- Email: [email protected]
Large magnetoresistance in double perovskite Sr2Cr1.2Mo0.8O6-δ
Pergamon
Materials Research Bulletin 36 (2001) 705–715
Large magnetoresistance in double perovskite Sr2Cr1.2Mo0.8O6-␦ Z. Zenga,1, I. D. Fawcetta, M. Greenblatta,*, M. Croftb a
Department of Chemistry, Rutgers State University of New Jersey, Piscataway, NJ 08854-8087, USA b Department of Physics, Rutgers State University of New Jersey, Piscataway, NJ 08854-8087, USA (Refereed) Received 12 July 2000; accepted 17 July 2000
Abstract Sr2Cr1.2Mo0.8O6-␦ (␦⫽0, 0.2) with a double perovskite structure was prepared by solid state reaction in evacuated quartz tubes. Cr and Mo are partially ordered on the B site. Oxygen vacancies decrease the ordering, but increase the lattice parameters. X-ray absorption spectroscopy results are consistent with Cr being 3⫹, and Mo being close to 5⫹ for ␦ ⫽ 0.2 and 5.5 for ␦ ⫽ 0. The spin of Cr3⫹ (d3) and Mo5⫹ (d1) order in an anti-parallel arrangement by superexchange interaction, and lead to ferrimagnetic ordering below 465 K. Both compounds are n-type narrow gap semiconductors. Large magnetoresistance (-43%) is observed in Sr2Cr1.2Mo0.8O6. The MR behavior is attributed to an intra-grain tunneling mechanism. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: A. Magnetic materials; B. Chemical synthesis; C. XANES; C. Electronic conductivity; D. Magnetic properties
1. Introduction Recent discovery of colossal magnetoresistance (CMR) in perovskite manganates has stimulated much interest [1–11]. CMR materials are potentially useful for applications in magneto-electronic devices. It is important to search for new magnetoresistance materials to improve the properties and understand the mechanism of CMR.
* Corresponding author. Fax: ⫹1-732-445-5312. E-mail address: [email protected] (M. Greenblatt). 1 Present address: Energy Technology Division, Argonne National Laboratory, Argonne, IL. 0025-5408/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 5 - 5 4 0 8 ( 0 1 ) 0 0 5 2 0 - 7
706
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 1. Magnetic structure of a ferrimagnetically ordered double perovskite.
Perovskites have the general formula ABX3, in which A represents a large electropositive cation, B is a small transition metal or main group ion, and X is commonly an oxide or halide ion. Double perovskites with the formula A’A”B’B”O6, with metal ions B’ and B”, may order on the perovskite B lattice site if the size/charge of the B’ and B” ions are different. The ordered double perovskites with paramagnetic transition metal, B’ and B” ions can exhibit magnetic ordering as well. Some double perovskite materials have high (⬎room temperature) magnetic ordering temperature [12–17]. Fig. 1 illustrates the magnetic structure of an aniferromagnetically (AF)-ordered double perovskite. In such materials, if the number of d electrons of the B’ and B” ions are different, the magnetic coupling leads to ferrimagnetic order, as in Fig. 1. Recently, Kobayashi et al. reported large low-field MR in the double perovskite Sr2FeMoO6 [18,19]; the MR was attributed to inter-grain tunneling mechanism. Subsequently, this mechanism was confirmed by the absence of MR in single crystals of Sr2FeMoO6 [20 –22]. Patterson et al reported on Sr2CrMoO6, which was contaminated by strontium molybdate [12]. More recently Moritomo et al investigated the electronic structure of Sr2MMoO6 (M ⫽ Cr, Mn, Fe, and Co), but the M ⫽ Cr phase was not well characterized [23]. In this paper we report the synthesis of single-phase
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
707
Sr2Cr1⫹xMoxO6-␦ (x ⫽ 0.1, 0.2; ␦ ⫽ 0, 0.2) and the properties of Sr2Cr1.2Mo0.8O6-␦ with relatively large MR.
2. Experimental Sr2Cr1⫹xMoxO6-␦ (x ⫽ 0.1, 0.2; ␦ ⫽ 0, 0.2) were prepared by solid-state reaction. All starting materials were purchased from Aldrich. Stoichiometric SrO (obtained by decomposition of SrCO3 at 1300 °C), Cr2O3, MO3 and Mo were ground in a dry box. The mixtures were pressed into pellets and sealed into evacuated quartz tubes. The pellets were sintered at 550 °C for 10 h, and then the temperature was increased to 1000 °C and the sample sintered for an additional 24 h. The resulting products were reground and pressed into pellets again and sintered at 1000 °C for 24 h in evacuated quartz tubes The powder X-ray diffraction (PXD) data were collected with a SCINTAG PAD V diffractometer with Ni-filtered CuK␣ radiation over the range 10° ⱕ 2 ⱕ 90° with a step size of 0.02 °. Silicon powder was used as an internal standard. Lattice parameters were refined by a least-square method. Rietveld refinement of the data was undertaken with the GSAS Rietveld refinement program [24]. A broad maximum seen in the background of the X-ray diffraction patterns is due to the glass slide and Vaseline used to hold the sample. The backgrounds of the PXD data in the Rietveld refinements were fitted with an 8-term Chebyshev polynomial and a pseudo-Voigt function was employed to model the peak shapes. The dc electrical resistivity and the magnetoresistance measurements were carried out by a standard four-probe technique from 400 to 4 K in a SQUID magnetometer (MPMS, Quantum Design). The magnetic properties were also measured with the SQUID magnetometer in the temperature range 4 – 400 K. The Mo L3-edge and Cr K-edge X-ray absorption spectroscopy (XAS) measurements were performed on beam line X-19A at the Brookhaven national synchrotron light source with a double crystal [InSb (111) or Si (111)] monochromator. The measurements were made in the fluorescence mode. A standard was run simultaneously with all the measurements for precise calibration of the Cr-spectra to a relative accuracy of about ⫾0.03 eV. Standard samples were run frequently in the Mo measurements for calibration with a lower accuracy of about ⫾0.8 eV. All spectra were normalized to unity step in the absorption coefficient from well below to well above the edge.
3. Results and discussions 3.1. Synthesis Attempts to prepare Sr2CrMoO6 double perovskite in an evacuated quartz tube resulted in a black powder contaminated with SrMoO4. Single-phase samples of Sr2Cr1⫹xMo1-xO6-␦, could be obtained only with Cr/Mo ⬎ 1.0 in a narrow range of composition (0.1 ⱕ x ⱕ 0.2), and the raw materials must be dried thoroughly. The quartz tubes should be flamed three
708
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 2. Cr K-edge X-ray absorption specta of Sr2Cr1.2Mo0.8O6, CrO2, LaCrO3, Cr2O3, and Cr-metal. In the Inset the pre-edge features of the oxide spectra are compared (with the Sr2Cr1.2Mo0.8O6 spectrum displaced downward for clarity).
times before sealing. The x ⫽ 0.1 material has orders of magnitude higher resistivity than the x ⫽ 0.2 compounds. Therefore the latter compounds were studied in detail 3.2. Oxidation state of Cr and Mo Figure 2 shows the Cr K-edge XAS spectra of Sr2Cr1.2Mo0.8O6, the Cr3⫹ standards LaCrO3 and Cr2O3, and the Cr4⫹ standard CrO2. The Cr-K edge spectra of both Sr2Cr1.2Mo0.8O6-␦ ␦ ⫽ 0 and 0.2 compounds were essentially identical so only one has been displayed. The rising part of the Sr2Cr1.2Mo0.8O6 main edge can be seen to have a chemical shift identical to that of the Cr3⫹ standards (see the box in the Figure) and well separated from that of the Cr4⫹ standard. The pre-edge features of the Cr-K edge spectra are sown in the inset of Figure 2. Specifically the Sr2Cr1.2Mo0.8O6-␦ pre-edge manifests a three-feature (a1, a2, and a3) structure, similar to the Cr3⫹ standards, whereas the Cr4⫹ standard has a quite distinct bimodal-feature structure. Among the Cr3⫹ materials the sharp-low-energy a1-feature is particularly distinctive with the intensities and positions of the a2 and a3
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
709
Fig. 3. (bottom) The Mo L-edge X-ray absorption spectra of Mo-metal, MoO2, Sm2Mo2O7 (from ref. 24), and MoO3 standard materials. (top) The Mo L-edge spectra of Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2) along with the MoO2 standard.
features varying between compounds. Thus the structure of the pre-edge further supports the assignment of a Cr3⫹ state for these Sr2Cr1.2Mo0.8O6-␦ materials. From the chemical formula, if the valence of Cr is 3⫹, Mo should be formally 5.5⫹ in Sr2Cr1.2Mo0.8O6, and 5⫹ in Sr2Cr1.2Mo0.8O5.8. In Figure 3-bottom, the Mo L3-edges of the Mo0, Mo4⫹O2, Sm2Mo4⫹2O7 [25] and Mo6⫹O3 standards are shown and in Figure 3-top the Sr2Cr1.2Mo0.8O6-␦; ␦ ⫽ 0 and 0.2 spectra are shown. Both the chemical shift and the structure of the Mo L3-edge can be used as Mo-valence indicators. The chemical shift of the spectral onset/centrum is seen to move to higher energy, with increasing Mo-valence, in the standard spectra. In view of varying 4d-orbital splitting contributions between materials (see discussion below) the chemical shift conclusions will be restricted to noting that the Sr2Cr1.2Mo0.8O6-␦ materials have a Mo-valence well above 4⫹. The intense peak features at the Mo L3-edge onset involve 2p-core-to-4d final-state transitions. These features can provide a probe of the 4d states, albeit modified by the transition matrix element, core-hole-interaction and multiplet effects [26]. Two strong spectral features are clearly visible in the MoO3 and the pyroclore-Sm2Mo2O7 standards: the
710
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 4. Observed, calculated and difference patterns from Rietveld refinement of Sr2Cr1.2Mo0.8O5.8. The difference plot is located at the bottom of the figure. Tic marks represent allowed reflections.
low energy peak (A), due to the 2p-to-4dt2g excitations; and the high-energy peak (B), due to 2p-to-4deg excitations. It should be noted that variations in the t2g-eg energy splitting and additional solid- state 4d-broadening effects contribute to these spectra. Nevertheless, the ratio of the to B spectral weight shows a dramatic decrease between the MoO3 (with a formal 6:4 ratio of t2g-to-eg available final states) and the pyroclore-Sm2Mo2O7 (with a formal 4:4 final state ratio) standards. Inspection of the A/B feature ratio of the Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0 and 0.2) spectra (in Figure 3-top) indicates a Mo-valence much greater than 4 and less than 6 for both the ␦ ⫽ 0 and 0.2 materials. Moreover, a clear decrease in the relative A-feature strength between the ␦ ⫽ 0 to 0.2 spectra strongly supports a direct Mo-valence reduction (on Table 1 Atomic parameters for Sr2Cr1.2Mo0.8O6 Atoms
Occupancy
x
y
z
Uiso(Å)2
Sr Cr(B⬘) Cr(B⬙) Mo(B⬘) Mo(B⬙) O
1.0 0.714(13) 0.486(13) 0.286(13) 0.514(13) 1.0
0.25 0 0.5 0 0.5 0.2578(1)
0.25 0 0.5 0 0.5 0
0.25 0 0.5 0 0.5 0
0.03045 0.01645 0.01927 0.01653 0.01927 0.02150
Note: Space group Fm3m: a ⫽ 7.7999(1) Å; V ⫽ 474.53(2) Å3; Rwp ⫽ 8.07%; Rp ⫽ 5.90%
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
711
Table 2 Atomic parameters for Sr2Cr1.2Mo0.8O5.8 at room temperature Atoms
Occupancy
x
y
z
Uiso (Å)2
Sr Cr(B⬘) Cr(B⬙) Mo(B⬘) Mo(B⬙) O
1.0 0.678(8) 0.523(8) 0.323(8) 0.477(8) 0.967
0.25 0 0.5 0 0.5 0.2571(2)
0.25 0 0.5 0 0.5 0
0.25 0 0.5 0 0.5 0
0.0400(7) 0.0114(12) 0.0174(11) 0.0114(12) 0.0174(11) 0.0326(11)
Note: Space group Fm3m: a ⫽ 7.8148(2) Å; V ⫽ 477.25(4) Å3; Rwp ⫽ 8.84%; Rp ⫽ 6.96%
the average) in response to the decreasing O-content. Thus it appears that the formal Mo-valence assignments of 5.5⫹ and 5⫹ for the ␦ ⫽ 0 and 0.2 materials are qualitatively consistent with the Mo XAS results. 3.3. Structure Sr2Cr1.2Mo0.8O6 crystallizes in a cubic lattice in space group Fm3m with a ⫽ 7.7999(1) Å in the latter. Table 1 shows the atomic parameters obtained from the Rietveld analysis. The refinement is relatively reliable with Rwp ⫽ 8.07% and Rp ⫽ 5.09%. Figure 4 shows the observed, calculated and difference room-temperature diffraction profiles for Sr2Cr1.2Mo0.8O5.8. The bond distance of Mo-O (1.898 Å) and Cr-O (2.009 Å) in Sr2Cr1.2Mo0.8O6 is close to the sums of the effective ionic radii of Mo5⫹ (0.75 Å), Cr3⫹ (0.755 Å), and O2- (1.26 Å) respectively [27]. As the ratio of Cr/Mo is 1.2:0.8, part of Cr has to enter the B’’ position after occupying all of the B’ sites. As the octahedral radius of Mo5⫹ (0.75 Å) is very close to that of Cr3⫹ (0.755 Å), Mo can also occupy both sites. However, the Rietveld analysis indicates that the B’ and B” ions are partially ordered. The Mo ratio of Table 3 Selected bond lengths (Å) and angles (°) for Sr2Cr1.2Mo0.8O6 and Sr2Cr1.2Mo0.8O5.8
SR-O Mo-O Cr-O Mo-O-Mo Cr-O-Cr Sr-O-Sr O-Cr-O O-Cr-O Cr-O-Mo O-Mo-O O-Mo-O O-Sr-O Sr-O-Cr Sr-O-Mo
Sr2Cr1.2Mo0.8O6
Sr2Cr1.2Mo0.8O5.8
2.758(2) 1.889(1) 2.011(1) 180 180 89.97(1) 179.980(0) 90.000(0) 180.000(0) 179.980(0) 90.000(0) 62.1(4) 88.74(23) 91.26(23)
2.763(1) 1.898(1) 2.009(1) 180 180 89.97(3) 180 90 180 180 90 61.874(2) 88.85(21) 91.15(33)
712
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 5. Log as a function of 1/T for Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6. Inset: log vs T
B’/B” is 0.286:0.514 from the Rietveld refinement. This suggests a ⬃28.5% Mo ordering [% ordering of Mo ⫽ (occupancy of Mo in B’ site - occupancy of Mo in the B” site)/total Mo content per unit] on the B site at room temperature. Sr2Cr1.2Mo0.8O5.8 has a slightly larger unit cell (a⫽7.8148 (2) Å) than Sr2Cr1.2Mo0.8O6 (a⫽ 7.7999(1) Å) consistent with the higher Mo5⫹ content of Sr2Cr1.2Mo0.8O5.8 [i.e., the ionic radius of Mo5⫹ (0.75 Å) is larger than that of Mo6⫹ (0.73 Å)]. The ordering of Cr3⫹ and Mo5/6⫹ is due to the charge difference of Cr and Mo, and this difference is greater in the ␦ ⫽ 0 compound, where the ordering is greater. 3.4. Electrical properties The temperature variations of resistivity of Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6 in Figure 5 indicate semiconducting behavior for both. Qualitative Seebeck measurements show that the majority carriers are electrons in these compounds. The log vs T plots are nearly linear (Figure 5 inset), while the log vs 1/T plots are not. This behavior is typical of magnetic semiconductors [12–17]. As noted previously and evidenced by XAS data, the oxidation state of Cr is formally 3⫹ in both compounds, while the molybdenum is mixed valent Mo5⫹/6⫹ in Sr2Cr1.2Mo0.8O6 and 5⫹ in Sr2Cr1.2Mo0.8O5.8. Thus the lower resistivity
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
713
Fig. 6. Magnetic susceptibility as a function of temperature for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2) at 100 G. Inset: Magnetization as a function of magnetic field for Sr2Cr1.2Mo0.8O6-␦ at 20 K.
of Sr2Cr1.2Mo0.8O6 is attributed to the presence of Mo5⫹/6⫹. In Sr2Cr1.2Mo0.8O5.8, the d1 electrons of Mo appear to be localized and the presence of oxygen vacancies also block electron transport; these factors lead to the three orders of magnitude higher resistivity in this compound compared to that of Sr2Cr1.2Mo0.8O6. 3.5. Magnetic properties Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6 order ferrimagnetically; the transition temperature, TN of both compounds is 465 K (Figure 6). The spins of Cr3⫹ (d3) and Mo5⫹ (d1) appear to order in an antiparallel arrangement by superexchange interaction, and lead to ferrimagnetic ordering below TN. Figure 6 (inset) shows hysteresis in the magnetization as a function of magnetic field for Sr2Cr1.2Mo0.8O6-␦. The coercivity (Hc) of Sr2Cr1.2Mo0.8O6 is 610 G at 20 K. 3.6. Magnetoresistance The magnetoresistance [defined as MR ⫽ (H-0)/0, where H and 0 are the resistivities at H and zero applied magnetic field] increases with decreasing temperature for both Sr2Cr1.2Mo0.8O6 and Sr2Cr1.2Mo0.8O5.8 (Figure 7). The largest MR observed is
714
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 7. MR as a function of temperature for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2). Inset: MR % as a function of field (H) for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2).
43% for Sr2Cr1.2Mo0.8O6 albeit at 20 K and 5T. In Figure 7 inset, the MR increases rapidly at low fields. This behavior is similar to that seen in Sr2FeMoO6. The large MR at relatively low fields is attributed to an intra-grain tunneling mechanism [28,29].
Acknowledgments We thank Prof. K.V. Ramanujachary for his suggestions with experimental problems and useful discussions and Prof. W.H. McCarroll for his critical reading of the manuscript. This work was supported by National Science Foundation-Solid State Chemistry Grant DMR96 –13106.
References [1] K. Chahara, T. Ohno, M. Kasai, Y. Kozono, Appl. Phys. Lett. 63 (1993) 1990; R.M. Kusters, J. Siug Letou, D.A. Keen, R. McGreeuy, W. Hayes, Phys Rev B. 155 (1993) 362; R. Von Helmholt, J. Wecker, B. Holzapfel, L. Schultz, K. Sanew Phys. Rev. Lett. 71 (1989) 2331. [2] S. Jin, T.H. Tiefel, M. McCormack, R.A. Fastnacht, R. Ramesh, L.H. Chen, Science 264 (1994) 413.
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
715
[3] A. Urushibara, Y. Moritomo, T. Arima, A. Asamitsu, G. Kido, Y.S.S. Manoharan, N.Y. Vasanthachaya, M.S. Hegde, K.M. Satyalakshmi, V. Prasad, S.V. Subramanyam, J. Appl. Phys. 76 (1994) 3923. [4] R. Mahendiran, R. Mahesh, A.K. Raychaudhuri, and C.N. Rao, J. Phys. D. 2 (8) (1995) 1743. [5] T. Shimura, T. Hayashi, Y. Inaguma, M. Itoh, J. Solids State Chem. 124 (1996) 250. [6] G. Briceno, H. Chang, X. Sun, P.G. Schultz, X.D. Xiang, Science 270 (1995) 273. [7] S. Jin, H.M. O’Bryan, T.H. Tiefel, M. McCormack, W.W. Rhodes, Appl. Phys. Lett. 66 (1995) 382. [8] J. Inou, S. Maekawa, Phys. Rev. 53B (1996) 52. [9] R. Mahendiran, A.K. Raychaudhuri, A. Chainani, D.D. Sarma, J. Phys. Condens. Matter. 7 (1995) L561. [10] A. P. Ramirez, R.J. Cava, J. Krajewski, Nature 386 (1997) 156. [11] T. Nakagawa, J. Phys. Soc. Japan 24 (1968) 806. [12] K.F. Patterson, C.W. Moeller, R. Ward, Inorg. Chem. 2 (1963) 196. [13] P. Woodward. R.D. Hoffmann, A.W. Sleight, J. Mater. Res. 9 (1994) 2118. [14] S. Nakayama, T. Nakagawa, S. Nomura, J. Phys. Soc. Japan 24 (1996) 219. [15] T. Nakawa, J. Phys. Soc. of Japan 24 (1968) 806. [16] F.S. Galasso, F.C. Douglas, R.J. Kasper, J. Chem. Phys. 44 (1996) 1672. [17] Y. Shimakuwa, Y. Kubo, T. Manako, Nature 379 (1996) 53. [18] K.-I. Kobayashi, T. Kimura, H. Sawada, K. Terakura, Y. Tokura, Nature 395 (1998) 677. [19] T. Manako, M. Izumi, Y. Konishi, K. Kobayashi, M. Kawasaki, Y. Tokura, Appl. Phys. Lett. 74 (1999) 2215. [20] Y. Tomioka, T. Okuda, Y. Okimoto, R. Kumai, K.-I. Kobayashi, Y. Tokura, Phys. Reb. B61 (2000) 422. [21] H.Q. Yin, J.-S. Zhou, J.-P. Zhou, R. Dass, J.T. Mcdevitt, J.B. Goodenough, Appl. Phys. Lett. 75 (1999) 2812. [22] C.L. Yuan, S.G. Wang, W.H. Song, T. Yu, J.M. Dai, S.L. Ye, Y.P. Sun, Appl. Phys. Lett. 75 (1999) 3853. [23] Y. Moritomo, Sh. Xu, A. Machida, T. Akimoto, E. Nishibori, M. Takata, M. Sakata, Phys. Rev. B 61 (2000) R7827. [24] A.C. Larson, R.B. von Dreele, General Structure Analysis System (GSAS), Los Alamos National Laboratories, Report LAUR, 1994, pp. 86 –748. [25] G. Veith, M. Greenblatt, M. Croft, to be published. [26] F.M.F. de Groot, Physica B 15 (1995) 208 –209. [27] R.D. Shannon, Acta Crystallogr. A32 (1976) 751. [28] H.W. Hwang, S.-W. Cheong, N.P. Ong, B. Batlogg, Phys. Rev. Lett. 77 (1996) 2041. [29] H.Y. Hwang, S.-W. Cheong, Colossal Magnetoresistance Oxides, Y. Tokura (Ed.), Gordon & Breach, Monograph in Condensed Matter Science, in press.
Materials Research Bulletin 36 (2001) 705–715
Large magnetoresistance in double perovskite Sr2Cr1.2Mo0.8O6-␦ Z. Zenga,1, I. D. Fawcetta, M. Greenblatta,*, M. Croftb a
Department of Chemistry, Rutgers State University of New Jersey, Piscataway, NJ 08854-8087, USA b Department of Physics, Rutgers State University of New Jersey, Piscataway, NJ 08854-8087, USA (Refereed) Received 12 July 2000; accepted 17 July 2000
Abstract Sr2Cr1.2Mo0.8O6-␦ (␦⫽0, 0.2) with a double perovskite structure was prepared by solid state reaction in evacuated quartz tubes. Cr and Mo are partially ordered on the B site. Oxygen vacancies decrease the ordering, but increase the lattice parameters. X-ray absorption spectroscopy results are consistent with Cr being 3⫹, and Mo being close to 5⫹ for ␦ ⫽ 0.2 and 5.5 for ␦ ⫽ 0. The spin of Cr3⫹ (d3) and Mo5⫹ (d1) order in an anti-parallel arrangement by superexchange interaction, and lead to ferrimagnetic ordering below 465 K. Both compounds are n-type narrow gap semiconductors. Large magnetoresistance (-43%) is observed in Sr2Cr1.2Mo0.8O6. The MR behavior is attributed to an intra-grain tunneling mechanism. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: A. Magnetic materials; B. Chemical synthesis; C. XANES; C. Electronic conductivity; D. Magnetic properties
1. Introduction Recent discovery of colossal magnetoresistance (CMR) in perovskite manganates has stimulated much interest [1–11]. CMR materials are potentially useful for applications in magneto-electronic devices. It is important to search for new magnetoresistance materials to improve the properties and understand the mechanism of CMR.
* Corresponding author. Fax: ⫹1-732-445-5312. E-mail address: [email protected] (M. Greenblatt). 1 Present address: Energy Technology Division, Argonne National Laboratory, Argonne, IL. 0025-5408/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 5 - 5 4 0 8 ( 0 1 ) 0 0 5 2 0 - 7
706
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 1. Magnetic structure of a ferrimagnetically ordered double perovskite.
Perovskites have the general formula ABX3, in which A represents a large electropositive cation, B is a small transition metal or main group ion, and X is commonly an oxide or halide ion. Double perovskites with the formula A’A”B’B”O6, with metal ions B’ and B”, may order on the perovskite B lattice site if the size/charge of the B’ and B” ions are different. The ordered double perovskites with paramagnetic transition metal, B’ and B” ions can exhibit magnetic ordering as well. Some double perovskite materials have high (⬎room temperature) magnetic ordering temperature [12–17]. Fig. 1 illustrates the magnetic structure of an aniferromagnetically (AF)-ordered double perovskite. In such materials, if the number of d electrons of the B’ and B” ions are different, the magnetic coupling leads to ferrimagnetic order, as in Fig. 1. Recently, Kobayashi et al. reported large low-field MR in the double perovskite Sr2FeMoO6 [18,19]; the MR was attributed to inter-grain tunneling mechanism. Subsequently, this mechanism was confirmed by the absence of MR in single crystals of Sr2FeMoO6 [20 –22]. Patterson et al reported on Sr2CrMoO6, which was contaminated by strontium molybdate [12]. More recently Moritomo et al investigated the electronic structure of Sr2MMoO6 (M ⫽ Cr, Mn, Fe, and Co), but the M ⫽ Cr phase was not well characterized [23]. In this paper we report the synthesis of single-phase
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
707
Sr2Cr1⫹xMoxO6-␦ (x ⫽ 0.1, 0.2; ␦ ⫽ 0, 0.2) and the properties of Sr2Cr1.2Mo0.8O6-␦ with relatively large MR.
2. Experimental Sr2Cr1⫹xMoxO6-␦ (x ⫽ 0.1, 0.2; ␦ ⫽ 0, 0.2) were prepared by solid-state reaction. All starting materials were purchased from Aldrich. Stoichiometric SrO (obtained by decomposition of SrCO3 at 1300 °C), Cr2O3, MO3 and Mo were ground in a dry box. The mixtures were pressed into pellets and sealed into evacuated quartz tubes. The pellets were sintered at 550 °C for 10 h, and then the temperature was increased to 1000 °C and the sample sintered for an additional 24 h. The resulting products were reground and pressed into pellets again and sintered at 1000 °C for 24 h in evacuated quartz tubes The powder X-ray diffraction (PXD) data were collected with a SCINTAG PAD V diffractometer with Ni-filtered CuK␣ radiation over the range 10° ⱕ 2 ⱕ 90° with a step size of 0.02 °. Silicon powder was used as an internal standard. Lattice parameters were refined by a least-square method. Rietveld refinement of the data was undertaken with the GSAS Rietveld refinement program [24]. A broad maximum seen in the background of the X-ray diffraction patterns is due to the glass slide and Vaseline used to hold the sample. The backgrounds of the PXD data in the Rietveld refinements were fitted with an 8-term Chebyshev polynomial and a pseudo-Voigt function was employed to model the peak shapes. The dc electrical resistivity and the magnetoresistance measurements were carried out by a standard four-probe technique from 400 to 4 K in a SQUID magnetometer (MPMS, Quantum Design). The magnetic properties were also measured with the SQUID magnetometer in the temperature range 4 – 400 K. The Mo L3-edge and Cr K-edge X-ray absorption spectroscopy (XAS) measurements were performed on beam line X-19A at the Brookhaven national synchrotron light source with a double crystal [InSb (111) or Si (111)] monochromator. The measurements were made in the fluorescence mode. A standard was run simultaneously with all the measurements for precise calibration of the Cr-spectra to a relative accuracy of about ⫾0.03 eV. Standard samples were run frequently in the Mo measurements for calibration with a lower accuracy of about ⫾0.8 eV. All spectra were normalized to unity step in the absorption coefficient from well below to well above the edge.
3. Results and discussions 3.1. Synthesis Attempts to prepare Sr2CrMoO6 double perovskite in an evacuated quartz tube resulted in a black powder contaminated with SrMoO4. Single-phase samples of Sr2Cr1⫹xMo1-xO6-␦, could be obtained only with Cr/Mo ⬎ 1.0 in a narrow range of composition (0.1 ⱕ x ⱕ 0.2), and the raw materials must be dried thoroughly. The quartz tubes should be flamed three
708
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 2. Cr K-edge X-ray absorption specta of Sr2Cr1.2Mo0.8O6, CrO2, LaCrO3, Cr2O3, and Cr-metal. In the Inset the pre-edge features of the oxide spectra are compared (with the Sr2Cr1.2Mo0.8O6 spectrum displaced downward for clarity).
times before sealing. The x ⫽ 0.1 material has orders of magnitude higher resistivity than the x ⫽ 0.2 compounds. Therefore the latter compounds were studied in detail 3.2. Oxidation state of Cr and Mo Figure 2 shows the Cr K-edge XAS spectra of Sr2Cr1.2Mo0.8O6, the Cr3⫹ standards LaCrO3 and Cr2O3, and the Cr4⫹ standard CrO2. The Cr-K edge spectra of both Sr2Cr1.2Mo0.8O6-␦ ␦ ⫽ 0 and 0.2 compounds were essentially identical so only one has been displayed. The rising part of the Sr2Cr1.2Mo0.8O6 main edge can be seen to have a chemical shift identical to that of the Cr3⫹ standards (see the box in the Figure) and well separated from that of the Cr4⫹ standard. The pre-edge features of the Cr-K edge spectra are sown in the inset of Figure 2. Specifically the Sr2Cr1.2Mo0.8O6-␦ pre-edge manifests a three-feature (a1, a2, and a3) structure, similar to the Cr3⫹ standards, whereas the Cr4⫹ standard has a quite distinct bimodal-feature structure. Among the Cr3⫹ materials the sharp-low-energy a1-feature is particularly distinctive with the intensities and positions of the a2 and a3
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
709
Fig. 3. (bottom) The Mo L-edge X-ray absorption spectra of Mo-metal, MoO2, Sm2Mo2O7 (from ref. 24), and MoO3 standard materials. (top) The Mo L-edge spectra of Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2) along with the MoO2 standard.
features varying between compounds. Thus the structure of the pre-edge further supports the assignment of a Cr3⫹ state for these Sr2Cr1.2Mo0.8O6-␦ materials. From the chemical formula, if the valence of Cr is 3⫹, Mo should be formally 5.5⫹ in Sr2Cr1.2Mo0.8O6, and 5⫹ in Sr2Cr1.2Mo0.8O5.8. In Figure 3-bottom, the Mo L3-edges of the Mo0, Mo4⫹O2, Sm2Mo4⫹2O7 [25] and Mo6⫹O3 standards are shown and in Figure 3-top the Sr2Cr1.2Mo0.8O6-␦; ␦ ⫽ 0 and 0.2 spectra are shown. Both the chemical shift and the structure of the Mo L3-edge can be used as Mo-valence indicators. The chemical shift of the spectral onset/centrum is seen to move to higher energy, with increasing Mo-valence, in the standard spectra. In view of varying 4d-orbital splitting contributions between materials (see discussion below) the chemical shift conclusions will be restricted to noting that the Sr2Cr1.2Mo0.8O6-␦ materials have a Mo-valence well above 4⫹. The intense peak features at the Mo L3-edge onset involve 2p-core-to-4d final-state transitions. These features can provide a probe of the 4d states, albeit modified by the transition matrix element, core-hole-interaction and multiplet effects [26]. Two strong spectral features are clearly visible in the MoO3 and the pyroclore-Sm2Mo2O7 standards: the
710
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 4. Observed, calculated and difference patterns from Rietveld refinement of Sr2Cr1.2Mo0.8O5.8. The difference plot is located at the bottom of the figure. Tic marks represent allowed reflections.
low energy peak (A), due to the 2p-to-4dt2g excitations; and the high-energy peak (B), due to 2p-to-4deg excitations. It should be noted that variations in the t2g-eg energy splitting and additional solid- state 4d-broadening effects contribute to these spectra. Nevertheless, the ratio of the to B spectral weight shows a dramatic decrease between the MoO3 (with a formal 6:4 ratio of t2g-to-eg available final states) and the pyroclore-Sm2Mo2O7 (with a formal 4:4 final state ratio) standards. Inspection of the A/B feature ratio of the Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0 and 0.2) spectra (in Figure 3-top) indicates a Mo-valence much greater than 4 and less than 6 for both the ␦ ⫽ 0 and 0.2 materials. Moreover, a clear decrease in the relative A-feature strength between the ␦ ⫽ 0 to 0.2 spectra strongly supports a direct Mo-valence reduction (on Table 1 Atomic parameters for Sr2Cr1.2Mo0.8O6 Atoms
Occupancy
x
y
z
Uiso(Å)2
Sr Cr(B⬘) Cr(B⬙) Mo(B⬘) Mo(B⬙) O
1.0 0.714(13) 0.486(13) 0.286(13) 0.514(13) 1.0
0.25 0 0.5 0 0.5 0.2578(1)
0.25 0 0.5 0 0.5 0
0.25 0 0.5 0 0.5 0
0.03045 0.01645 0.01927 0.01653 0.01927 0.02150
Note: Space group Fm3m: a ⫽ 7.7999(1) Å; V ⫽ 474.53(2) Å3; Rwp ⫽ 8.07%; Rp ⫽ 5.90%
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
711
Table 2 Atomic parameters for Sr2Cr1.2Mo0.8O5.8 at room temperature Atoms
Occupancy
x
y
z
Uiso (Å)2
Sr Cr(B⬘) Cr(B⬙) Mo(B⬘) Mo(B⬙) O
1.0 0.678(8) 0.523(8) 0.323(8) 0.477(8) 0.967
0.25 0 0.5 0 0.5 0.2571(2)
0.25 0 0.5 0 0.5 0
0.25 0 0.5 0 0.5 0
0.0400(7) 0.0114(12) 0.0174(11) 0.0114(12) 0.0174(11) 0.0326(11)
Note: Space group Fm3m: a ⫽ 7.8148(2) Å; V ⫽ 477.25(4) Å3; Rwp ⫽ 8.84%; Rp ⫽ 6.96%
the average) in response to the decreasing O-content. Thus it appears that the formal Mo-valence assignments of 5.5⫹ and 5⫹ for the ␦ ⫽ 0 and 0.2 materials are qualitatively consistent with the Mo XAS results. 3.3. Structure Sr2Cr1.2Mo0.8O6 crystallizes in a cubic lattice in space group Fm3m with a ⫽ 7.7999(1) Å in the latter. Table 1 shows the atomic parameters obtained from the Rietveld analysis. The refinement is relatively reliable with Rwp ⫽ 8.07% and Rp ⫽ 5.09%. Figure 4 shows the observed, calculated and difference room-temperature diffraction profiles for Sr2Cr1.2Mo0.8O5.8. The bond distance of Mo-O (1.898 Å) and Cr-O (2.009 Å) in Sr2Cr1.2Mo0.8O6 is close to the sums of the effective ionic radii of Mo5⫹ (0.75 Å), Cr3⫹ (0.755 Å), and O2- (1.26 Å) respectively [27]. As the ratio of Cr/Mo is 1.2:0.8, part of Cr has to enter the B’’ position after occupying all of the B’ sites. As the octahedral radius of Mo5⫹ (0.75 Å) is very close to that of Cr3⫹ (0.755 Å), Mo can also occupy both sites. However, the Rietveld analysis indicates that the B’ and B” ions are partially ordered. The Mo ratio of Table 3 Selected bond lengths (Å) and angles (°) for Sr2Cr1.2Mo0.8O6 and Sr2Cr1.2Mo0.8O5.8
SR-O Mo-O Cr-O Mo-O-Mo Cr-O-Cr Sr-O-Sr O-Cr-O O-Cr-O Cr-O-Mo O-Mo-O O-Mo-O O-Sr-O Sr-O-Cr Sr-O-Mo
Sr2Cr1.2Mo0.8O6
Sr2Cr1.2Mo0.8O5.8
2.758(2) 1.889(1) 2.011(1) 180 180 89.97(1) 179.980(0) 90.000(0) 180.000(0) 179.980(0) 90.000(0) 62.1(4) 88.74(23) 91.26(23)
2.763(1) 1.898(1) 2.009(1) 180 180 89.97(3) 180 90 180 180 90 61.874(2) 88.85(21) 91.15(33)
712
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 5. Log as a function of 1/T for Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6. Inset: log vs T
B’/B” is 0.286:0.514 from the Rietveld refinement. This suggests a ⬃28.5% Mo ordering [% ordering of Mo ⫽ (occupancy of Mo in B’ site - occupancy of Mo in the B” site)/total Mo content per unit] on the B site at room temperature. Sr2Cr1.2Mo0.8O5.8 has a slightly larger unit cell (a⫽7.8148 (2) Å) than Sr2Cr1.2Mo0.8O6 (a⫽ 7.7999(1) Å) consistent with the higher Mo5⫹ content of Sr2Cr1.2Mo0.8O5.8 [i.e., the ionic radius of Mo5⫹ (0.75 Å) is larger than that of Mo6⫹ (0.73 Å)]. The ordering of Cr3⫹ and Mo5/6⫹ is due to the charge difference of Cr and Mo, and this difference is greater in the ␦ ⫽ 0 compound, where the ordering is greater. 3.4. Electrical properties The temperature variations of resistivity of Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6 in Figure 5 indicate semiconducting behavior for both. Qualitative Seebeck measurements show that the majority carriers are electrons in these compounds. The log vs T plots are nearly linear (Figure 5 inset), while the log vs 1/T plots are not. This behavior is typical of magnetic semiconductors [12–17]. As noted previously and evidenced by XAS data, the oxidation state of Cr is formally 3⫹ in both compounds, while the molybdenum is mixed valent Mo5⫹/6⫹ in Sr2Cr1.2Mo0.8O6 and 5⫹ in Sr2Cr1.2Mo0.8O5.8. Thus the lower resistivity
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
713
Fig. 6. Magnetic susceptibility as a function of temperature for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2) at 100 G. Inset: Magnetization as a function of magnetic field for Sr2Cr1.2Mo0.8O6-␦ at 20 K.
of Sr2Cr1.2Mo0.8O6 is attributed to the presence of Mo5⫹/6⫹. In Sr2Cr1.2Mo0.8O5.8, the d1 electrons of Mo appear to be localized and the presence of oxygen vacancies also block electron transport; these factors lead to the three orders of magnitude higher resistivity in this compound compared to that of Sr2Cr1.2Mo0.8O6. 3.5. Magnetic properties Sr2Cr1.2Mo0.8O5.8 and Sr2Cr1.2Mo0.8O6 order ferrimagnetically; the transition temperature, TN of both compounds is 465 K (Figure 6). The spins of Cr3⫹ (d3) and Mo5⫹ (d1) appear to order in an antiparallel arrangement by superexchange interaction, and lead to ferrimagnetic ordering below TN. Figure 6 (inset) shows hysteresis in the magnetization as a function of magnetic field for Sr2Cr1.2Mo0.8O6-␦. The coercivity (Hc) of Sr2Cr1.2Mo0.8O6 is 610 G at 20 K. 3.6. Magnetoresistance The magnetoresistance [defined as MR ⫽ (H-0)/0, where H and 0 are the resistivities at H and zero applied magnetic field] increases with decreasing temperature for both Sr2Cr1.2Mo0.8O6 and Sr2Cr1.2Mo0.8O5.8 (Figure 7). The largest MR observed is
714
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
Fig. 7. MR as a function of temperature for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2). Inset: MR % as a function of field (H) for Sr2Cr1.2Mo0.8O6-␦ (␦ ⫽ 0, 0.2).
43% for Sr2Cr1.2Mo0.8O6 albeit at 20 K and 5T. In Figure 7 inset, the MR increases rapidly at low fields. This behavior is similar to that seen in Sr2FeMoO6. The large MR at relatively low fields is attributed to an intra-grain tunneling mechanism [28,29].
Acknowledgments We thank Prof. K.V. Ramanujachary for his suggestions with experimental problems and useful discussions and Prof. W.H. McCarroll for his critical reading of the manuscript. This work was supported by National Science Foundation-Solid State Chemistry Grant DMR96 –13106.
References [1] K. Chahara, T. Ohno, M. Kasai, Y. Kozono, Appl. Phys. Lett. 63 (1993) 1990; R.M. Kusters, J. Siug Letou, D.A. Keen, R. McGreeuy, W. Hayes, Phys Rev B. 155 (1993) 362; R. Von Helmholt, J. Wecker, B. Holzapfel, L. Schultz, K. Sanew Phys. Rev. Lett. 71 (1989) 2331. [2] S. Jin, T.H. Tiefel, M. McCormack, R.A. Fastnacht, R. Ramesh, L.H. Chen, Science 264 (1994) 413.
Z. Zeng et al. / Materials Research Bulletin 36 (2001) 705–715
715
[3] A. Urushibara, Y. Moritomo, T. Arima, A. Asamitsu, G. Kido, Y.S.S. Manoharan, N.Y. Vasanthachaya, M.S. Hegde, K.M. Satyalakshmi, V. Prasad, S.V. Subramanyam, J. Appl. Phys. 76 (1994) 3923. [4] R. Mahendiran, R. Mahesh, A.K. Raychaudhuri, and C.N. Rao, J. Phys. D. 2 (8) (1995) 1743. [5] T. Shimura, T. Hayashi, Y. Inaguma, M. Itoh, J. Solids State Chem. 124 (1996) 250. [6] G. Briceno, H. Chang, X. Sun, P.G. Schultz, X.D. Xiang, Science 270 (1995) 273. [7] S. Jin, H.M. O’Bryan, T.H. Tiefel, M. McCormack, W.W. Rhodes, Appl. Phys. Lett. 66 (1995) 382. [8] J. Inou, S. Maekawa, Phys. Rev. 53B (1996) 52. [9] R. Mahendiran, A.K. Raychaudhuri, A. Chainani, D.D. Sarma, J. Phys. Condens. Matter. 7 (1995) L561. [10] A. P. Ramirez, R.J. Cava, J. Krajewski, Nature 386 (1997) 156. [11] T. Nakagawa, J. Phys. Soc. Japan 24 (1968) 806. [12] K.F. Patterson, C.W. Moeller, R. Ward, Inorg. Chem. 2 (1963) 196. [13] P. Woodward. R.D. Hoffmann, A.W. Sleight, J. Mater. Res. 9 (1994) 2118. [14] S. Nakayama, T. Nakagawa, S. Nomura, J. Phys. Soc. Japan 24 (1996) 219. [15] T. Nakawa, J. Phys. Soc. of Japan 24 (1968) 806. [16] F.S. Galasso, F.C. Douglas, R.J. Kasper, J. Chem. Phys. 44 (1996) 1672. [17] Y. Shimakuwa, Y. Kubo, T. Manako, Nature 379 (1996) 53. [18] K.-I. Kobayashi, T. Kimura, H. Sawada, K. Terakura, Y. Tokura, Nature 395 (1998) 677. [19] T. Manako, M. Izumi, Y. Konishi, K. Kobayashi, M. Kawasaki, Y. Tokura, Appl. Phys. Lett. 74 (1999) 2215. [20] Y. Tomioka, T. Okuda, Y. Okimoto, R. Kumai, K.-I. Kobayashi, Y. Tokura, Phys. Reb. B61 (2000) 422. [21] H.Q. Yin, J.-S. Zhou, J.-P. Zhou, R. Dass, J.T. Mcdevitt, J.B. Goodenough, Appl. Phys. Lett. 75 (1999) 2812. [22] C.L. Yuan, S.G. Wang, W.H. Song, T. Yu, J.M. Dai, S.L. Ye, Y.P. Sun, Appl. Phys. Lett. 75 (1999) 3853. [23] Y. Moritomo, Sh. Xu, A. Machida, T. Akimoto, E. Nishibori, M. Takata, M. Sakata, Phys. Rev. B 61 (2000) R7827. [24] A.C. Larson, R.B. von Dreele, General Structure Analysis System (GSAS), Los Alamos National Laboratories, Report LAUR, 1994, pp. 86 –748. [25] G. Veith, M. Greenblatt, M. Croft, to be published. [26] F.M.F. de Groot, Physica B 15 (1995) 208 –209. [27] R.D. Shannon, Acta Crystallogr. A32 (1976) 751. [28] H.W. Hwang, S.-W. Cheong, N.P. Ong, B. Batlogg, Phys. Rev. Lett. 77 (1996) 2041. [29] H.Y. Hwang, S.-W. Cheong, Colossal Magnetoresistance Oxides, Y. Tokura (Ed.), Gordon & Breach, Monograph in Condensed Matter Science, in press.