Magnetic behavior of epitaxial SrRuO<sub>3</sub> thin films under pressure up to 23 GPa

(1)

Article

Reference

Magnetic behavior of epitaxial SrRuO

₃

thin films under pressure up to 23 GPa

LE MARREC, Françoise, et al.

Abstract

We report on measurements of electrical resistivity and determination of the Curie temperature of single-crystal epitaxial ferromagnetic SrRuO3 thin films under hydrostatic pressures up to 23 GPa. The SrRuO3 thin film was grown on vicinal (001) SrTiO3 substrate using 90° off-axis magnetron sputtering. At atmospheric pressure, the thin-film Curie temperature is about 150 K. To generate the high pressure, the Bridgman anvil technique was used. Up to 13 GPa, a linear decrease of the Curie temperature was observed at a rate of 5.9 KGPa-1. Above 13 GPa, a striking saturation of the magnetic transition temperature at 77 K was observed.

LE MARREC, Françoise, et al . Magnetic behavior of epitaxial SrRuO

₃

thin films under pressure up to 23 GPa. Applied Physics Letters , 2002, vol. 80, no. 13, p. 2338

DOI : 10.1063/1.1459484

Available at:

http://archive-ouverte.unige.ch/unige:35753

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Magnetic behavior of epitaxial SrRuO

₃

thin films under pressure up to 23 GPa

F. Le Marrec,^a)A. Demuer,^b)D. Jaccard, and J.-M. Triscone

DPMC, University of Geneva, 24 quai Ernest Ansermet, 1211 Geneva 4, Switzerland M. K. Lee and C. B. Eom

Department of Materials Science and Engineering, University of Wisconsin–Madison, 1500 Engineering Drive, Madison, Wisconsin 53706

共Received 19 November 2001; accepted for publication 14 January 2002兲

We report on measurements of electrical resistivity and determination of the Curie temperature of single-crystal epitaxial ferromagnetic SrRuO₃ thin films under hydrostatic pressures up to 23 GPa.

The SrRuO₃ thin film was grown on vicinal 共001兲SrTiO₃ substrate using 90° off-axis magnetron sputtering. At atmospheric pressure, the thin-film Curie temperature is about 150 K. To generate the high pressure, the Bridgman anvil technique was used. Up to 13 GPa, a linear decrease of the Curie temperature was observed at a rate of 5.9 K GPa^⫺¹. Above 13 GPa, a striking saturation of the magnetic transition temperature at 77 K was observed. © 2002 American Institute of Physics.

关DOI: 10.1063/1.1459484兴

Over the last decade, perovskite oxide materials have attracted increasing attention due to their variety of physical properties, ranging from high T_C superconductivity and co- lossal magnetoresistance, to ferroelectricity and ferromagnetism.¹These diverse electronic properties, occur- ring in materials with very similar crystal structures, are promising for new applications and devices. For some spe- cific devices, it might be an advantage to use materials in epitaxial thin-film form. A potential problem, however, is that lattice and thermal expansion mismatch between the thin film and the substrate can cause strain which can modify the genuine physical properties of the material. In order to use these new materials in an optimal way, it is particularly im- portant to understand and control the role of strain on the properties of epitaxial perovskite thin films. The most com- monly used approach to study the influence of strain is to grow epitaxial thin films, varying both the thickness and the nature of the substrate. For instance, Locquet et al. achieved a doubling of the critical temperature of the copper oxide superconductor La_1.9Sr_0.1CuO₄ using compressive epitaxial strain caused by SrLaAlO₄ substrates.²Another method car- ried out by Gan et al. on epitaxial SrRuO₃ thin films con- sisted of studying the same sample before and after strain relaxation by film lift off from the substrate.³ In this latter work, the relaxation of the lattice strain resulted in a 10 K increase of the Curie temperature, up to the bulk value. In epitaxial SrRuO₃ films on BaTiO₃ substrates, the structural phase transitions of BaTiO₃ were also used to study the influence of different strain states on the magnetic and trans- port properties of SrRuO₃.⁴ In all of these examples, the epitaxial pressure leads to biaxial strains and relatively small volume cell changes. Large biaxial strains in thin films re- quire coherent growth on a substrate with a large lattice mismatch. Since dislocations appear above a critical thickness,

which goes down with increasing lattice mismatch, the range of possible film thicknesses is limited for large strain.

In this letter, we report an approach to generate very high pressure 共up to 23 GPa兲 in thin films, allowing us to study the influence of large strain on the physical properties of the material. We note that experiments under hydrostatic pressure have previously been carried out on high T_Ccuprate thin films with a piston cylinder technique allowing a maximum pressure of about 2 GPa.⁵To generate very high pressure, we adapted for thin films the conventional Bridgman technique developed to investigate bulk materials.⁶ This pressure technique used here to study SrRuO₃can be applied to other oxide thin films and heterostructures, opening up the possibility of finely tuning the electronic properties of a par- ticular system.

Electrical resistivity measurements were performed on the ferromagnetic metallic oxide SrRuO₃. The chemical sta- bility of this material is an advantage during the delicate preparation of the very small samples required for the pressure setup. SrRuO₃ is a metallic perovskite with a GdFeO₃-type structure 共the lattice parameter of the pseudocubic cell is 3.93 Å兲which displays an itinerant ferromagnetism with a bulk Curie temperature of 160 K.⁷Mag- netic ordering manifests itself at T_C as a kink in the resistivity versus temperature curve due to reduced spin scattering below the Curie temperature. Measurements were performed under hydrostatic pressures up to 23 GPa and temperatures down to 1.25 K.

The epitaxial SrRuO₃thin film, 350 Å thick, was depos- ited on miscut 共001兲 SrTiO₃ substrate by 90° off-axis sputtering.⁸ X-ray analysis revealed that epitaxial SrRuO₃ thin films are single domains with the pseudocubic direction 关100兴normal to the substrate. The out-of-plane lattice parameter is 3.95 Å, a value slightly larger than the 3.93 Å mea- sured for bulk materials⁹ demonstrating a uniaxial tensile strain along the 关100兴direction in the film. This can be un- derstood if the film growth is coherent resulting in a com-

a兲Electronic mail: [email protected]

b兲Author to whom all correspondence should be addressed; electronic mail:

[email protected]

APPLIED PHYSICS LETTERS VOLUME 80, NUMBER 13 1 APRIL 2002

2338

Downloaded 14 Aug 2002 to 129.194.8.73. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp

(3)

pressive stress in the plane and an elongated out-of-plane parameter.

For the pressure experiment, the SrRuO₃ sample was reduced by very slow repeated cutting with a thin diamond saw and by a final polishing共with the diamond coated side of the saw兲 of the substrate down to a rectangular shape of 300␮^m⫻230␮m and a thickness of 30␮m. The Bridgman anvil technique used to generate pressure consists of pressing a high pressure cell between two opposed anvil made of sintered diamond. The diameter of the anvil flat (␾

⫽2 mm兲allowed a maximum pressure of around 25 GPa to be generated.

The high pressure chamber is realized inside a pyrophyl- lite ring gasket, as shown in Fig. 1. The sample is squeezed between steatite disks used as the pressure transmitting medium. Pressure was determined via the resistive superconducting transition of a lead manometer.¹⁰ Six annealed Au wires were placed on the sample allowing different four- point dc configurations for the resistivity measurements. One of these configurations was chosen for all the measurements reported here. An additional configuration was used both at 293 and 1.25 K to calculate the absolute resistivity,␳^{, by the} van der Pauw method.¹¹ Upon increasing the pressure from 5.19 to 23.11 GPa, the ambient temperature resistivity fell continuously from 230 to 214 ␮⍀ cm, suggesting that the film is not damaged during pressurization.¹²␳(T) is obtained using a four-point measurement and the room temperature resistivity value.

Figure 2共a兲shows ␳(T) curves for SrRuO₃ thin film for three selected pressures: 0 共atmospheric pressure兲, 5.19 and 21.86 GPa. At atmospheric pressure, the measurements were performed on the 1 cm⫻1 cm specimen from which our sample was cut. The ambient temperature resistivity was obtained by extrapolating ␳^{( P,T}⫽293 K) to P⫽0.

For T⬎T_C, the resistivity ␳(T) linearly increases with temperature with a slope almost independent of pressure. At T_C, the kink becomes less pronounced as the pressure increases, probably because of the pressure gradient in the transmitting medium. This gradient is estimated by the total width of the superconducting transition of the lead manometer. We found that the gradient increased with pressure from about 0.2 GPa at P⫽5.19 GPa to 0.6 GPa at P

⫽23.11 GPa.

The effect of pressure on the Curie temperature can be more clearly seen in Fig. 2共b兲 where the derivative of the temperature dependence of the resistivity is plotted at vari- ous pressures. The d␳/dT plots were obtained by fitting the

resistivity at each pressure using a ninth-order polynomial and further differentiation. As the pressure is increased up to 13.30 GPa, the maximum in d␳/dT versus T shifts towards lower temperatures. For pressures above 13.30 GPa, the maximum in the d␳/dT versus T becomes independent of pressure. For each pressure, we defined T_C as the tempera- ture of the midpoint of the jump in d␳/dT observed in Fig.

2共b兲共see filled squares兲. This definition gives a Curie temperature very close to that determined using the break point on the ␳versus T plot关arrows in Fig. 2共a兲兴.

Figure 3 shows the pressure dependence of the ferromagnetic Curie temperature for the SrRuO₃ thin film. Up to 13 GPa, we observed a fairly linear decrease in the transition temperature at a rate of 5.9 K GPa^⫺¹共an extrapolation would lead to T_C⫽0 K for P⫽27 GPa兲. This value is in good agreement with the 5.7 K GPa^⫺¹observed for bulk polycrys- talline SrRuO₃investigated up to 6 GPa.¹³A striking feature of the data is the saturation of the magnetic transition temperature around 77 K for pressures higher than 13 GPa. This behavior does not depend on the choice of criterion for the determination of the transition temperature.

In our experiment, increasing the pressure produces a reduction of the cell volume. The nature of the pressure, purely hydrostatic or partially uniaxial, depends on the rela- tive values of the bulk moduli for SrRuO₃ and SrTiO₃. The

FIG. 1. Top view of the high pressure cell before closing.

FIG. 2. 共a兲 Resistivity versus temperature of SrRuO3 thin film at three pressures共0, 5.19, and 21.86 GPa兲. The arrows indicate the Curie temperature. 共b兲 Derivative of the temperature dependence of the resistivity for various pressures. TC, defined as the temperature of the midpoint of the jump in d␳/dT plot, is indicated by the filled symbols.

FIG. 3. Pressure dependence of the magnetic transition temperature for the SrRuO3thin film. The solid line represents the bulk behavior. The arrow indicates the pressure at which the cell volume is equal to that of CaRuO3. The dotted lines are a guide for the eye.

2339

Appl. Phys. Lett., Vol. 80, No. 13, 1 April 2002 Le Marrecet al.

(4)

value is unfortunately only known for SrTiO₃ 共176 GPa兲.¹⁴ Assuming that the bulk moduli of SrRuO₃ and SrTiO₃ are close and thus that the pressure is essentially hydrostatic共this is supported by the close comparison between the bulk data of Ref. 13 and our data, see Fig. 3兲, the change in cell volume produced by the highest pressure achieved, 23.11 GPa, is 10% according to the Murnaghan equation of state.¹⁵

A reduction of the Curie temperature has been observed in the alloy compounds Ca_xSr₁_⫺_xRuO₃ where T_C seems to vanish for x⫽0.7.¹⁶ The substitution by Ca²^⫹ induces a chemical pressure, so that the lattice parameter changes from 3.93 Å (x⫽0) for pure SrRuO₃ to 3.84 Å (x⫽1) for pure CaRuO₃, corresponding to a 6.7% reduction in cell volume.

Simultaneously, this substitution induces a larger orthorhombic distortion. Band structure calculations by Mazin and Singh have suggested that the rotation of the oxygen octahe- dra, responsible for this distortion, is a key parameter con- trolling the magnetism in this compound.¹⁷ The change in cell volume between SrRuO₃ (T_C⫽160 K) and Ca_0.7Sr_0.3RuO₃(T_C⫽0 K) is about 4.5%, corresponding to a pressure of about 11 GPa. The data, up to 13 GPa, can thus not be simply ‘‘mapped’’ onto the Ca_xSr₁_⫺_xRuO₃ behavior.

Such an explanation would lead to a 兩dT_C/d P兩 of 15 K GPa^⫺¹ whereas we observe a rate of 5.9 K GPa^⫺¹ in our experiment.

The most striking feature of the data is the sudden saturation of the Curie temperature at 77 K above 13 GPa. This behavior might be related to the complex magnetism which develops in CaRuO₃ below 70 K.^18,19 This compound was proposed to be poised between ferromagnetic and paramag- netic ground states, and the large distortion共rotation兲, twice that of SrRuO₃, could eventually lead to this complex weak magnetism. The pressure dependence of the orthorhombic distortion in our measurements is unknown. However, the persistence of a magnetic signature in our resistivity measurements suggests long range magnetic ordering. A pressure induced phase transition could be at the origin of the ob- served change in T_C( P).

In conclusion, we have performed measurements of electrical resistivity on epitaxial ferromagnetic metallic oxide SrRuO₃ thin films under hydrostatic pressures up to 23 GPa.

Below 13 GPa, a linear decrease of the Curie temperature is observed at a rate of 5.9 K GPa^⫺¹. For higher pressures, the Curie temperature is found to saturate at 77 K.

The authors thank T. Jarlborg for helpful discussions and A. T. Holmes for a careful reading of the manuscript. This work was supported by the Swiss National Science Founda- tion through the National Center of Competence in Research

‘‘Materials with Novel Electronic Properties-MaNEP’’ and Division II.

1A. R. West, Solid State Chemistry and its Applications共Wiley, New York, 1992兲.

2J. P. Loquet, J. Perret, J. Fompeyrine, E. Ma¨chler, J. W. Seo, and G. Van Tendeloo, Nature共London兲394, 453共1998兲.

3Q. Gan, R. A. Rao, C. B. Eom, J. L. Garret, and M. Lee, Appl. Phys. Lett.

72, 978共1998兲.

4M. K. Lee, T. K. Nath, C. B. Eom, M. C. Smoak, and F. Tsui, Appl. Phys.

Lett. 77, 3547共2000兲.

5S. L. Bud’ko, J. Guimpel, O. Nakamura, M. B. Maple, and I. K. Schuller, Phys. Rev. B 46, 1257共1992兲.

6D. Jaccard, E. Vargoz, K. Alami-Yadri, and H. Wilhelm, Rev. High Pres- sure Sci. Technol. 17, 412共1998兲.

7A. Kanbayasi, J. Phys. Soc. Jpn. 41, 1876共1976兲.

8C. B. Eom, R. J. Cava, R. M. Fleming, J. M. Phillips, R. B. van Dover, J.

H. Marshall, J. W. P. Hsu, J. J. Krajewski, and W. F. Peck, Jr., Science 258, 1766共1992兲.

9R. J. Bouchard and J. L. Gillson, Mater. Res. Bull. 7, 873共1972兲.

10B. Bireckhoven and J. Witting, J. Phys. E 21, 841共1988兲.

11L. J. Van der Pauw, Philips Res. Rep. 13, 1共1958兲.

12One can not exclude however a progressive introduction of dislocations.

13J. J. Neumeier, A. L. Cornelius, and J. S. Schilling, Physica B 198, 324 共1994兲.

14L. R. Edwards and R. W. Lynch, J. Phys. Chem. Solids 31, 573共1969兲.

15F. D. Murnaghan, Proc. Natl. Acad. Sci. U.S.A. 30, 244共1944兲.

16A. Kanbayasi, J. Phys. Soc. Jpn. 44, 108共1978兲.

17I. I. Mazin and D. J. Singh, Phys. Rev. B 56, 2556共1997兲.

18T. He and R. J. Cava, J. Phys.: Condens. Matter 13, 8347共2001兲.

19I. Felner, I. Nowik, I. Bradaric, and M. Gospodinov, Phys. Rev. B 62, 11332共2000兲.

2340 Appl. Phys. Lett., Vol. 80, No. 13, 1 April 2002 Le Marrecet al.