 0 1) as well as STO(00 l) reflections, and (b) reciprocal space map in the vicinity of STO(11 4) showing the reflections YMO(2 0 8) and YMO(0 2 8). (c) Profiles extracted at L ¼ 4.22 in the reciprocal space maps for a selection of samples. Vertical lines show the bulk location for the (0 2 8) and (2 0 8) reflections. peaks are observed, corresponding to the YMO(2 0 8) and YMO (0 2 8) reflections from the film. As discussed elsewhere [17], the presence of these two peaks indicates the existence of two crystal domains, 901 in-plane rotated, with the same out-of-plane parameter. Profiles of the reciprocal space maps along a fixed L (signaled by arrow in the map) gives straight forward the in-plane parameters of the film as aSTOO2/Qx. Fig. 1c shows the profiles from a selection of the samples, with vertical lines that mark the positions of YMO(0 2 8) and YMO(2 0 8) reflections in bulk. It is observed that the a parameter is relaxed, whereas, b parameter always presents a certain degree of strain caused by the compressive epitaxial stress, which varies depending on growth conditions and film thickness. It is worth commenting here that the sinusoidal spin modulation is located along the b direction anticipating that a modification of the AF magnetic structure is expected. Fig. 2 displays the dependence of the lattice parameters with the oxygen pressure in two series of films (120 and 30 nm). Panel (a) shows that for both series of samples the b-axis is compressed compared to the bulk value and when increasing thickness the b-axis gradually expands, approaching the bulk value, indicating a progressive relaxation. A similar trend is observed when comparing films grown at increasingly higher ARTICLE IN PRESS X. Marti et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 1719–1722 1721 Fig. 3. X-ray photoelectron spectroscopy scans around the position of Mn 3s showing the splitting between low (L) and high (H) spin states. Figure shows that the change in O2 pressure during deposit in the range 0.1–0.4 mbar does not introduce changes in the chemical state of Mn. Fig. 2. Lattice parameters of films grown at different pressure, 30 nm thick (closed symbols) and 120 nm thick (open symbols): (a) b-axis, (b) c-axis and (c) unit cell volume. Horizontal dashed lines signal the bulk values. pressure. We note in Fig. 2a that overall variation of b with PO2 is smaller than its variation with increasing thickness, being 1.5% and 1.1%, respectively. When analyzing the variation of the c-parameter (Fig. 2b) it turns out that it is expanded compared to bulk value. Upon cell relaxation, either by increasing thickness or PO2, it gradually contracts approaching the bulk value. The progressive relaxation by increasing either thickness or PO2 is also observed for the unit cell volume as shown in Fig. 2c. Data in Fig. 2 indicates that the variation of oxygen pressure during the deposition is an appropriate tool to modify the cell parameters of films. This could result from substrate-induced effects or from chemical modifications associated to the different growth conditions. To get a deeper information about the relative role of these effects and eventually, to discriminate among them, we have performed XPS analysis of some films. The Mn-3s spectra collected from samples grown at PO2 ranging from 0.1 and 0.4 mbar are reported in Fig. 3. It is well known that in manganites the energy splitting DE3s between the high-spin and the low-spin final state configuration is roughly proportional to the Mn formal valence, with a decrease of roughly 0.7 eV per unitary decrease in valence [23–25]. Data in Fig. 3 shows that samples grown at different pressure present the same exchange splitting, i.e. the same oxidation sate of Mn. Moreover, the value DE3s ¼ 5.5 eV is compatible with the oxidation state 3+, as expected for stoichiometric YMnO3. A possible misbalance of oxygen stoichiometry should have instead induced the presence of a certain amount of Mn4+ and/or Mn2+ atoms for electrical charge compensation which, at the same time, would have been detected as merging of the Mn-3s peaks or a larger splitting, respectively. Therefore, within the experimental resolution, the oxygen pressure during deposition has no influence on the chemical state of Mn. Now, we address the magnetic measurements performed in the films. In Fig. 4a, we show, as a typical example, the zero-fieldcooled, and field-cooled, magnetization curves of the o-YMnO3 film prepared at 0.1 mbar, 120 nm thick, measured with the field either in-plane (solid symbols) and out-of-plane (empty symbols). It is worth to recall that in bulk the magnetic moments should be Fig. 4. (a) Temperature dependence of the magnetic moment after ZFC and FC procedures with a sample grown at 0.1 mbar, 120 nm thick. A magnetic field of 500 Oe is applied in-plane (solid symbols) and out-of-plane (empty symbols). Inset shows magnetization loops collected at 2 K with magnetic field applied in in-plane (solid symbols) and out-of-plane (empty symbols). (b) FC magnetization curves (field applied in-plane) of samples grown at different conditions. Magnetization at T ¼ 25 K as a function of (c) epitaxial strain and (d) surface roughness of the films. Spin arrangement is sketched in panel (c) showing the modulation along the [0 1 0] direction. Symbols used to identify samples in panel (b) are kept in panels (c) and (d). confined in the ab plane as shown in the sketch in Fig. 4c. With this spin ordering, at temperatures below TN, one should expect a drop of the magnetization without thermal hysteresis if the magnetic field is applied in-plane of the sample. In contrast, measurements (Fig. 4a, solid symbols) show a remarkable thermal hysteresis below T50 K with a splitting which gradually increases as temperature decreases, thus indicating the existence ARTICLE IN PRESS 1722 X. Marti et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 1719–1722 of net magnetic moment in the film. The low-temperature magnetic susceptibility is larger for the field applied perpendicular to the ab plane (empty symbols) than for field applied in-plane (solid symbols). The existence of an anisotropic ferromagnetic component is also illustrated by the magnetization curves vs field shown in Fig. 4a (inset), measured at 2 K. The existence of an hysteresis ZFC–FC has been observed in all strained films, thus suggesting that the ferromagnetic component is associated to strain. We show FC curves in Fig. 4b, and in Fig. 4c we summarize the values of the FC magnetization measured at 25 K (M25 K) vs the in-plane strain e[0 1 0] along the [0 1 0] direction. In these measurements, the external magnetic field was applied in-plane of the sample. A similar trend is observed when comparing data at other – lower – temperatures. Data in this figure provide a clear insight into the physical origin of the induced ferromagnetic response. It clearly shows that the induced ferromagnetic response is directly related to the strain: the magnetization increasing with increasing strain. We should mention that the existence of ZFC–FC hysteresis had been earlier observed in thin films [22] being more pronounced than in bulk samples [14]. One of the essential differences to evoke when comparing data from the polycrystalline bulk samples (no single crystals of metastable o-YMnO3 have been produced yet) with the epitaxial thin films, is the modification of the magnetic structure due to the strain induced by the substrate. Therefore, a plausible picture could be that the epitaxial strain may introduce some degree of canting in the spin modulation that gives rise to net magnetic moment in the magnetically ordered state. Moreover, in Fig. 4b one notices that the onset of magnetization, occurs at about 60–70 K, which is above the Néel temperature (40 K) reported for bulk o-YMnO3 [14,19–21]. This shift could also reflect the strain-induced unbalance of the antiferromagnetic competing interactions in the film. Finally, one may imagine the scenario where the crystal defects (dislocations, etc), grain boundaries or even surface roughness, may be a source of magnetic disorder and/or uncompensated spins. To address this issue, in Fig. 4d we have plotted M25 K against the surface roughness (RMS) as determined by AFM for films prepared under different conditions. All films display a grain-like pattern morphology with lateral grain size in the 50–100 nm range. Data in this plot indicate that the ferromagnetic component M25 K – and the ZFC–FC hysteresis – is reduced when increasing the RMS. It suggests that uncompensated spins at grain boundaries do not play a dominant role on the total magnetic response of the film. Clearly, films having rougher surfaces present reduced magnetic response. Another possible scenario compatible with the observed ferromagnetic behavior is a possible coexistence of Mn3+ and Mn4+ ions, driving to partial superparamagnetic or spin-glass behavior as reported by Muñoz et al. [19]. However, this possibility does not appear to be supported by the XPS experiments, which reveal constant 3+ manganese valence in all the samples. 4. Conclusions Epitaxial orthorhombic YMnO3(0 0 1) thin films were grown on SrTiO3(0 0 1) substrates by pulsed laser deposition. The analysis by X-ray diffraction reveals that pressure and thickness are good parameters to control the strain state of the films. Magnetic measurements on the films showed that, close to the antiferromagnetic ordering temperature, a clear and unexpected thermomagnetic hysteresis develops. Correlation with structural data reveals that strain state of the films is a crucial parameter inducing the existence of an unexpected ferromagnetic response. Microstructural roughness has been considered as a potential source for magnetic disorder, but left behind after atomic force microscopy characterization indicated that rougher films do not present larger magnetization. In summary, data indicate that the strain state of the films is the main driving force on the appearance of net magnetic moment in orthorhombic YMnO3 thin films. Acknowledgements Financial support by the Ministerio de Ciencia e Innovación of the Spanish Government Projects: NAN2004-9094-C03, MAT20055656-C04, MAT2008-06761-C03 and NANOSELECT CSD200700041, and by the European Union [Project MaCoMuFi (FP6-03321) and FEDER] is acknowledged. References [1] N.A. Hill, J. Phys. Chem. B 104 (2000) 6694. [2] X. Marti, F. Sánchez, D. Hravobsky, L. Fabrega, A. Ruyter, J. Fontcuberta, V. Laukhin, V. Skumryev, M.V. Garcı́a-Cuenca, C. Ferrater, M. Varela, A. Vilà, U. Lüders, J.F. Bobo, Appl. Phys. Lett. 89 (2006) 032510. [3] J. Dho, M.G. Blamire, Appl. Phys. Lett. 87 (2005) 252504. [4] V. Laukhin, V. Skumryev, X. Marti, D. Hravobsky, F. Sánchez, M.V. Garcı́aCuenca, C. Ferrater, M. Varela, J. Fontcuberta, Phys. Rev. Lett. 97 (2006) 227201. [5] Y. Chu, L.W. Martini, M.B. Holcomb, M. Gajek, S.J. Han, Q. He, N. Balke, C. Yang, D. Lee, W. Hu, Q. Zhan, P.L. Yang, A. Fraile-Rodrı́guez, A. Scholl, S.X. Wang, R. Ramesh, Nat. Mater. 7 (2008) 478. [6] N. Fujimura, T. Ishida, T. Yoshimura, T. Ito, Appl. Phys. Lett. 69 (1996) 1011. [7] M. Fiebig, D. Fröhlich, K. Kohn, St. Leute, Th. Lottermoser, V.V. Pavlov, R.V. Pisarev, Phys. Rev. Lett. 84 (2000) 5620. [8] D. Ito, N. Fujimura, T. Yoshimura, T. Ito, J. Appl. Phys. 93 (2003) 5563. [9] X. Marti, F. Sánchez, D. Hrabovsky, J. Fontcuberta, V. Laukhin, V. Skumryev, M.V. Garcı́a-Cuenca, C. Ferrater, M. Varela, U. Lüders, J.F. Bobo, S. Estradé, J. Arbiol, F. Peiró, J. Cryst. Growth 299 (2007) 288. [10] A.R. Kaul, O.Y. Gorbenko, A.A. Kamenev, Russ. Chem. Rev. 73 (2004) 861. [11] T. Kimura, G. Lawes, T. Goto, Y. Tokura, A.P. Ramirez, Phys. Rev. B 71 (2005) 224425. [12] M. Mostovoy, Phys. Rev. Lett. 96 (2006) 067601. [13] T. Arima, A. Tokunaga, T. Goto, H. Kimura, Y. Noda, Y. Tokura, Phys. Rev. Lett. 96 (2006) 097202. [14] B. Lorenz, Y.Q. Wang, Y.Y. Sun, W. Chu, Phys. Rev. B 70 (2004) 212412. [15] P.A. Salvador, T.D. Doan, B. Mercey, B. Raveau, Chem. Mater. 10 (1998) 2592. [16] O.Y. Gorbenko, O. Yu, S.V. Samoilenkov, I.E. Graboy, A.R. Kaul, Chem. Mater. 14 (2002) 4026. [17] X. Marti, F. Sánchez, V. Skumryev, V. Laukhin, C. Ferrater, M.V. Garcı́a-Cuenca, M. Varela, J. Fontcuberta, Thin Solid Films 516 (2008) 4899. [18] Powder Diffraction File, Joint Committee on Powder Diffraction Standards, American Society for Testing of Materials (ASTM), Philadelphia, PA, 2000, Card 20-0732. [19] A. Muñoz, J.A. Alonso, M.J. Martı́nez-Lope, M.T. Casáis, J.L. Martı́nez, M.T. Fernández-Dı́az, J. Phys.: Condens. Matter 14 (2002) 3285. [20] H.W. Brinks, H. Fjellvag, A. Kjekshus, J. Solid State Chem. 129 (1997) 334. [21] M.N. Iliev, B. Lorenz, A.P. Litvinchuk, Y.Q. Wang, Y.Y. Sun, C.W. Chu, J. Phys.: Condens. Matter 17 (2005) 3333. [22] X. Marti, V. Skumryev, V. Laukhin, F. Sánchez, C. Ferrater, M.V. Garcı́a-Cuenca, M. Varela, J. Fontcuberta, J. Mater. Res. 22 (2007) 2096. [23] V.R. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N.A. Ovechkina, E.Z. Kurmaev, N.I. Lobachevskaya, Ya.M. Mukovskii, J. Mitchell, D.L. Ederer, Phys. Rev. B 65 (2002) 113102. [24] L. Zhong, V. Young, J. Electron Spectrosc. Rel. Phenom. 34 (1984) 45. [25] R. Bertacco, A. Tagliaferri, M. Riva, L. Signorini, M. Cantón, A. Cattoni, F. Cicacci, Phys. Rev. B 78 (2008) 035448.