Chemistry of Materials
Accepted for publication
The valence state of iron in the Sr2Fe(Mo,W,Ta)O6.0 double-perovskite system: An Fe K-edge and L2,3-edge XANES study
M. Karppinen,1* H. Yamauchi,1 Y. Yasukawa,1 J. Lindén,2 T. S. Chan,3 R. S. Liu,3 and J. M. Chen4
Materials and Structures Laboratory, Tokyo Institute of Technology, Yokohama 226-8503, Japan
Physics Department, ?bo Akademi, FIN-20500 Turku, Finland
Department of Chemistry, National Taiwan University, Taipei, Taiwan, Republic of China
National Synchrotron Radiation Research Center, Hsinchu, Taiwan, Republic of China
Corresponding author (e-mail: firstname.lastname@example.org)
Here we employ both Fe K and L2,3 edge x-ray absorption near-edge structure (XANES) spectroscopy techniques to clarify that iron in the B-site ordered double-perovskite halfmetal, Sr2FeMoO6.0, possesses a mixed-valence state, FeII/III, and accordingly molybdenum a mixed MoV/VI valence state. A reliable interpretation of the spectral features has been made possible by using a series of samples of the Sr2Fe(Mo,W/Ta)O6.0 system. Replacing MoV/VI gradually with WVI causes increasing amount of Fe to adopt the FeII state, whereas TaV substitution shifts the valence of iron towards FeIII. As the valence of Fe increases from II to III in the Sr2Fe(Mo,W/Ta)O6.0 system, the absorption energy at the Fe K-edge gradually shifts towards the higher energy side. Similarly, in the L2,3-edge XANES spectra intermediate spectral features are revealed for the Sr2FeMoO6.0 sample in comparison with those for samples heavily substituted with either W or Ta. Keywords: Fe valence, XANES spectroscopy, Double perovskite, Halfmetal
Introduction The B-site ordered double perovskites, A2BB’O6-w, are derived from the simple perovskite compounds, ABO3, upon co-occupation of the octahedral cation site with two different metal species of different charges. Such category of compounds has been re-highlighted since roomtemperature halfmetallicity and tunneling-type magnetoresistance behavior were revealed for Sr2FeMoO6-w,1 Sr2FeReO6-w,2 and related A2Fe(Mo,Re)O6-w (A = Ca, Sr, Ba) compounds. Long before the present rush in research of these compounds they were known as conductive roomtemperature ferrimagnets.3 In an ideal double perovskite the two B-site cations, B and B’, are perfectly ordered such that each BO6 octahedron is surrounded by six corner-sharing B’O6 octahedra, and vice versa. For the different pairs of B and B’, the larger the charge difference is, the higher is the degree of order at the B site.4 Actual valence states of the B-site cations in A2Fe(Mo,Re)O6-w were discussed for the first time in 1970’s.5 Renewed debates on this topic have been actively going on for a couple of years.6,7 The Sr2FeMoO6-w phase forms only under strongly reduced oxygen partial pressures, underlining the fact that either one (or both) of the B-site cation constituents possesses a relatively low valence value, i.e., lower than III for Fe and/or VI for Mo.8 Initially a picture was assumed based on high-spin FeIII with localized 3d5 electrons (t2g3eg2; S = 5/2) and MoV with an itinerant 4d1 electron (t2g1; S = 1/2). However, we recently proposed that iron is in a mixed-valence or valence-fluctuating state expressed as FeII/III, based on the
spectra revealing intermediate hyperfine parameters for Fe in Sr2FeMoO6.0 as compared with those typically obtained for high-spin FeIII and high-spin FeII.9 Note that molybdenum should be at a mixed MoV/VI valence state accordingly, as the itinerant d electron of Mo transfers part of its charge and spin density to Fe. The mixed valence interpretation has been accepted in other
Fe M?ssbauer studies,10,11 though divergent interpretations have been reported as
well.12 The reported L2,3-edge x-ray absorption near-edge structure (XANES) spectra for Sr2FeMoO6-w have been ascribed to both FeIII suggested a mixed-valence state, MoV/VI.16 Here, using a series of samples of the Sr2Fe(Mo,W/Ta)O6-w system we are able to show that the Fe XANES spectra not only at the L2,3 edge but also at the K edge may be interpreted on the basis of actual II/III mixed-valence state for iron in Sr2FeMoO6.0. The end-members of the sample series, Sr2FeWO617 and Sr2FeTaO618, provide us with optimal references for iron of pure divalent and pure trivalent state, respectively, in an oxide environment as akin as possible and FeII/III
. For Mo, an NMR study
to that in Sr2FeMoO6.0. Replacing MoV/VI with increasing amount of WVI thus makes the valence of Fe gradually approach the II state, whereas that with increasing amount of TaV causes continuous shift of the Fe valence towards III. The very same samples used in the present study previously had undergone thorough characterizations by means of
M?ssbauer spectroscopy for the hyperfine parameters of different Fe species present in the samples and transport measurements under different magnetic fields for the magnetoresistance characteristics, the results of which were reported elsewhere.19,20 For the samples heavily substituted with either W or Ta enhanced low-temperature MR values were achieved. For this reason the Sr2Fe(Mo,W)O6-w system was already highlighted in earlier studies.21,22
Experimental Section Sample Preparation. The Sr2Fe(Mo,W/Ta)O6-w samples used in this study were synthesized by means of an oxygen-getter-controlled low-O2-pressure encapsulation technique.8 As starting materials, stoichiometric mixtures of high-purity SrCO3, Fe2O3, MoO3 and WO3/Ta2O5 powders were used. Calcinations were carried out for the thoroughly mixed powders in an Ar atmosphere at 900 oC for 15 hours. The calcined powders were pelletized and sintered in evacuated fused-quartz ampoules containing sample pellets together with Fe grains (99.9 % up, under 10 mesh) as a getter of excess oxygen. The empty space inside the ampoule was filled with a fused-quartz rod. The synthesis was carried out at 1150 oC for 50 hours. At 1150 oC the oxygen partial pressure equilibrates in the presence of the Fe/FeO redox couple at ?2.6 × 10-13 atm.23 Characterization. The synthesized samples were checked for phase-purity and lattice parameters by x-ray diffraction (XRD; MAC Science M18XHF22; CuKα radiation). For the oxygen-content determination, an analysis method8 based on coulometric titration of FeII and/or MoV species formed upon acidic dissolution of the sample was applied. The samples were also characterized for the dc magnetization (from 5 K to 400 K, and –5 T to 5 T) using a superconductivity-quantum-interface-device magnetometer (SQUID; Quantum Design: MPMSR-5S). From the measured magnetization data the magnitude of saturation magnetization (MS) was determined as the magnetization value per formula unit at 5 T and 5 K. X-ray Absorption Measurements. The Fe K- and L2,3-edge XANES measurements were
performed at National Synchrotron Radiation Research Center (NSRRC) in Hsinchu, Taiwan with an electron beam energy of 1.5 GeV and a maximum stored current of 240 mA.24 For the XANES measurement the sample was ground to pass through a 400-mesh sieve to fulfil the requirement that the size of the particles is smaller than the absorption length in the material, i.e. ?d < 1, where d is the particle size and ? is the total absorption coefficient. The resultant fine powder was rubbed homogeneously onto Scotch tape. Then the thickness of the sample was carefully adjusted by folding the Scotch tape several times to achieve ??x ≈ 1, where ??x is the edge step. The measurements at the Fe K edge were performed in transmission mode at the Wiggler beamline BL-17C with a double-crystal Si (111) monochromator. Gas-ionization chambers filled with gas mixtures of N2-He and N2-Ar were used as detectors to measure (with a scan step of ~0.4 eV in the XANES region), respectively, the incident (I0) and transmitted (I) photon intensities. As a reference for energy calibration, the spectrum of Fe metal foil was simultaneously monitored. The higher x-ray harmonics were minimized by detuning the double-crystal Si(100) monochromator to 80 % of the maximum. The Fe L2,3-edge spectra were recorded by measuring the sample drain current in an ultrahigh vacuum chamber (10-9 torr) at the 6-m high-energy spherical grating monochromator (HSGM) beamline. The incident photon flux (I0) was monitored simultaneously by using a Ni mesh located after the exit slit of the monochromatic beam. The absorption measurements were normalized to I0. The reproducibility of the adsorption spectra of the same sample in different experimental runs was found to be extremely good. All the measurements were performed at room temperature.
Results and Discussion The Sr2Fe(Mo,W/Ta)O6-w samples were found to be phase pure for the whole Ta-for-Mo substitution range and for W-for-Mo substitution up to ?70 % substitution level. For the higher W contents traces of the non-magnetic Scheelite-type Sr(Mo,W)O4 phase were detected in the x-ray diffraction patterns. On the other hand, for the Ta-substituted samples the superlattice peaks due to doubling of the lattice parameter of the simple perovskite cell were clearly distinguished up to ?80 % substitution level only, indicating that Sr2FeTaO6-w rather possesses a simple-perovskite structure with random occupation of the B-lattice site by Fe and Ta atoms. For the non-substituted Sr2FeMoO6-w sample the measured saturation magnetization, MS, was as high as 3.5 ?B reflecting the high degree of order among the Fe
and Mo atoms. (The reduction in MS as compared with the maximum value of 4?B is commonly attributed to the decreased degree of order at the B-cation site.) Parallel with the high MS value, our
Fe M?ssbauer spectroscopy investigation carried out for the same
Sr2FeMoO6-w sample had revealed that the fraction of Fe atoms occupying the “wrong” Mo site is as low as 3 %.20 Here it should also be noted that M?ssbauer data have clearly shown that the misplaced or so-called antisite Fe atoms in the Sr2Fe(Mo,W/Ta)O6-w system are trivalent.9,11,20 Whereas the Ta substitution was found to decrease the degree of B-site order, substituting Mo with W enhanced the ordering. This trend is rather what one expects, since for the B-cation pair of FeII and WVI the charge difference is 4 while it is only 2 for FeIII and TaV. For the both sample series, Sr2Fe(Mo,W)O6-w and Sr2Fe(Mo,Ta)O6-w the unit volume of the lattice as determined from the XRD patterns showed monotonous evolution as the W/Ta substitution proceeded. The volume versus substitution level plots were given in Ref. 20 and are therefore not repeated here. The small deviations seen in these plots from the completely linear behavior are well explained by the changes in the degree of B-site cation order. (For Bsite ordered double perovskites the lattice dimension(s) are typically found to expand with decreasing degree of order.4) For all the W-substituted samples and also for the Ta-substituted samples up to 40 % substitution level, the oxygen content was precisely determined by means of wet-chemical analysis. For the heavily Ta-substituted samples the analysis was not possible due to poor solubility of the sample material. For all the samples investigated, the analysis yielded a stoichiometric value of 6.00 for the oxygen content per formula unit within the error limits of ±0.03. Consistently with this result our previous M?ssbauer data had not revealed any indication of five-fold coordinated Fe species for any other sample except for that substituted with 100 % Ta.20 For Sr2FeTaO6-w, 4 % of Fe atoms were found to possess the (fivecoordinated) divalent state,20 thus suggesting oxygen deficiency of w ≈ 0.02 (that is not significant either). In Figure 1 we show the L3 portion of the Fe L2,3-edge XANES spectra for a series of Sr2Fe(Mo,W/Ta)O6.0 samples. The main spectral features of the L2,3 edge of Fe originate from dipole transitions from the core Fe 2p level to the empty Fe 3d states.25,26 The spectra are separated into two regions due to core-hole spin-orbit interaction: Fe 2p3/2 (L3 edge; 705 ? 715 eV) and Fe 2p1/2 (L2 edge; 715 ? 730 eV). Transitions 2p → 4s are also allowed but much weaker contributing only to the smooth background at higher energies. Both the edges, L3 and L2, are further divided into two peaks. The splitting and intensity ratio between the two peaks
is determined by the interplay of crystal-field effects and electronic interactions. The L3 absorption edge of FeII species in an octahedral crystal field typically exhibits a main peak at a lower energy (?707 eV), followed by a weaker peak or a shoulder at a higher energy (?709 eV).25,26 The order of the peaks is reversed for FeIII species. This is what is precisely seen for the Sr2Fe(Mo,W/Ta)O6.0 sample series: the lower-energy peak is stronger than the higherenergy peak for heavily WVI-substituted samples and weaker for the heavily TaV-substituted ones. The Sr2FeMoO6.0 sample possesses intermediate spectral features as compared with those for the strongly W- and Ta-substituted samples, as a manifestation of the FeII/III mixedvalence state in it. For a more quantitative illustration, we approximate the intensities, I707 and I709, of the two L3-edge peaks by the peak heights and plot the intensity ratio, I709/I707, against the W/Ta substitution level in Figure 1(b). In Figure 2(a), the Fe K-edge absorption spectra are shown for selected Sr2Fe(Mo,W/Ta)O6.0 samples together with those for three simple iron oxides as references: FeIIO (wüstite of the rock-salt structure with an octahedral site for FeII), FeII/III3O4 (magnetite of the inverse-spinel structure with an octahedral site for FeII and tetrahedral and octahedral sites for FeIII) and α-FeIII2O3 (hematite of the corundum structure with an octahedral site for FeIII). For the reference oxides, it is concluded that the main peak that is due to transitions 1s → 4p clearly shifts to the higher energy with an increasing valence state of iron. A small prepeak is seen at the low-energy side of the main peak for all the three oxides. This is probably due to transitions 1s → 3d, though the exact origin of the fine structure seen at the Fe K edge is still being debated.26 The spectra of the Sr2Fe(W0.9Mo0.1)O6.0, Sr2Fe(W0.2Mo0.8)O6.0, Sr2Fe(Mo0.8Ta0.2)O6.0 and Sr2FeTaO6.0 samples also shown in Figure 2(a) are located roughly between those for FeIIO and α-FeIII2O3, obeying the trend seen for the reference oxides, i.e., with increasing valence state of iron the main absorption edge shifts to the higher energy. The details of the spectral features of the Sr2Fe(Mo,W/Ta)O6.0 samples somewhat differ from those of the simple iron oxides, whereas within the Sr2Fe(Mo,W/Ta)O6.0 double perovskite series, the spectral features evolve smoothly. This underlines the fact that the sample series itself provides us with the best reference system. The very edge area of the spectra from 7115 to 7130 eV is shown in Figure 2(b) for the whole series of Sr2Fe(Mo,W/Ta)O6.0 samples. Clearly, the absorption edge monotonously shifts to the higher energy first with decreasing amount of WVI that replaces MoV/VI in Sr2Fe(Mo,W)O6.0 and then with increasing amount of TaV to replace MoV/VI in Sr2Fe(Mo,Ta)O6.0, i.e., with the expected valence of Fe increasing from II to III. Thus, for the Mo-rich samples, absorption energy values that are intermediate between
those for strongly WVI- or TaV-substituted samples are seen, which manifests the intermediate valence state of iron in these samples. From Figure 2(b), even though the overall trend is clear within the whole Sr2Fe(Mo,W/Ta)O6.0 system, the shift of the K-edge absorption energy is larger for the WVI-substituted samples as compared with those substituted with TaV. The small oxygen-deficiency of w ≈ 0.02 concluded for the 100 % Ta-substituted sample from the M?ssbauer data provides us with partial (though not complete) explanation. The observed behavior at the K edge might thus indicate that the actual valence of Fe in Sr2FeMoO6.0 is not precisely 2.5 but a little higher. In conclusion, we employed Fe XANES spectroscopy at both K and L2,3 edges to show that iron in the halfmetallic Sr2FeMoO6.0 magnetoresistor possesses an FeII/III mixed-valence state. The key for the reliable interpretation of the spectral features was to use a full series of doubleperovskite samples, Sr2Fe(Mo,W/Ta)O6.0, as a reference system where the valence of Fe gradually varies from II to III. At both the edges, K and L2,3, intermediate spectral features were revealed for Mo-rich samples as compared with those for the strongly W- and Tasubstituted samples. The mixed-valence state of iron as thus confirmed for Sr2FeMoO6.0 is in line not only with the M?ssbauer data9,19,20 but also with the result of band-structure calculation1, suggesting a strong mixing of the itinerant d electron from nominally pentavalent Mo and the minority spin t2g band of nominally trivalent Fe. It also agrees with the neutron diffraction data showing reduced magnetic moment values at both the Fe and Mo sites from the values expected for high-spin FeIII (t2g3eg2; S = 5/2) and MoV (t2g1; S = 1/2).10 Finally we like to emphasize that perovskite-derived mixed-valent Fe oxides are gaining considerable research interest owing to the phenomena such as charge separation and ordering in (La,Sr)FeO327 and in REBaFe2O528 (RE is rare earth element) and large magnetoresistance in SrFeO2.9529 that are intimately related to the valence of iron. Thus the results of the present work are expected to be of wider importance in understanding the (mixed) valence state of iron in a larger group of perovskite oxides, as probed by various experimental techniques. Acknowedgments. The present work has been supported by a Grant-in-Aid for Scientific Research (contract No. 11305002) from the Ministry of Education, Science and Culture of Japan, and also through the JSPS Research Fellowship Program for Young Scientists (No. 14006635; Y.Y.). T. Yamamoto is thanked for his contribution in sample preparation.
References (1) (2) (3) Kobayashi, K.-I.; Kimura, T.; Sawada, H.; Terakura, K.; Tokura, Y. Nature (London) 1998, 395, 677. Kobayashi, K.-I.; Kimura, T.; Tomioka, Y.; Sawada, H.; Terakura, K.; Tokura, Y. Phys. Rev. B 1999, 59, 11159. Longo, J. M.; Ward, R. J. Am. Chem. Soc. 1961, 83, 2816; Sleight, A. W.; Longo, J. M.; Ward, R. Inorg. Chem. 1962, 1, 245; Pattersson, F. K.; Moeller, C. W.; Ward, R. Inorg. Chem. 1963, 2, 196; Nakagawa, T. J. Phys. Soc. Jpn. 1968, 24, 806. (4) (5) (6) (7) (8) (9) Woodward, P.; Hoffmann, R.-D.; Sleight, A. W. J. Mater. Res. 1994, 9, 2118. Sleight, A. W.; Weiher, J. F. J. Phys. Chem. Solids 1972, 33, 679. Goodenough, J. B.; Dass, R. I. Int. J. Inorg. Mater. 2000, 2, 3. Sarma, D. D. Current Opinion in Solid State and Materials Science 2001, 5, 261. Yamamoto, T.; Liimatainen, J.; Lindén, J.; Karppinen, M.; Yamauchi, H. J. Mater. Chem. 1999, 10, 2342. Lindén, J.; Yamamoto, T.; Karppinen, M.; Yamauchi, H.; Pietari, T. Appl. Phys. Lett. 2000, 76, 2925. (10) Chmaissem, O.; Kruk, R.; Dabrowski, B.; Brown, D. E.; Xiong, X.; Kolesnik, S.; Jorgensen, J. D.; Kimball, C. W. Phys. Rev. B 2000, 62, 14197. (11) Greneche, J. M.; Venkatesan, M.; Suryanarayanan, R.; Coey, J. M. D. Phys. Rev. B 2001, 63, 174403. (12) Algarabel, P. A.; Morellon, L.; De Teresa, J. M.; Blasco, J.; Garcia, J.; Ibarra, M. R.; Hernández, T.; Plazaola, F.; Barandiarán, J. M. J. Magn. Magn. Mater. 2001, 226-230, 1089. (13) Ray, S.; Kumar, A.; Sarma, D. D.; Cimino, R.; Turchini, S.; Zennaro, S.; Zema, N. Phys. Rev. Lett. 2001, 87, 97204. (14) Moreno, M. S.; Gayone, J. E.; Abbate, M.; Caneiro, A.; Niebieskikwiat, D.; Sánchez, R. D.; de Siervo, A.; Landers, R.; Zampieri, G. Solid State Commun. 2001, 120, 161. (15) Kang, J.-S.; Kim, J. H.; Sekiyama, A.; Kasai, S.; Suga, S.; Han, S. W.; Kim, K. H.; Muro, T.; Saitoh, Y.; Hwang, C.; Olson, C. G.; Park, B. J.; Lee, B. W.; Shim, J. H.; Park, J. H.; Min, B. I. Phys Rev. B 2002, 66, 113105. (16) Kapusta, Cz.; Riedi, P. C.; Zajac, D.; Sikora, M.; De Teresa, J. M.; Morellon, L.; Ibarra,
M. R. J. Magn. Magn. Mater. 2002, 242-245, 701. (17) Blasse, G. J. Inorg. Nucl. Chem. 1965, 27, 993; Nakagawa, T.; Yoshikawa, K.; Nomura, S. J. Phys. Soc. Jpn. 1969, 27, 880; Kawanaka, H.; Hase, I.; Toyama, S.; Nishihara, Y. J. Phys. Soc. Jpn. 1999, 68, 2890. (18) Galasso, F.; Katz, L.; Ward, R. J. Am. Chem. Soc. 1959, 81, 820. (19) Lindén, J.; Yamamoto, T.; Karppinen, M.; Yamauchi, H. Appl. Phys. Lett. 2001, 78, 27. (20) Lindén, J.; Yamamoto, T.; Nakamura, J.; Yamauchi, H.; Karppinen, M. Phys. Rev. B 2002, 66, 184408. (21) Kobayashi, K.-I.; Okuda, T.; Tomioka, Y.; Kimura, T.; Tokura, Y. J. Magn. Magn. Mater. 2000, 218, 17. (22) Dass, R. I.; Goodenough, J. B. Phys. Rev. B 2001, 63, 64417. (23) Richardson, F. D.; Jeffes, J. H. E. J. Iron Steel Inst. 1948, 160, 261. (24) Chan, T. S.; Liu, R. S.; Guo, G. Y.; Hu, S. F.; Lin, J. G.; Chen, J. M.; Attfield, J. P. Chem. Mater. 2003, 15, 425. (25) Crocombette, J. P.; Pollak, M.; Jollet, F.; Thromat, N.; Gautier-Soyer, M. Phys. Rev. B 1995, 52, 3143. (26) Gautier-Soyer, M. J. Eur. Cer. Soc. 1998, 18, 2253. (27) Li, J. Q.; Matsui, Y.; Park, S. K.; Tokura, Y. Phys. Rev. Lett. 1997, 79, 297. (28) Lindén, J.; Karen, P.; Kjekshus, A.; Miettinen, J.; Pietari, T.; Karppinen, M. Phys. Rev. B 1999, 60, 15251; Karen, P.; Woodward, P. M.; Lindén, J.; Vogt, T.; Studer, A.; Fischer, P. Phys. Rev. B 2001, 64, 214405. (29) Zhao, Y. M.; Mahendiran, R.; Nguyen, N.; Raveau, B.; Yao, R. H. Phys. Rev. B 2001, 64, 24414.
Figure Captions Fig. 1. (a) Fe L3-edge XANES spectra for a series of Sr2Fe(Mo,W/Ta)O6.0 samples in the energy range of 704 - 714 eV, and (b) the intensity ratio, I709/I707, of the two peaks at ?707 and ?709 eV (approximated by the peak heights) as plotted against the W/Ta substitution level. Fig. 2. Fe K-edge XANES spectra (a) for the selected Sr2Fe(Mo,W/Ta)O6.0 samples and the reference oxides, FeIIO, FeII/III3O4 and FeIII2O3 in the energy range of 7100 - 7150 eV, and (b) for all the Sr2Fe(Mo,W/Ta)O6.0 samples in the energy range of 7115 - 7130 eV.
90 % W 70 % 60 %
50 % 20 % 10 % Mo 40 % 50 % 60 % 80 % 100 % Ta
Karppinen et al: Fig. 1(a).
0.0 100 % W
100 % Ta
Karppinen et al: Fig. 1(b).
7110 7120 7130 7140 7150
Karppinen et al: Fig. 2(a).
90% W 80% W 70% W 60% W 50% W 20% W 10% W 20% Ta 40% Ta 50% Ta 60% Ta 80% Ta 100% Ta
Karppinen et al: Fig. 2(b).
copyright ©right 2010-2020。