Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.cplire.ru/html/oxide233/publication09/GAO09JETPlsmo_en.pdf Äàòà èçìåíåíèÿ: Sun Jun 12 19:34:50 2011 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 12:45:41 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 5

ISSN 1063-7761, Journal of Experimental and Theoretical Physics, 2009, Vol. 108, No. 1, pp. 48­55. © Pleiades Publishing, Inc., 2009. Original Russian Text © G.A. Ovsyannikov, A.M. Petrzhik, I.V. Borisenko, A.A. Klimov, Yu.A. Ignatov, V.V. Demidov, S.A. Nikitov, 2009, published in Zhurnal èksperimental'nooe i Teoreticheskooe Fiziki, 2009, Vol. 135, No. 1, pp. 56­64.

ORDER, DISORDER, AND PHASE TRANSITION IN CONDENSED SYSTEMS

Magnetotransport Characteristics of Strained La0.7Sr0.3MnO3 Epitaxial Manganite Films
G. A. Ovsyannikova, b, *, A. M. Petrzhika, I. V. Borisenkoa, A. A. Klimova, Yu. A. Ignatova, V. V. Demidova, and S. A. Nikitova
a

Institute of Radio-Engineering and Electronics, Russian Academy of Sciences, Mokhovaya 11, Moscow, 125009 Russia * e-mail: gena@hitech.cplire.ru b Chalmers University of Technology, GÆteborg, SE-41296 Sweden
Received April 4, 2008

Abstract--The electrical and magnetic characteristics of La0.7Sr0.3MnO3 (LSMO) epitaxial manganite films are investigated by different methods under conditions when the crystal structure is strongly strained as a result of mismatch between the lattice parameters of the LSMO crystal and the substrate. Substrates with lattice parameters larger and smaller than the nominal lattice parameter of the LSMO crystal are used in experiments. It is shown that the behavior of the temperature dependence of the electrical resistance for the films in the lowtemperature range does not depend on the strain of the film and agrees well with the results obtained from the calculations with allowance made for the interaction of electrons with magnetic excitations in the framework of the double-exchange model for systems with strongly correlated electronic states. Investigations of the magneto-optical Kerr effect have revealed that an insignificant (0.3%) orthorhombic distortion of the cubic lattice in the plane of the NdGaO3(110) substrate leads to uniaxial anisotropy of the magnetization of the film, with the easy-magnetization axis lying in the substrate plane. However, LSMO films on substrates (((LaAlO3)0.3 + (Sr2AlTaO6)0.7)(001)) ensuring minimum strain of the films exhibit a biaxial anisotropy typical of cubic crystals. The study of the ferromagnetic resonance lines at a frequency of 9.76 GHz confirms the results of magnetooptical investigations and indicates that the ferromagnetic phase in the LSMO films is weakly inhomogeneous. PACS numbers: 74.45.+c, 74.50.+r, 75.70.Cn DOI: 10.1134/S1063776109010075

1. INTRODUCTION Rare-earth manganite perovskites Re1 ­ xAxMnO3 (where Re is a rare-earth element of the La or Nd type and A is an alkaline-earth metal of the Sr or Ca type) exhibit a wide spectrum of unusual electrical and magnetic properties, including the colossal magnetoresistance effect (see reviews [1­4]). The parameters of epitaxial films of these materials very often differ substantially from the characteristics of single crystals. As was shown in a number of works [2, 3, 5­12], the factor responsible for the change in the electrical and magnetic parameters is a strain arising in the films due to the mismatch with the substrate used for depositing the film. It was demonstrated that the three-dimensional compression of the crystal lattice increases the hopping probability amplitude within the double-exchange model, which results in an increase in the Curie temperature TC [12], whereas biaxial distortions of the Jahn­ Teller type lead to an enhancement of electron localization and decrease the Curie temperature TC [5, 6, 11]. The magnetic properties of films are substantially affected by the phase separation phenomena and the presence of a nonmagnetic layer at the substrate­film interface [9]. These effects most strongly manifest themselves in very thin films (thinner than 10 nm).
48

However, a number of problems associated with the influence of the strain on the electrical and magnetic properties of LSMO films have remained unclear and call for further investigation. There is no experimentally confirmed model of the temperature dependence of the electrical resistance for films at low temperatures [4], the problems regarding the influence of the film strain on the magnetic anisotropy [3] and the Curie temperature [5, 6] remain unsolved, etc. In the present paper, we described the results of combined experimental investigation of epitaxial La0.7Sr0.3MnO3 (LSMO) films deposited onto substrates with different crystal lattice parameters. We chose substrates for which the mismatch between the crystal lattice parameters of the substrate and the LSMO film varied from ­2.2 to 2.0%. Owing to the use of identical deposition conditions, it was possible to avoid the effect of other factors (except for the film strain) on the magnetotransport characteristics of films. The film thickness exceeded the thickness (3­5 nm) of the "dead" (nonmagnetic) layer at the film­substrate interface [3, 9] but was smaller than the critical thickness (on the order of 100 nm) at which the strain relaxation is observed in the film. The electrical conductivity, the ferromagnetic transition temperature, and the


MAGNETOTRANSPORT CHARACTERISTICS OF STRAINED La0.7Sr0.3MnO3

49

magnetization anisotropy of the films were studied as a function of the strain induced in the manganite films by interaction with the substrate. 2. SAMPLE PREPARATION AND EXPERIMENTAL TECHNIQUE Epitaxial LSMO films 40­100 nm thick were deposited onto LaAlO3(001) (LAO), SrTiO3(001) (STO), NdGaO3(110) (NGO), ((LaAlO3)0.3 + (Sr2AlTaO6)0.7)(001) (LSAT), GdSrO3(110) (GSO), and DyScO3(110) (DSO) substrates by laser ablation (KrF laser, = 248 nm) at a temperature of 600­800°C and an oxygen pressure of 0.2 mbar.1 All films were cooled under identical conditions (10°C/min, atmospheric pressure) in order to avoid a variation in the oxygen content in the films.2 The crystallographic parameters of the films and substrates were determined by measuring the X-ray diffraction patterns (2/ and scan mode) and the rocking curves on a four-circle X-ray diffractometer. The resistance of the films was measured by the four-point probe method, which excluded the influence of the contact resistance. Contact electrodes were applied by sputtering of platinum through a metal mask. The presence of the ferromagnetic phase in the LSMO films at room temperature makes it possible to investigate the magnetic parameters with the use of the meridional Kerr effect without additional cooling, which simplifies the measurements of the spatial anisotropy of magnetic parameters. The magneto-optical setup for measuring the characteristics of the samples consists of a semiconductor laser ( = 0.63 µm, P = 5 mW), a beam-splitting glass plate for the formation of the reference and signal beams, a phase halfwave plate for the selection of the s or p polarization of light incident on the sample, a polarizer­analyzer for the separation of the signal associated with the meridional Kerr effect, and two photodiodes, one of which is located in the reference channel and the other diode is positioned in the signal channel. The photodiodes operating in the bias mode are connected to a differential resonance amplifier tuned to the frequency of modulation of the laser radiation intensity. The signal (proportional to the change in the angle of rotation of the plane of polarization of reflected light) is fed from the differential amplifier to an analog-to-digital converter (NI6221) and then is processed with the LabView software package. The sample to be studied is located on a
1

rotary table in the gap of an electromagnet. In the setup, the sample can be rotated around the normal to the substrate plane by an angle in the range 0°­360°. The magnetic characteristics were also measured using the magnetic resonance method. The magnetic resonance spectra were recorded on a Bruker ER-200 EPR spectrometer operating in the 3-cm range of electromagnetic waves (frequency, 9.76 GHz). The signal recording was performed by synchronous detection at the frequency of modulation of the external magnetic field (100 kHz). The samples under investigation were located in a microwave cavity of the spectrometer so that the plane of the sample was always parallel to the direction of the constant external magnetic field and the magnetic component of the microwave field. The parallelism of the plane of the samples to the constant magnetic field was controlled from the minimum of the resonance field of the ferromagnetic resonance lines. This location of the samples excludes the change in the magnetic resonance spectra due to the anisotropy of the shape. The samples were rotated around the axis perpendicular to the plane of the samples. 3. CRYSTALLOGRAPHIC INVESTIGATIONS The grown LSMO films were precisely oriented with respect to both the normal to the plane of the substrate and the preferred direction in the plane of the substrate. The interplanar distance a in the LSMO films along the normal to the plane of the substrate and the lattice constant as of the substrate were determined from the X-ray diffraction patterns (measured in the 2/ scan mode). Figure 1 shows the X-ray diffraction patterns (measured in the 2/ scan mode) of the LSMO/NGO and LSMO/STO films. It can be seen from Fig. 1 that the parameter a of the LSMO film depends strongly on the parameter as of the substrate (see also the inset to Fig. 1). The dashed line indicates the position of the hypothetical reflection (002) LSMO for the bulk material (without regard for strains) with the unit cell parameter (in the pseudocubic representation) aLSMO = 0.3876 nm obtained from the measurements of LSMO polycrystals [12]. The lattice constant (coinciding within the limits of experimental error) of individual LSMO film was determined from the experimental data presented in the inset to Fig. 1, which depicts the dependence of the measured interplanar distance a of the LSMO films on the lattice parameter as of the substrates. These parameters were determined from the (002), (003), and (004) peaks of the LSMO films and substrates. The intersection of the dependence a(as) with the straight line a = as leads to the lattice constant of the LSMO film aLSMO = 0.387 ± 0.014 nm, which, within the limits of experimental error, coincides with the results obtained in [12]. It is assumed that the lattice constant a|| of the LSMO film in the plane of the substrate coincides with the lattice constant as to within the experimental error. This is confirmed by the results of investigations of oblique cross
Vol. 108 No. 1 2009

For the most part, the substrates were manufactured by CrysTec GmbH (Germany). 2 The results of low-temperature measurements of the resistance of the LSMO films deposited onto the DSO substrates were irreproducible, probably due to cracking of the films upon cooling. In this paper, we do not discuss the results of the measurements performed for these films.

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


50 I, counts/s 105 (002)STO 10 10 10 10
4
0.38 0.38

OVSYANNIKOV et al.

(220)NGO

a, nm 0.40

3

0.40 as, nm

(002)LSMO/NGO

(002)LSMO/STO

2

1

a = 3.875 å

made for possible tetragonal distortions is equal to 10-3 nm in the plane of the substrate and 10­4 nm in the perpendicular direction. Therefore, the corresponding errors in the calculations of the strains in the plane of the substrate and in the perpendicular direction are equal to 0.20 and 0.02%, respectively. It should be noted that the volume Veff of the crystal lattice Veff of the LSMO film is not retained with a change in the substrate. At small lattice distortions, the Curie temperature TC and other magnetotransport parameters of manganites depend on two parameters: the relative change in the unit cell volume b = 2|| + and the biaxial distortion * = ( ­ ||) 2/3 [6, 11]. The volume distortions of the lattice lead to a decrease or an increase in the Curie temperature TC depending on the distortion sign, whereas the biaxial distortions always result in a decrease in the Curie temperature TC . Below, experimental changes in the Curie temperature TC will be more thoroughly compared with the model [11] which is applicable to small distortions of the cubic lattice. 4. ELECTRICAL PARAMETERS

46.0

46.5

47.0

47.5 2, deg

Fig. 1. X-ray diffraction patterns (measured in the 2/ scan mode) of the LSMO films deposited onto the NGO and STO substrates. The dashed line indicates the position of the hypothetical reflection (002)LSMO for the bulk material (without regard for strains) [12]. The corresponding lattice parameter aLSMO within the experimental error coincides with the lattice parameter obtained from the intersection of two lines (solid and dashed) in the inset to the figure, which depicts the dependence of the interplanar distance a of the LSMO films on the lattice parameter as of the substrates.

sections for thin (less than 100 nm) films and the data obtained in [5]. Consequently, the grown manganite films are under the action of the mechanical strains induced by interaction with the substrate: the compressive strains for the NGO and LAO substrates and the tensile strains for the STO, LSAT, and GSO substrates. The biaxial strains in the plane of the substrate lead to a distortion in the crystal lattice of the film in the perpendicular direction. The strains arising in the film in the plane of the substrate || = (a|| ­ aLSMO)/aLSMO and in the perpendicular direction = (a ­ aLSMO)/aLSMO are listed in the Table 1. The values of the effective volume Veff = a || a are also presented in this table. It can be seen from Table 1 that the compressive strains in the plane of the substrate for the LSMO films deposited onto the NGO and LAO substrates increase the interplanar distance a, whereas the tensile strains (for the STO substrates) decrease this interplanar distance. The accuracy in determining the lattice parameters with allowance
Table 1. Crystallographic parameters of films Substrate NGO LAO STO LSAT GSO as, nm 0.3862 0.3787 0.3903 0.38748 0.3958 a, nm 0.3906 0.4003 0.3845 0.38753 0.3836 , deg 0.06 0.15 0.05 0.06 0.36
2

The resistance of all the films under investigation decreases drastically with a decrease in the temperature T < TC. At low temperatures (T < 10 K), the resistivity of all samples approaches the asymptotic value = 0 . The term 0 independent of the temperature is determined by the processes of scattering from impurities, defects, grain boundaries, and domain walls [13­16]. The minimum resistivity 0 8 â 10­5 cm is observed for the LSMO/NGO and LSMO/LSAT films, which have the most perfect crystal structure (minimum ). The values obtained do not exceed 0 = 1.1 â 10­4 cm for high-quality LSMO/LSAT films [13] but are somewhat larger than those for single crystals [4]. An increase in the quantity 0 for the LSMO/LAO films can be associated with the twinning of the substrate, which results in a broadening of the rocking curve (see Table 1); however, the factors responsible for the increase in the quantity 0 for the LSMO/STO films were not revealed. It should be noted that the rocking curves for the LSMO/STO and LSMO/NGO films coincide with each other, which suggests a high crystallographic quality of the LSMO/STO structures (see

Veff, nm3 0.058 0.057 0.059 0.058 0.060

, % 0.78 3.28 ­ 0.80 ­ 0.02 ­1.04

||, % ­ 0.34 ­2.30 0.70 ­ 0.03 2.11

b, % ­ 0.10 ­1.32 0.59 ­ 0.08 3.17
Vol. 108 No. 1

*, % 0.53 2.63 ­ 0.71 0.01 ­1.48
2009

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


MAGNETOTRANSPORT CHARACTERISTICS OF STRAINED La0.7Sr0.3MnO3

51

Table 1). The temperature dependences (T) observed for manganites at low temperatures can be explained in the framework of the double-exchange model [1, 4, 14, 16]. A strong electron­electron scattering, as a rule, leads to the term 1 proportional to T2. According to the Kubo calculations (see reviews [1, 4]), the scattering of electrons by magnons results in the term 2 T4.5. The relationship 2 T4.5 was actually observed for the La0.67Ba0.33MnO3 manganite films deposited onto the LSAT(001) substrates [16]. Analysis of the temperature dependences of the resistance for our films on the log­ log scale (Fig. 2) demonstrates that, in the low-temperature range 4.2­200 K, the dependence (T) is described with a high accuracy by the expression ( T ) = 0 + 1 T
n

, 10 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0

­3

cm
LSMO/NGO LSMO/LAO LSMO/STO LSMO/LSAT LSMO/GSO

50 100 150 200 250 300 350 400 T, K

(1)

with the exponent n = 2.5 ± 0.1. The same expression was also observed for this material in [13, 14]. The relationship (T) T4.5 [4] associated with the electron­magnon interaction does not describe the experimental dependence (T) but can modify the exponent in the temperature dependence. A recent theoretical analysis of the resistance of manganite films [18] showed that, owing to the double-exchange mechanism, the coupling between the static conductivity fluctuations and the spin fluctuations is responsible for the temperature dependence of the resistance (T) T2.5, which is indeed observed in our experiment. However, the quantity 1 in the experimental temperature dependence (T) T2.5 appears to be one order of magnitude larger than that predicted by the theory [18]. In manganite films, the transition to the ferromagnetic state with a decrease in temperature at T < TC is accompanied by a changeover from a hopping behavior of the temperature dependence of the resistance to a metallic behavior at the temperature T = Tp.3 As a rule, the temperature Tp coincides with the Curie temperature TC to an accuracy of a few degrees. At temperatures in the vicinity of the Curie point TC, the temperature and magnetic field dependences of the resistance are determined by the activation processes [19]. 5. MAGNETO-OPTICAL CHARACTERISTICS OF FILMS The magnetic characteristics of the LSMO films in the external magnetic field applied parallel to the plane of the sample were investigated using the meridional magneto-optical Kerr effect. The magnetization curves, the coercive force Hc , the saturation field Hs, and the anisotropy field Ha were determined at room temperature.
3

Fig. 2. Temperature dependences of the resistivity of the films. Solid lines indicate the approximations (T) = 0 + 1T2.5. The parameters obtained for the approximation (T) from the analysis of the dependences (T) on a logarithmic scale are given in Table 2.

The behavior of the field dependences and the magnitude of the magneto-optical Kerr response for the LSMO films depend substantially on the substrate material (Fig. 3). No anisotropy of the magneto-optical Kerr response for the LSMO films on the LAO and STO substrates is observed upon rotation of the substrate around the normal to the substrate plane. The magnetization curves for the samples grown on the STO and LAO substrates are shown in Fig. 3a. Upon rotation of the samples around the normal to their surface, the shape of the curves remains unchanged, which indicates the absence of a preferred magnetization direction in the plane. A small value of the coercive force for the LSMO/STO structure (4­5 Oe) suggests that this system belongs to the class of "soft magnetic" materials. The magnetization of the LSMO/LAO films does not reach complete saturation in the external magnetic field with a strength of up to 1 kOe (see the inset to Fig. 3a). Most likely, this circumstance is associated with the occurrence of the easy magnetization axis deviating from the film plane [8]. The magneto-optical Kerr signal in the LSMO/GSO films is at the noise level due to the low Curie temperature (TC < 290 K). The LSMO/NGO film is characterized by a magnetic anisotropy of the easy-axis type as a function of the direction of the external magnetic field lying in the plane of the substrate. Upon magnetization along the easy axis, the magnetization curve exhibits a rectangular hysteresis loop when the coercive force Hc = 15 Oe coincides with the saturation field Hs . The inset to Fig. 3b shows the angular dependence of the saturation field Hs for the LSMO/NGO film. The magnetization along the hard axis is attended by the rotation of the magnetic moment in the plane of the sample, and the corresponding curve is characterized by a narrow (no more than 1 to 2 Oe) hysteresis loop. In this case, the
Vol. 108 No. 1 2009

The temperature Tp is determined from the maximum value of the resistance.

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


52 M/Ms, arb. units 1.0 0.5 0 ­ 0.5 ­ 1.0 ­ 100 ­50 0 (a)

OVSYANNIKOV et al.

M/Ms, arb. units 1

­ 800 ­1

0

800 H, Oe

50

100

1.0

(b) 90°
0° 270°

0.5 180° 0 ­ 0.5 ­ 1.0 ­ 300

saturation field Hs is considerably higher than that upon magnetization along the easy axis; it is equal to the anisotropy field Ha and amounts to 120 Oe. The anisotropy of the magnetic properties of the LSMO/NGO films is probably associated with the crystallographic anisotropy of the NGO(110) plane of the substrate. The NGO substrate has an orthorhombic unit cell with the following parameters: a = 0.5426 nm, b = 0.5503 nm, and c = 0.7706 nm (see, for example, [17]). The parameter corresponding to the pseudocubic unit cell of the NGO substrate is equal to 0.3851 nm. A more accurate inclusion of orthorhombic distortions for the NGO(110) plane leads to anisotropy along the NGO[001] and NGO[110] directions (forming the NGO(110) plane): 2(a001 ­ a110)/(a001 + a110) = 0.3%. This anisotropy is most likely responsible for 80% uniaxial anisotropy of the saturation field Hs (Fig. 3b). The crystallographic anisotropy in the LSMO/LSAT films is absent because the LSAT substrates in the temperature range 150­1200 K have cubic symmetry [20]. However, the dependences Hs() for the LSMO/LSAT structures are characterized by the biaxial anisotropy (Fig. 3c) determined by the cubic structure of the LSMO crystal [21, 22]. 6. MAGNETIC RESONANCE INVESTIGATION OF FILMS The ferromagnetic resonance spectra were recorded in order to determine the paramagnetic­ferromagnetic transition temperatures and to study the in-plane magnetization anisotropy of the films. Figure 4 depicts the ferromagnetic resonance spectra at room temperature for the LSMO films grown on the substrates of four types. For each film, Fig. 4 shows two spectra that differ from each other by the rotation of the substrate through 90° around the normal to the substrate. It can be seen from Fig. 4 that the maximum change in the resonance magnetic field H0 upon rotation of the sample by 90° is observed for the LSMO/NGO films. The shape of all ferromagnetic resonance spectra at room temperature can be described by two Lorentzian lines. This fact uniquely indicates that the ferromagnetic phases in all films are inhomogeneous. Since the strain can gradually decrease with increasing thickness and the Curie temperature TC and the g factor depend on the strain , two ferromagnetic phases most likely correspond to film layers with different strains. The difference between the g factors of the Lorentzian lines used for describing the experimental ferromagnetic resonance spectra does not exceed the width pp of the ferromagnetic resonance line. The nonmagnetic layer, which, according to [9], is located at the film­substrate interface, does not substantially affect the ferromagnetic resonance. Figure 5 shows the angular dependences of the resonance magnetic field H0 for two films LSMO/NGO and LSMO/LSAT under investigation at room temperature upon rotation of the sample around the normal to
Vol. 108 No. 1 2009

­ 100

100

300

1.0 0.5 0 ­ 0.5 ­ 1.0 ­ 100

(c)

90° 45°
180°

90°

0°

270°

­50

0 H, Oe

50

100

Fig. 3. Magnetization curves for the structures under investigation: (a) LSMO/STO (triangles), LSMO/LAO (circles), (b) LSMO/NGO, and (c) LSMO/LSAT films. In panels (b) and (c), triangles and circles correspond to different magnetic field directions. The inset to panel (a) shows the magnetization curve for the LSMO/LAO film on an enlarged scale with respect to the external magnetic field. The insets to panels (b) and (c) show the angular dependences of the saturation field Hs in the plane of the substrate for the LSMO/NGO and LSMO/LSAT films, respectively.

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


MAGNETOTRANSPORT CHARACTERISTICS OF STRAINED La0.7Sr0.3MnO3 ESR, arb. units 200 LSMO/NGO 0 ­ 200 1900 2000 4 0 ­4 0.2 0 ­ 0.2 0 0.2 0 ­ 0.2 1000
LSMO/LSAT LSMO/STO

53

H0, Oe 2250 2200 2150 2100 2430 2400 2370 2340 2310 0 90 180 , deg 270 360

2100

2200

2300

1600 1700 1800 1900 2000 2100 2200 2050
LSMO/LAO

2000 2000 3000 4000 5000

1000 1500 2000 2500 3000 3500 4000 H, Oe
Fig. 4. Ferromagnetic resonance spectra of the four structures under investigation at room temperature for two mutually perpendicular orientations of the external magnetic field lying in the plane of the substrate (solid and dashed lines).

Fig. 5. The angular dependences of the resonance field H0 for the LSMO/NGO (squares) and LSMO/LSAT (triangles) films upon rotation of the sample by the angle around the normal to the plane of the substrate. Solid and dashed lines correspond to the relationship H0 2H1cos(2) + 2H2 cos(4) for constant quantities H1 and H2 fitted according to the best agreement with the experimental data.

the plane of the film. The constant magnetic field and the magnetic component of the microwave field lie in the plane of the film. Since the substrates with the film represent square plates 5 â 5 mm2 in size, the influence of the shape anisotropy is minimum and the shift in the resonance field is completely determined by the magnetic anisotropy of the LSMO film. It can be seen from Fig. 5 that the angular dependence of the resonance field is well described by the relationship H0 2H1cos(2) + 2H2 cos(4) with the amplitudes H1 = 55.8 Oe and H2 = 0.9 Oe for the LSMO/NGO film. For the LSMO/LSAT structure, we have H1 = 3.0 Oe and H2 = 8.4 Oe. It should be noted that the angular dependences of the magnetic anisotropy of the LSMO/NGO and LSMO/LSAT films are in good agreement with the corresponding dependences obtained from the magneto-optical measurements (see Section 5). The magnetic anisotropy of the LSMO film is determined by the combination of the cubic anisotropy (in the crystal) proportional to K1sin 4 and the anisotropy induced by the uniaxial strain Ku sin 2 [21, 22]. According to the measurements performed in [22] for the LAO, STO, and MgO substrates, the quantity K1 for the LSMO film decreases sharply with an increase in the temperature and the effect of the uniaxial strains Ku is almost independent of the temperature [22]. As a result, the influence of the cubic anisotropy of the LSMO structure at room temperature manifests itself only for the LSMO/LSAT films, for which the effect of the uniaxial anisotropy is minimum. With a decrease in the temper-

ature, Tsui et al. [6] experimentally observed the biaxial anisotropy for the LSMO films deposited onto three substrates NGO, LSAT, and STO, which provide the minimum mismatch between the lattice parameters. The temperature dependences of the resonance magnetic field for the three samples under investigation are plotted in Fig. 6. The paramagnetic­ferromagnetic transition temperature was identified as the abscissa corresponding to the inflection point of the above dependences. The Curie temperatures TC thus determined are presented in Table 2. The analysis of the ferromagnetic resonance spectra and their temperature dependences clearly indicates that there is a correlation between the ferromagnetic properties of the films and the crystallographic matching of the substrate and film materials. A comparison of the data presented in Tables 1 and 2 allows us to trace that a deterioration of the crystal quality results in an increase in the ferromagnetic resonance width pp. However, at identical crystal quality (equality of the widths of peaks in the rocking curves), the Curie temperature TC depends on the bulk strains [11]. It should be noted that the ferromagnetic resonance width pp = 20 Oe for our LSMO/NGO film is close to a record width for the epitaxial LSMO films. Figure 7 shows the dependences of the Curie temperature TC on the bulk and biaxial strains. The plane in the three-dimensional space describes the dependence
Vol. 108 No. 1 2009

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


54 H0, Oe 3600 3400 3200 3000 LSMO/LAO 2800 LSMO/LSAT 2600 2400 2200 LSMO/STO 2000 1800 260 280 300 320 340 360

OVSYANNIKOV et al. LSMO/STO LSMO/NGO LSMO/LSAT

TC , K 340

325

310 380 T, K 0.2 0 b, %

0 *2, % ­ 0.2 0.1 LSMO/LAO

Fig. 6. Temperature dependences of the resonance magnetic field for three films ( ) LSMO/LAO, ( ) LSMO/LSAT, and ( ) LSMO/STO. The Curie temperatures were determined from the maximum of the derivative dH0/dT.

Fig. 7. Dependences of the Curie temperature TC on the bulk strain b and the square of the biaxial strain *2 for the LSMO films. The plane is constructed using relationship (2).

of the Curie temperature TC on the arising strain according to the relationship obtained by Millis et al. [11]: T C ( ) = T C ( 0 ) ( 1 ­ b ­ * /2 ) ,
2

(2)

where = (1/Tk)dTC/db and = (1/TC)d2TC/d*2. In our case, the best agreement between the experimental M Curie temperature TC and the temperature T C in relationship (2) is achieved at = 0.66 ± 0.07, = 103 ± 2, and TC(0) = 336 ± 0.2 K (see Table 2). A comparison of the calculated temperatures T C and the experimental Curie temperatures TC demonstrates that the deviation of the experimental Curie temperatures from the calculated temperatures is rather small (1 K) for the four most widely used substrates. The largest deviation observed for the Curie temperature TC of the LSMO/GSO films is possibly associated with the considerable bulk distortions of the LSMO film. A higher defectiveness (larger width ) of the LSMO/GSO film leads to a broad ferromagnetic resonance line. It
Table 2. Electrical and magnetic characteristics of films Substrate NGO LAO STO LSAT GSO 0, cm 0.82 â 10 2.12 â 10 3.2 â 10 0.85 â 10 1.3 â 10
­4 ­4 ­4 ­4 ­4
M

should be noted that the parameters obtained for relationship (2) that provide the best agreement between the theoretical and experimental data somewhat differ from those determined by Tsui et al. [6]: = 2.2, = 70, and TC(0) = 334 K. 7. CONCLUSIONS Thus, epitaxial manganite films deposited onto six substrates providing compression and tension of the crystal lattice of the films were investigated. It was demonstrated that the temperature dependence of the resistance of all films at low temperatures is well described by the power relationship T2.5, which follows from the results of recent calculations of the electrical conductivity of manganites in the framework of the double-exchange model. A weak anisotropy (on the order of 0.3%) of the NGO substrate leads to the observed anisotropy of the magnetic properties of the films in the plane of the substrate at room temperature. However, an insignificant crystallographic mismatch

1, cm/T

2.5 ­10 ­10 ­10 ­10 ­10

Hc, Oe 15/<1* 10 4 10 ­

Hs, Oe 120 100 10 50 ­

TC, K 335 315 333 336 251

TC , K 335 316 334 336 328
Vol. 108 No. 1

M

pp, Oe 65 340 64 100 330*
2009

4.7 â 10 12 â 10 18 â 10 4.9 â 10 9.1 â 10

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS


MAGNETOTRANSPORT CHARACTERISTICS OF STRAINED La0.7Sr0.3MnO3

55

between the LSMO film and the LSAT substrate is responsible for the magnetic anisotropy typical of cubic ferromagnetic materials. The ferromagnetic resonance lines of all films are adequately described by a superposition of two Lorentzian lines, which indicate the presence of two ferromagnetic phases. The dependence of the Curie temperature determined from the ferromagnetic resonance spectra is qualitatively described by the Millis theory, which takes into account both the bulk and biaxial strain of the LSMO films. ACKNOWLEDGMENTS We would like to thank V.I. Atsarkin, Yu.A. Booekov, D. Winkler, A.L. Vlasyuk, T. Claeson, K.Y. Constantinyan, I.M. Kotelyanskii, Ya.M. Mukovskii, and A. V. Shadrin for helpful discussions of the results and for assistance in performing the investigations. This study was performed within the programs "Strongly Correlated Electrons in Semiconductors, Metals, Superconductors, and Magnetic Materials" and "Spin-Dependent Effects in Solids and Spintronics" of the Branch of General Physics and Astronomy of the Russian Academy of Sciences; Program P-03-2.25 of the Presidium of the Russian Academy of Sciences; the Program of the European Union (project no. NMP3CT-2006-033191); and the programs "Arrays of Quantum Dots and Josephson Junctions" and "Thin Films for Novel Oxide Devices" of the European Science Foundation; it also receive support from the International Science and Technology Center (project no. 3743); the Ministry of Education and Science of the Russian Federation; the Council on Grants from the President of the Russian Federation for the Support of Leading Scientific Schools (project no. NSh5408.2008.2); and the Russian Foundation for Basic Research (project no. 08-02-00487). REFERENCES
1. Yu. A. Izyumov and Yu. N. Skryabin, Usp. Fiz. Nauk 171 (2), 121 (2001) [Phys.--Usp. 44 (2), 109 (2001)]. 2. W. Prellier, Ph. Lecoeur, and B. Mercey, J. Phys.: Condens. Matter 13, R915 (2001). 3. A.-M. Haghiri-Cosnet and J. P. Renard, J. Phys. D: Appl. Phys. 36, R127 (2003). 4. M. Ziese, Rep. Prog. Phys. 65, 143 (2002).

5. P. Dey, T. K. Nath, and A. Tarapher, Appl. Phys. Lett. 91, 012 511 (2007). 6. F. Tsui, M. C. Smoak, T. K. Nath, and C. B. Eom, Appl. Phys. Lett. 76, 2421 (2000). 7. Y. P. Lee, S. Y. Park, Y. H. Hyun, and J. B. Kim, Phys. Rev. B: Condens. Matter 73, 224 413 (2006). 8. Yan Wu, Y. Suzuki, U. RÝdiger, J. Yu, A. D. Kent, T. K. Nath, and C. B. Eom, Appl. Phys. Lett. 75, 2295 (1999). 9. M. Bibes, S. Valencia, L. Balcells, B. Martoenez, J. Fontcuberta, M. Wojcik, S. Nadolski, and E. Jedryka, Phys. Rev. B: Condens. Matter 66, 134 416 (2002). 10. H. Y. Hwang, T. T. M. Palstra, S.-W. Cheong, and B. Batlogg, Phys. Rev. B: Condens. Matter 52, 15 046 (1995). 11. A. J. Millis, T. Darling, and A. Migliori, J. Appl. Phys. 83, 1588 (1998). 12. M. C. Martin, G. Shirane, Y. Endoh, K. Hirota, Y. Moritomo, and Y. Tokura, Phys. Rev. B: Condens. Matter 53, 14 285 (1996). 13. Yu. A. Booekov, T. Claeson, and V. A. Danilov, Fiz. Tverd. Tela (St. Petersburg) 47 (12), 2189 (2005) [Phys. Solid State 47 (12), 2281 (2005)]. 14. G. J. Snyder, R. Hiskes, S. DiCarolis, M. R. Beasley, and T. H. Geballe, Phys. Rev. B: Condens. Matter 53, 14 434 (1996). 15. P. Schiffer, A. P. Ramirez, W. Bao, and S.-W. Cheong Phys. Rev. Lett. 75, 3336 (1995). 16. Yu. A. Booekov and V. A. Danilov, Fiz. Tverd. Tela (St. Petersburg) 50 (1), 92 (2008) [Phys. Solid State 50 (1), 95 (2008)]. 17. I. K. Bdikin, P. B. Mozhaev, G. A. Ovsyannikov, F. V. Komissinskioe, I. M. Kotelyanskioe, and E. I. Raksha, Fiz. Tverd. Tela (St. Petersburg) 43 (9), 1548 (2001) [Phys. Solid State 43 (9), 1611 (2001)]. 18. M. J. Calderon and L. Brey, Phys. Rev. B: Condens. Matter 64, 140 403 (2001). 19. N. G. Bebenin, R. I. Zainullina, V. V. Mashkautsan, V. V. Ustinov, and Ya. M. Mukovskii, Phys. Rev. B: Condens. Matter 69, 104 434 (2004). 20. B. C. Chakoumakos, D. G. Schlom, M. Urbanik, and J. Luine, J. Appl. Phys. 83, 1979 (1998). 21. Y. Suzuki, H. Y. Hwang, S.-W. Cheong, T. Siegrist, R. B. van Dover, A. Asamitsu, and Y. Tokura, J. Appl. Phys. 83, 7064 (1998). 22. K. Steenbeck and R. Hiergeist, Appl. Phys. Lett. 75, 17 778 (1999).

Translated by O. Borovik-Romanova

JOURNAL OF EXPERIMENTAL AND THEORETICAL PHYSICS

Vol. 108

No. 1

2009