Thin Solid Films 413 (2002) 32–40
Magnetoresistant La1yxSrxMnO3 films by pulsed injection metal organic chemical vapor deposition: effect of deposition conditions, substrate material and film thickness A. Abrutisa,*, V. Plausinaitienea,b, V. Kubiliusa, A. Teiserskisa, Z. Saltytea, R. Butkuteb, J.P. Senateurc a
Department of General and Inorganic Chemistry, Vilnius University, Naugarduko 24, LT-2006 Vilnius, Lithuania b Semiconductor Physics Institute, A. Gostauto 11, LT-2600 Vilnius, Lithuania c ` LMGP, ENS de Physique de Grenoble, INPG, UMR CNRS 5628, Saint Martin d’Heres, France Received 1 August 2001; received in revised form 16 February 2002; accepted 22 March 2002
Abstract High quality epitaxial La1yxSrxMnO3 films were deposited by pulsed injection metal organic chemical vapor deposition, using as precursor materials metal-2,2,6,6-tetramethyl-3,5-heptandionates dissolved in monoglyme. The influence of various deposition conditions, substrate material and film thickness on film properties was investigated. The best films were deposited on LaAlO3 substrates at 825 8C: films exhibited a sharp semiconductor–metal transition and high magnetoresistance (;40%) at a close-toroom temperature and in a rather low field of 1.5 T. In-situ or ex-situ high-temperature annealing in oxygen increased the temperature of semiconductor–metal transition, but decreased the magnetoresistance and the temperature coefficient of resistance. Biaxial strain imposed by the lattices’ mismatch was clearly observed in thinner La1yx Srx MnO3 films on perovskite substrates. Tensile strain was present in the films on SrTiO3 and compressive strain in the films on LaAlO3 (and less clearly on NdGaO3). Both tensile and compressive strains decreased the temperature of electric transition and the values of magnetoresistance. The strain was completely relaxed in the films more than ;100 nm thick. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Chemical vapor deposition; Manganites; Electrical properties; Magnetic properties
1. Introduction Over the last few years much attention has been focused on colossal magnetoresistance observed in films of perovskite manganites of the general formula A1yxBxMnO3 (AsLa, Nd, Pr, BsCa, Sr, Ba, Pb). This attention was justified by a wide range of possible applications of these films in magnetic sensors, magnetoresistive read heads and magnetoresistive random access memory. The attempts are focused on increasing the magnetoresistance at room temperature in small magnetic fields. Moreover, the large temperature coefficient of resistance near the resistivity peak makes these materials real candidates for infrared detectors suitable for uncooled bolometric applications. Among these *Corresponding author. Tel.: q370-2-331004; fax: q370-2630987. E-mail address:
[email protected] (A. Abrutis).
materials, special attention has been paid to thin films of La0.67Ca0.33MnO3 and La0.67Sr0.33MnO3 due to their high Curie point values and colossal magnetoresistance near room temperature. The A1yxBxMnO3 perovskites also have advantages over other magnetoresistant materials, because their temperature of paramagnetic–ferromagnetic transition (Tc) and insulator–metal transition (Tr) can be tuned by varying the doping rate x and the type of A and B ions. The electrical and magnetic properties of perovskite manganite films are closely related to their composition, crystallinity and epitaxy, oxygen content, i.e. to the characteristics that strongly depend on the deposition technique and deposition conditions. Contrary to a large number of publications on physical vapor depositions, reports on the preparation of manganite thin films by metal organic chemical vapor deposition (MOCVD) are not numerous w1–9x. Due to difficulties related to
0040-6090/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 0 - 6 0 9 0 Ž 0 2 . 0 0 3 5 2 - 8
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
33
the deposition of oxide layers and multilayers. Earlier this technique has been optimized for growing high quality YBCO films on various substrates w11–13x and YBCOyPrBCO superlattices w14x. In the present work, the influence of various deposition conditions, substrate material and film thickness on La1yxSrxMnO3 film properties was studied. 2. Experimental details
Fig. 1. Schematic representation of injection MOCVD reactor.
multichannel MO-vapor transport in the growth process of multicomponent oxide film by classical MOCVD, in all the works cited only CVD modifications based on a single source continuous liquid injection and flash evaporation principles were used for the deposition of La1yxCaxMnO3 w1–6x, La1yxSrxMnO3 w4–8x and La0.8MnO3-d w9x films. It is not very difficult to obtain, by single source MOCVD techniques, La1yxSrxMnO3 films with a Tr value above 300 K, especially on highly matched perovskite substrates. It is more problematic to achieve a sharp insulator–metal transition and high magnetoresistance values in low or moderate magnetic fields. In this case the features of deposition technique become important; moreover, the process must be well optimized by studying the influence of deposition conditions on the film properties. Probably mainly due to these reasons La1yxSrxMnO3 films of mediocre quality are described in w4,7,8x: even on perovskite substrates the films had a large metal–insulator transition and exhibited significant magnetoresistance only in high magnetic fields (5–7 T). In the other cited publications on CVD of manganite films the bulk of attention is given to calcium-doped material, while information on deposited La1yxSrxMnO3 films is marginal and insufficient. In the present work we report on the deposition of high quality La1yxSrxMnO3 (LSMO) films by a recently elaborated MOCVD technique—pulsed liquid injection MOCVD w10x. It is a simple and versatile technique for
A scheme of the vertical hot wall injection CVD reactor used in this work is presented in Fig. 1. In this reactor, films on substrates up to 2 inches in diameter can be deposited. The reactor has three evaporators, each can connect one injector. The injector injects microdoses (a few microlitres) of an organic solution containing a dissolved mixture of organometallic precursors. After flash evaporation of a microdose, the resulting vapor mixture is transported by ArqO2 gas towards the heated substrate. The injector is a computerdriven, high speed and precision electromagnetic valve operating in repetitive pulsing (usually used for fuel injection in internal-combustion engines). The main differences of our method compared with other single source methods based on flash evaporation principle are pulsed character of deposition process, simplicity of operation, precision and reproducibility of precursor delivery, ensuring the reproducibility of the CVD process. There are a lot of possibilities to vary the precursor delivery rate and the growth rate (pulse duration, injection frequency, pressure in the solution reservoir, solution concentration). The method is very versatile for in-situ deposition of complicated multilayers where several injectors can be used and their functioning parameters can be changed during the deposition process. Principal deposition conditions are presented in Table 1. Under these conditions films with a typical thickness of approximately 140 nm were deposited on LaAlO3 (100) substrates with the growth rate of approximately 0.6 mmyh. To study the biaxial strain, films of various thickness were deposited on LaAlO3, SrTiO3 and NdGaO3 substrates. The microstructure of the deposited films was invesTable 1 Deposition conditions Substrate temperature Evaporator temperature Total gas flow rate (ArqO2) Total pressure O2 pressure Precursors Solvent Solution concentration Injection frequency Mass of injected microdose a
650–825 8C 290 8C 105 lyh 5 torr 1.1–2.1 torr La(tmhd)3, Sr(tmhd)2, Mn(tmhd)3a Monoglyme 0.03 molyl (LaqSr) 0.5–3 Hz 4 mg
Abbreviation: tmhd, 2,2,6,6-tetramethyl-3,5-heptandionate.
34
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
tigated by measuring X-ray diffraction (XRD) spectra in Bragg–Brentano and Schultz geometry by means of a Siemens D5000 texture XRD diffractometer. The cparameter of the LSMO lattice (considered as cubic) was calculated from the position of peak (002) in uy2u scans. In order to remove systematic errors, the position of the peak was determined using substrate peaks as internal standard. The accuracy of XRD measurements ˚ The surface was estimated to be no less than 0.01 A. morphology of the films was studied by SEM. The SEM operates with EDX analysis, but the composition of La1yxSrxMnO3 films on LaAlO3 and SrTiO3 substrates cannot be determined by EDX, because the same elements are present in the film and in the substrate. Uncertain results are also obtained for films on other perovskite substrates. For this reason the film composition is not presented in this work; it could be approximately estimated from transport measurements, Curie temperatures and lattice constants. It is worth noting that these measurements suggested in most deposited La1yxSrxMnO3 films the x-values to be lower than the usually reported xs0.33. Electrical resistance measurements were carried out in a conventional four-probe configuration in the temperature range Ts100–400 K and in a magnetic field of 0–1.5 T. The thickness of the deposited YBCO films was measured by profilometry. The thickness of growing layers was also in-situ monitored by the optical reflection interference method, using a system of laser (ls670 nm) and silicon photodetector installed in the injection CVD reactor. 3. Results and discussion
Fig. 2. Effect of Sr content in solution on electrical properties of LSMO films on LaAlO3 substrate: (a) resistance (normalized to Rmax value) and temperature resistance coefficient (TCR) vs. temperature, (b) magnetoresistance (MR) vs. temperature. Inset in (b) shows the variation of the c-parameter of the LSMO lattice.
3.1. Effect of deposition conditions and annealing A series of LSMO films was prepared on LaAlO3 (100) substrates at 750 8C from solutions with the compositions La(1yx)SrxMn0.83, where x varied in the range of 0–0.5. After deposition the films were cooled to room temperature in oxygen (760 torr). In Fig. 2 the effect of solution composition on the semiconductor– metal transition temperature Tr, temperature coefficient of resistance (TCR) and magnetoresistance (MR) is presented for the deposited films. More Sr in solution (and in film) leads to a higher Tr but lowers TCR and MR values. Such dependence is typical for La1yxSrxMnO3 films with a variable Sr content, which determines the Mn4q yMn3q ratio. The figure demonstrates that Tr and MR values can be easily tuned by changing the Sr concentration in solution. In other series of depositions the influence of deposition temperature was studied. Films were deposited using two compositions of solutions differing in the Mny(LaqSr) ratio. Only slow cooling in pure oxygen atmosphere was used for film oxygenation. Fig. 3a,b demonstrates that the deposition temperature increased
the Tr values, but Tr was less sensible to Tdep when a solution more rich in Mn was used. The best films were deposited at 825 8C from solutions with the composition La0.6Sr0.4Mn0.63—they had Tr;320 K, TCR;4.5% (;300 K) and MR;43% (1.5 T, ;300 K). The peak resistivity of the films was ;10 mV cm and the resistivity at 100 K was ;130 mV cm. An interesting result is the fact that the TCR (and MR) values increased together with Tr values, i.e. in the opposite way compared to the effect of Sr contents. This fact suggests that the variation in film properties vs. deposition temperature does not result from a simple change in Mn4q y Mn3q ratio. The exact reason for such a dependence is not yet known. Several reasons may be taken into account, such as the improved epitaxial quality or the relaxation of biaxial strain at higher temperatures. However, all films deposited at different temperatures were of good epitaxial quality; only a small decrease in the width of the rocking curves and w-scan peaks was observed when the deposition temperature increased. The variation of the values of the lattice c-parameter (in ˚ did not allow to affirm that the range 3.885–3.875 A)
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
35
The importance of post-annealing procedure for the electrical properties of manganite films is underlined in the majority of works, therefore, the influence of annealing on the properties of our samples was also studied. The ex-situ annealing in oxygen (760 torr, 825 8C, 40 min) of films deposited at different temperatures (650– 825 8C) increased their Tr values by 15–25 K, but decreased both MR and TCR values (Fig. 4a). All annealed films had a slightly lower and approximately ˚ independently of the the same c-parameter (;3.87 A) deposition temperature. We also studied the effect of annealing duration on film properties (Fig. 4b). In this case films deposited at 750 8C were in-situ post-annealed at the same temperature in oxygen (760 torr). Increase of Tr and decrease of MR was observed for the films annealed for 30 min, and this result did not change noticeably after a prolonged (90 min) annealing procedure (Fig. 4b). This means that 30 min heating at 750 8C in pure oxygen is enough to reach an equilibrium
Fig. 3. Effect of deposition temperature (a,b) and Mn content in solution (c) on electrical properties of LSMO films on LaAlO3 substrates.
biaxial strain was clearly present and influenced the properties of these films deposited at different temperatures. One of the possible reasons may be the influence of the substrate temperature on the decomposition reaction of precursors, and probably this influence is greater in the case of Mn precursor. Consequently, the deposition temperature in MOCVD processing of LSMO films may determine the degree of deviation from Mn stoichiometry in them. Such explanation is supported by the fact that similar behavior wTCR (and MR) values increased together with Tr valuesx was also observed for the series of films deposited at the same temperature from solutions with a variable Mn content (Fig. 3c).
Fig. 4. Effect of annealing on electrical properties of LSMO films on LaAlO3 substrates: (a) influence of ex-situ annealing on Tr and MR values for films deposited at different temperature, (b) influence of the duration of in-situ annealing on electrical properties of films. Solution composition La0.63Sr0.37Mn0.83.
36
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
oxygen content in a manganite film. Annealing decreased the film resistivity; however, as in the case of ex-situ annealing, the MR and TCR values of in-situ annealed films were lower as compared to as-deposited films cooled in oxygen atmosphere. It is expected that annealing should increase oxygen content in manganite structure and consequently the Mn4q yMn3q ratio and Tr value. However, the same effect may be achieved by increasing the Sr content in films. We considered that annealing did not improve the properties of deposited La1yxSrxMnO3 films, and this procedure has been omitted in most deposition experiments. If the Tr value needs to be increased, a simple change of Sr content in solution is more convenient than the annealing procedure, which prolongs the film preparation time. However, cooling the film in oxygen atmosphere is crucial for LSMO film properties, and this procedure is obligatory. Such a conclusion is supported by Fig. 4b, where curve 4 represents a resistance vs. temperature dependence for the film deposited at 825 8C and cooled to room temperature in Ar atmosphere (Ps760 torr). One can see that such a treatment dramatically deteriorates the film properties as compared to those of a film cooled in oxygen atmosphere. The resistivity of a film cooled in Ar was by two orders higher than of a film cooled in oxygen. This effect can be caused by a significant oxygen deficiency in the film and demonstrates the importance of oxygen content for obtaining high Tr values w15x. A series of depositions was done to study the influence of oxygen partial pressure during deposition on the electrical properties of LSMO films. The results are presented in Fig. 5a. An increase of oxygen pressure in the reactor during the deposition decreased the Tr values and increased the TCR (and MR) values. The decrease of Tr by increasing the oxygen pressure is rather surprising, considering a possible enrichment of the structure by oxygen in the oxygen-richer environment. This suggests that the variation in oxygen stoichiometry is not the main reason for such a dependence of film properties vs. oxygen pressure during deposition. It is more likely that the variation of oxygen concentration changes the yields of precursor decomposition reactions and consequently the film composition. Most probably it is mainly the variation of Sr content that is responsible for such a behavior, because the changes of Tr, TCR and MR are similar to those caused by the effect of Sr content in the films. As an extension of our previous study on the influence of oxygen partial pressure, we present a series of films deposited with a different injection frequency of precursor solution (i.e. with a different solution delivery rate). La1yxSrxMnO3 films were deposited at an injection frequency 0.5, 1, 2 and 3 Hz. The electrical properties of the obtained films are presented in Fig. 5b which shows that the Tr values increase and the TCR values
Fig. 5. Effect of oxygen partial pressure (a) and injection frequency (b) on electrical properties of LSMO films deposited on LaAlO3 substrates at 825 8C, (c) electrical properties of films vs. the ratio of oxygen and vapor flow rates (calculated from the data of a and b). Solution composition La0.6Sr0.4 Mn0.63 . Insets in a and b show the variation of c-parameter of LSMO lattice and FWHM of rocking curve for LSMO (200) reflection.
decrease with an increase in injection frequency. The c˚ of deposited films was almost parameter (;3.875 A) independent of injection frequency (inset in Fig. 5b) as well as of oxygen partial pressure (inset in Fig. 5a). The injection frequency is directly related to the film growth rate, but it is rather difficult to realize why the
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
37
films are strongly governed by an excess of oxygen in respect to vapor concentration. 3.2. Effect of substrate material and film thickness
Fig. 6. Effect of substrate material on electrical properties of LSMO films deposited at 700 and 8258C. Solution composition La0.6Sr0.4Mn0.63.
electrical properties of epitaxial La1yxSrxMnO3 films should be so much dependent on the film growth rate. We tried to compare the film behavior presented in Fig. 5a,b and made an attempt to relate such behavior in both cases with an excess of oxygen concentration in respect to solvent and precursor vapor concentration. For this purpose we calculated the ratio of oxygen flow rate (Voxygen, lyh) to the total debit of solvent and precursor vapor (Vvapor, lyh) and presented the dependencies of Tr, TTCR(max), TCR (max) and MR (max) vs. Voxygen yVvapor for the samples obtained in two series— one with variable oxygen partial pressure and the other with variable injection frequency (Fig. 5c). One can see that both dependencies are rather similar. This means that the excess of oxygen in respect to vapor concentration is a very important factor in MOCVD of La1yxSrxMnO3 films. Most probably it influences film properties by influencing the precursor decomposition reaction and consequently the film composition. We performed some depositions with different microdose volume of precursor solution (which also corresponds to the variation of Voxygen yVvapor) and obtained a similar result to confirm that the electrical properties of LSMO
The substrate material undoubtedly influenced the electrical properties of LSMO films. In Fig. 6 the electrical properties of LSMO films deposited on various perovskite substrates at 700 and 825 8C are presented. We see that the Tr values slowly increase in the range SrTiO3 (001)–NdGaO3 (001)–LaAlO3 (001), i.e. simultaneously with a decrease of the substrate lattice ˚ respectively). Such a parameter (3.905–3.85–3.79 A, result was well reproducible and almost the same for the films deposited at two temperatures (700 and 825 8C). This effect of perovskite substrate is difficult to explain. NdGaO3 substrate and La1yxSrxMnO3 films have the best matched lattice parameters. Thus one can suppose that the higher Tr of the film on LaAlO3 may be caused by a compressive strain imposed by the lattice mismatch with the substrate, while the lower Tr on SrTiO3 may be determined by the tension strain in the film. However, let us remind that the film thickness was approximately 140 nm and in such rather thick films deposited at a high temperature (825 8C) the biaxial strain could hardly remain not relaxed. The fact that the c-parameter determined for the films grown at 825 8C on LaAlO3 and SrTiO3 substrates is almost the same ˚ on SrTiO3 and ;3.875 A ˚ on LaAlO3) is in (;3.87 A favor of this conclusion. For the films grown at 700 8C, the difference in c-parameters is not great, but more ˚ on SrTiO3 and 3.885 A ˚ on pronounced (3.87 A LaAlO3), and in this case we can suppose that biaxial strains in these films may not be completely relaxed (at a lower temperature the kinetics more strongly influences the crystal growth and thicker strained films can be grown). In order to study in more detail the possible strain and its influence on the film properties, we deposited a series of films with different thickness (;8–300 nm) at 825 8C on three perovskite substrates. The results of this study are presented in Figs. 7 and 8. Fig. 7c shows that in the range of thickness ;60–300 nm the Tr values for the films on three different perovskite substrates almost did not depend on film thickness and always followed the same order indicated above (Tr(LAO))Tr(NGO))Tr(STO)). The curves of lattice parameter vs. film thickness (Fig. 7a) demonstrate that only at small thicknesses (less than approx. 60 nm) the strain is certainly present in the films on perovskite substrates and influences the film properties (Fig. 7c and d). Tensile strain is present in films on SrTiO3, and compressive strain is observed in films on LaAlO3 (and less visible on NdGaO3) substrates. Both tensile and compressive strains decreased the values of magnetoresistance (except on NdGaO3 substrates, where the strain was less) and the temperature of electrical transition.
38
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
Fig. 7. Effect of film thickness on the properties of LSMO films deposited on perovskite substrates at 825 8C. Solution composition La0.6Sr0.4Mn0.63.
The strain is completely relaxed in the films thicker than ;100 nm. There are numerous literature sources on the biaxial strain and its influence on the properties of epitaxial La1yx(Ca,Sr)xMnO3 films w15–22x, but the published results are very contradictory, especially those concerning the critical thickness up to which the films can be grown coherently strained on LaAlO3 and SrTiO3 substrates. According to some references w19,20x, the biaxial strain influencing the film properties was found only in very thin films (up to 10–12 nm), but the possibility to grow films without misfit dislocations even up to 100 nm thick under well optimized conditions was not excluded w19x. Other authors w18,21x found that thicker films (100–200 nm) contained a coherently strained bottom layer up to 40 or 60 nm and a relaxed top layer. Some authors present the lattice parameters undoubtedly
demonstrating the presence of biaxial strain in thick manganite films (150 nm w15x or even 450 nm w22x). Returning to our films, the reason for the difference in Tr of rather thick LSMO films on SrTiO3, NdGaO3 and LaAlO3 is not clearly understood at the moment and cannot be explained solely by the effect of the residual biaxial strain. A more detailed study is needed to elucidate the exact reason for this effect of perovskite substrate material on film properties. La1yxSrxMnO3 films on perovskite substrates SrTiO3, NdGaO3 and LaAlO3 were grown epitaxially. In all deposited films the (001) texture and ‘cube-on-cube’ inplane orientation were found. No double LSMO (002) peak structure indicating the presence of a strained and a relaxed part of the film with different lattice parameters was observed in the uy2u spectra, contrary to the results presented in Konishi et al. w18x and Wiedenhorst et al.
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
39
Fig. 8. Effect of film thickness on magnetoresistance of LSMO films deposited on LaAlO3 (a,b) and SrTiO3 (c,d) substrates at 825 8C. Solution composition La0.6Sr0.4Mn0.63.
w21x for the films of similar thickness. Film thickness influenced the width of the rocking curves—the lowest values of the full width at half maximum (FWHM) were obtained in biaxially strained thin films (Fig. 7b). The magnetoresistance of LSMO films of different thickness on LaAlO3 and SrTiO3 substrates was measured in moderate magnetic fields from 0.25 to 1.5 T. The results of such measurements are presented in Fig. 8. In the whole range of the magnetic fields the influence of strain on magnetoresistance was observed, which was expressed by evidently lower MR values for thin films as compared to thicker ones. A surprising behavior was demonstrated by films on LaAlO3 substrates; contrary to films on SrTiO3, the MR values did not saturate up to a film thickness of 300 nm, while the Tr values and the lattice parameter became almost constant above the thickness of approximately 60 nm (Fig. 7a,c). It is
difficult to explain such a behavior, especially as the strain seems to be relaxed in thicker films on LaAlO3. It is possible that such a behavior is defined by a greater contribution of intergranular magnetoresistance in thicker films whose epitaxial quality is lower as compared with thinner films. In order to compare perovskite and non-perovskite substrates, some La1yxSrxMnO3 films were deposited under similar conditions on non-perovskite substrates YSZ (001), MgO (001) and sapphire (R-plane). Even the best films had a broader semiconductor–metal transition and distinctly lower Tr values (;270 K on YSZ, ;240 K on MgO and ;210 K on sapphire), possibly because of an incomplete epitaxy of films on these less matched substrates. The determined lattice c-parameter ˚ did not indicate the presence of biaxial strain (;3.87 A) in the films.
40
A. Abrutis et al. / Thin Solid Films 413 (2002) 32–40
LSMO films deposited on LaAlO3 substrates had a smooth surface up to the thickness of approximately 150–160 nm. The surface of thicker films ()200 nm) was rather rough. We chose the film thickness between 100 and 150 nm as most suitable for the preparation of multilayered heterostructures. This allowed us to deposit heterostructures LaAlO3 yyLa1yxSrxMnO3 ySrTiO3 y YBa2Cu3O7-d, in which the effect of spin-polarized quasiparticle injection from ferromagnetic to superconducting layers was observed w23x. 4. Conclusions In conclusion, the pulsed liquid injection MOCVD technique allows to deposit high quality epitaxial La1yxSrxMnO3 films. MOCVD film properties are very sensitive to deposition conditions and only well optimized conditions provide for the best properties. The factors that exert the strongest influence on film properties are deposition temperature, solution composition and oxygenyvapor ratio. In-situ or ex-situ high-temperature annealing of the deposited films in pure oxygen increased the temperature of semiconductor–metal transition, but decreased the magnetoresistance and the temperature coefficient of resistance. The best films were deposited at 825 8C on LaAlO3 substrates: they exhibited sharp semiconductor–metal transition and high magnetoresistance (;40%) at room temperature and in a rather low field of 1.5 T (sensibility approx. 0.003%y Oe). To our knowledge, this is the best result for La1yxSrxMnO3 films deposited by MOCVD, which successfully competes with the best results obtained by physical vapor deposition techniques. Good epitaxial quality of La1yxSrxMnO3 films is necessary to obtain a high transition temperature, sharp transition and high magnetoresistance in a medium field in the same time, so in this case perovskite substrates are more suitable. A study of relation between the film thickness (8–300 nm) and film properties was performed, and the biaxial strain imposed by the lattice mismatch was found in the thinner La1yxSrxMnO3 films on perovskite substrates. Tensile strain was present in the films on SrTiO3 and compressive strain was observed in the films on LaAlO3 (and NdGaO3). Both tensile and compressive strains decreased the temperature of electric transition and the values of magnetoresistance. The strain was completely relaxed in the films more than ;100 nm thick.
Acknowledgments The authors acknowledge partial financial support from European Community (FP5, MULTIMETOX network, contract C5RT-CT-2000-05001). References w1x Y.Q. Li, J. Zhang, S. Pombric, D. Mascio, W. Stevens, Y.F. Yan, N.P. Ong, J. Mater. Res. 10 (1995) 2166. w2x O.Y. Gorbenko, A.R. Kaul, N.A. Babushkina, L.M. Belova, J. Mater. Chem. 7 (1997) 747. w3x P.B. Tavares, V.S. Amaral, J.P. Araujo, J.B. Sousa, A.A.C.S Lourenco, J.M. Vieira, J. Appl. Phys. 85 (1999) 5411. w4x G.J. Snyder, R. Hiskes, S. DiCarolis, M.R. Beasley, T.H. Gebale, Phys. Rev. B 53 (1996) 14434. w5x D. Studebaker, M. Todd, C. Seegel, T.H. Baum, Mater. Sci. Eng. B 56 (1998) 168. w6x V. Moshnyaga, I. Khoroshun, A. Sidorenko, P. Petrenko, A. Weidinger, M. Zeitler, B. Rauschenbach, R. Tidecks, K. Samwer, Appl. Phys. Lett. 74 (1999) 2842. w7x K.H. Dachmen, M.W. Carris, Chem. Vapor Depos. 3 (1997) 27. w8x J.J. Heremans, M. Carris, S. Yu, X. Watts, K.H. Dahmen, S. von Molnar, J. Appl. Phys. 81 (1997) 4967. w9x S. Pignard, H. Vincent, J.P. Senateur, J. Pierre, A. Abrutis, J. Appl. Phys. 82 (1997) 4445. w10x F. Felten, J.P. Senateur, F. Weiss, R. Madar, A. Abrutis, J. Phys. IV C 5 (1995) 1079. w11x A. Abrutis, J.P. Senateur, F. Weiss, V. Kubilius, V. Bigelyte, Z. ˇ Saltye, B. Vengalis, A. Jukna, Supercond. Sci. Technol. 10 (1997) 959. w12x A. Abrutis, J.P. Senateur, F. Weiss, V. Bigelyte, A. Teiserskis, ˇ V. Kubilius, V. Galindo, S. Balevicius, J. Cryst. Growth 191 (1998) 79. w13x A. Abrutis, V. Plausinaitiene, A. Teiserskis, V. Kubilius, J.P. Senateur, F. Weiss, Chem. Vapor Depos. 5 (1999) 171. w14x V. Galindo, J.P. Senateur, A. Abrutis, A. Teiserskis, F. Weiss, J. Cryst. Growth 208 (2000) 357. w15x J.R. Sun, C.F. Yeung, K. Zhao, L.Z. Zhou, C.H. Leung, H.K. Wong, B.G. Shen, Appl. Phys. Lett. 76 (2000) 1164. w16x S. Jin, T.H. Tiefel, M. McCormack, H.M. O’Bryan, L.H. Chen, R. Ramesh, D. Schurig, Appl. Phys. Lett. 67 (1995) 557. w17x A.J. Millis, T. Darling, A. Migliori, J. Appl. Phys. 83 (1998) 1588. w18x Y. Konishi, M. Kasai, M. Izumi, M. Kawasaki, Y. Tokura, Mater. Sci. Eng. B 56 (1998) 158. w19x M. Kawasaki, M. Izumi, Y. Konishi, T. Manako, Y. Tokura, Mater. Sci. Eng. B 63 (1999) 49. w20x J.Z. Sun, D.W. Abraham, R.A. Rao, C.B. Eom, Appl. Phys. Lett. 74 (1999) 3017. w21x B. Wiedenhorst, C. Hofener, Y. Lu, J. Klein, L. Alff, R. Gross, B.H. Freitag, W. Mader, Appl. Phys. Lett. 74 (1999) 3636. w22x Y. Wu, Y. Suzuki, U. Rudiger, J. Yu, A.D. Kent, T.K. Nath, C.B. Eom, Appl. Phys. Lett. 75 (1999) 2295. w23x V. Plausinaitiene, A. Abrutis, B. Vengalis, R. Butkute, J.P. Senateur, Z. Saltyte, V. Kubilius, Physica C 351 (2001) 13.
