J. Phys. Chem. B 2006, 110, 23817-23820
23817
Large Low-Field Magnetoresistance in Nanocrystalline Magnetite Prepared by Sol-Gel Method Z. L. Lu, W. Q. Zou, L. Y. Lv, X. C. Liu, S. D. Li, J. M. Zhu, F. M. Zhang,* and Y. W. Du Jiangsu ProVincial Laboratory for Nanotechnology, National Laboratory of Solid State Microstructure and Department of Physics, Nanjing UniVersity, Nanjing 210093, China ReceiVed: February 8, 2006; In Final Form: July 27, 2006
Nanocrystalline magnetite Fe3O4 samples with a grain size of about 40 nm have been synthesized by an optimized sol-gel method. The single phase of spinel magnetite was confirmed by both X-ray diffraction and transmission electron microscopy. It has been found that the magnetoresistance of the samples at low field (LFMR) is relatively large, and with the decrease of temperature its value at a field of 0.5 T changes dramatically from -2.5% at 300 K to -17.0% at 55 K. With the further decrease of temperature a sharp drop occurs for the magnitude of the magnetoresistance (MR), regarded as a spin (cluster) glass transition in the surface region of the grains that can be confirmed by the zero-field-cooled and field-cooled magnetization and ac susceptibility measurement. The mechanism of the magnetic and transport properties was discussed.
1. Introduction
2. Experimental Methods
Recently half-metallic ferromagnets with 100% spin polarization at the Fermi level have been of considerable interest in spin electronics applications such as magnetic random access memory (MRAM), magnetic reading heads, and magnetic sensors concerning the magnetoresistance that depends on the spin polarization of the materials used.1-3 Among them, most of the half-metallic ferromagnets are not suitable for magnetoelectronic applications in an extended temperature range due to their low Curie temperatures, i.e., 395 K for CrO2, 360 K for La0.7Sr0.3MnO3, and 265 K for La0.67Ca0.33MnO3.3-5 On the contrary, magnetite, the ferromagnetic inverse spinel Fe3O4, has shown the highest known Curie temperature of 860 K, and band structure calculation for the system indicated that at the Fermi level electrons are present with one spin orientation only.6,7 Therefore, a lot of work has been carried out for the variety forms of magnetite materials, including single crystal, epitaxial and polycrystalline films, particles, colloidal nanocrystal arrays, multiple films heterostructure, and nanocontacts.8-16 A relatively large low-field magnetoresistance (LFMR) was indeed found in Fe3O4 nanocontact, colloidal nanocrystal arrays and Fe3O4/SrTiO3/La0.7Sr0.3MnO3 heterostructure.8,9,16 However, it was disappointing that in most cases large magnetoresistance could only be obtained at a high field and the corresponding LFMR was always fairly low for both bulk and film crystalline magnetite.10-13 The possible reasons including the nonstoichiometry grain boundaries, antiphase boundaries effect, surface spin arrangement, and strong coupling between Fe3O4 grains have been issues for study.12,13 In this work, we report the observation of a large LFMR in nanocrystalline magnetite Fe3O4 samples prepared by an optimized sol-gel method. It is believed that the large LFMR is caused due to the spin-polarized tunneling of conducting electrons through the insulating boundaries in nanocrystalline magnetite.
The nanocrystalline Fe3O4 samples were prepared by an optimized sol-gel method, a superior technique for synthesizing samples with good homogeneous and small size grains. First, N mol of Fe(NO3)3‚9H2O was dissolved in deionized water at 50 °C to form a clean solution. Then 3N mol of citric acid and 6N mol of ethylene glycol were added, and the pH of the solution was adjusted to be about 7 using ammonia. The mixed solution was heated at 60 °C for 24 h for slow evaporation under constant stirring to allow the gel to be formed and dried at 150 °C in an oven to form gelatinous precursor. The prepared gel was first precalcined at 850 °C for 4 h in air to form fine particles. After the powders were cold-pressed into disk-shaped pellets under a pressure of about 1000 kg/cm2, the pellets were finally sintered at 1300 °C for 2.5 h in pure nitrogen atmosphere. Phase and microstructure analysis were carried out using X-ray diffraction and transmission electron microscopy. The sintered disks were polished before electric measurements to ensure a good electrical contact. The magnetic properties, resistivity, and MR were measured using a MPMS (Quantum Design XL-7T) incorporated with a standard four-probe measurement.
* To whom correspondence should be addressed. E-mail: fmzhang@ nju.edu.cn.
3. Results and Discussion Figure 1a shows the typical X-ray diffraction pattern of the samples, indicating the single phase of Fe3O4 in the samples. The average grain size can be calculated as about 40 nm using the Scherror formula from the full width at half-maximum of the peak (311). This is confirmed by the corresponding transmission electron micrographs with bright-field image and selected area diffraction (SAED), as shown in Figure 1b. An average grain size of about 40 nm is actually demonstrated, consistent with the results from the XRD. All diffraction rings can be attributed to polycrystalline spinel Fe3O4, further confirming the single phase of our nanocrystalline magnetite. The field dependence of magnetization for the nanocrystalline Fe3O4 samples at both 5 and 300 K is shown in Figure 2a. The saturated magnetization at 5 K is found as 3.84 µB/f.u., a value slightly less than 4.0 µB/f.u. for the ferromagnetic configuration
10.1021/jp0608325 CCC: $33.50 © 2006 American Chemical Society Published on Web 10/17/2006
23818 J. Phys. Chem. B, Vol. 110, No. 47, 2006
Figure 1. (a) X-ray diffraction pattern of the nanocrystalline magnetite samples and (b) corresponding plane view of TEM with the corresponding SAED pattern inserted.
Figure 2. (a) Magnetic hysteresis loops of the nanocrystalline Fe3O4 at 5 and 300 K, respectively; inset shows low-field details. (b) Magnetization as a function of temperature measured under ZFC and FC conditions.
of Fe3+ and Fe2+ ions, possibly due to the small grain size and the nonstoichiometry grain boundaries.17 It is noted that the coercivity of the samples is quite low, with a value of about 60 Oe at 5 K and 30 Oe at 300 K, because of the low magnetocrystalline anisotropy.18 The Curie temperature mea-
Lu et al.
Figure 3. Temperature dependence of the in-phase (real) component χ′ (a) and out-of-phase (imaginary) component χ′′ (b) of the magnetic susceptibility for the nanocrystalline magnetite at three different frequencies around 55 K; inset shows corresponding patterns of whole temperature range. Data taken with a driving field of H0 ) 3 Oe.
sured as about 850 K is consistent with the value of 860 K for the system. Figure 2b shows the typical temperature dependence of magnetization for the nanocrystalline magnetite. A magnetic field of 100 Oe was used for both the zero-field-cooled (ZFC) and field-cooled (FC) measurement. The ZFC data clearly show the Verwey transition28 at about 115 K with the characteristic sharp drop, while no obvious change is observed in the FC curve. For Fe3O4 nanocrystalline films, a similar phenomenon has been observed and was explained by the hysteresis behavior.19 Moreover, a step observed at about 55 K (Tf) from the ZFC data indicates a spin (cluster) transition, which has often been seen in other nanocrystalline ferrites because of the frozen canted spins in the grain surface region.20-23 To better understand the spin dynamics near the transition temperature, the real (in-phase) and imaginary (out-of-phase) components of the ac susceptibility for the nanocrystalline Fe3O4 were investigated as a function of temperature with a driving field of amplitude of H0 ) 3 Oe, as shown in Figure 3. Data were collected on warming from 5 to 200 K after zero-field cooling of the samples. It can be seen that around 50 K the in-phase susceptibility component χ′ depends strongly on frequency, reflecting a marked slowing down of the dynamics from Figure 3a. Figure 3b indicates the temperature dependence of the out-of-phase susceptibility component χ′′, and it can be seen that the temperature for the maximum shifts toward higher values with increasing frequency. This behavior recalls the features of a reentrant spin glass and suggests a transition from a high-temperature ferromagnetic state to a low-temperature disordered frozen magnetic state at the grain surfaces near the boundary in this case.24,25 It may be noted that no significant increase was observed in coercivity below this temperature. However, when the spin glass transition at the grain surface occurs the surface magnetic anisotropy of the grain would significantly increase and a significant increase in coercivity would be expected due to the strong coupling between the core
Magnetoresistance in Nanocrystalline Magnetite
J. Phys. Chem. B, Vol. 110, No. 47, 2006 23819
Figure 4. Resistivity as a function of temperature (9) and log F vs T-1/2 (inset) with the resistivity of the bulklike thick Fe3O4 films 660 nm in thickness (-).
and the disordered surface of the grain. The lack of significant increase in coercivity below the transition temperature for nanocrystalline Fe3O4 in this work is believed to be due to the weak exchange coupling effect between the core and surface of the grain. Furthermore, hysteresis loops were also measured at 5 K after being field cooled in a 5 T field, and no obvious loop shift was observed, indicating the lack of strong exchange coupling between the grain surface and core, consistent with the results obtained in magnetite nanoparticles by Goya et al.26 The resistivity of the nanocrystalline Fe3O4 samples was measured as about 0.3 Ω cm at 300 K, which is nearly two orders larger than that of Fe3O4 single crystal or epitaxial films. Figure 4 shows the temperature dependence of the resistivity for the samples with inset for the log F - T-1/2 relation, demonstrating a typical grain boundary tunneling conductance mechanism in most of the temperature range studied.27,28 It is known that the Verwey transition has an abrupt resistivity change. However, no obvious abrupt change was revealed in the Fe3O4 system. It could be reasonably interpreted as suppression of the effect by the large boundary resistivity because resistivity change caused during the Verwey transition is less than 1% of the total resistivity in the nanocrystalline Fe3O4 samples. For comparison, the resistivity of bulklike Fe3O4 thick films 660 nm in thickness14 is also shown. The field dependence of MR, defined as [FH - F0]/F0, has been measured at various temperatures and is shown in Figure 5a. It can be seen that the MR values first increase quickly with the field applied (H e 0.5 T) and then increase slowly with the further increase of field. The results are similar to the MR effect of some polycrystalline perovskite oxide but are significantly different from the results in magnetite, which always shows large MR only at high field. As discussed above, the conduction mechanism here is the tunneling of conduction electrons between adjacent grains. For this case, the extrinsic MR can be expressed as MR ∝ 〈cos φ〉 ∝ 〈cos θ〉 ) a(M/MS)2, where a is a constant, φ is the angle between the magnetization directions of two adjacent grains, and θ is the angle between the magnetization directions of a grain and the applied field.29 According to this relation the MR can be simulated by choosing a ) -0.12, and the result is shown in Figure 5b. It is found that the MR at low field follows the simulated value well. This fact indicates that the magnetoresistance at a temperature over 55 K is predominantly determined by the magnetic tunneling process. To our knowledge this is the first time the large LFMR following the (M/MS)2 relationship in single-phase nanocrystalline magnetite has been observed.
Figure 5. (a) Field dependence of MR ratio for nanocrystalline Fe3O4 at 45, 55, 100, and 300 K, and (b) MR dependence on field at 100 K (symbols) compared with the simulated results; inset shows low-field details.
Figure 6. Temperature dependence of MR at 0.5 T and the corresponding deduced spin polarization.
The further slow increase of the MR at high field is attributed to the more ordered spin alignment in the boundary barriers between grains at the field, representing less scattering at the tunneling barriers for the conduction electrons.2 Furthermore, the temperature dependence of the MR of the samples at 0.5 T was investigated and is shown in Figure 6. Clearly, the value of the MR for the nanocrystalline Fe3O4 samples increases with the decrease of temperature down to about 55 K and then turns into a sharp decrease with the further decrease of temperature. Interestingly, by calculating the spin polarization of the nanocrystalline Fe3O4 from MR ) P2/(1 + P2) for the granular ferromagnets,30 the deduced temperature change of the spin polarization is found to be similar to that of the MR, which indicates the MR change with temperature is due to the spin polarization change and little spin flip occurs
23820 J. Phys. Chem. B, Vol. 110, No. 47, 2006 for the tunneling of electrons between nanocrystalline Fe3O4 grains. A large low field (0.5 T) a MR value of about 17.0% at 55 K mainly originates from the boundary tunneling of spinpolarized electron with high-spin polarization of the nanocrystalline Fe3O4 samples. For the drop of the MR occurring with the further decrease of temperature below 55 K, as discussed above, there is a reentrant spin glass which suggests a transition from a high-temperature ferromagnetic state to a low-temperature disordered frozen magnetic state in the grain surfaces near the boundary. The spin glass transition causes the spins in the grain surface regions to be “frozen” disordered or the spins could hardly be rotated by the applied field and a state of disorder for the spins is kept. The “frozen” disordered spins in the surface regions give rise to more scattering for the transport of polarized electrons and cause some spin-polarized electrons to flip before tunneling from one grain to the neighboring one. Consequently, the decrease of the spin polarization of the conduction electrons when transporting through the surface regions finally reduces the MR with decreasing temperature below 55 K. 4. Conclusions Investigation of the magnetic and transport properties of single-phase nanocrystalline Fe3O4 samples synthesized by solgel method has been carried out, and large low-field MR was revealed for the samples. It was found that the MR at low field is due to alignment of the magnetization between adjacent grains, while the one at high field is caused from the more ordered spin alignment at the boundary barriers. The large lowfield MR in nanocrystalline Fe3O4 samples at low temperature is due to high-spin polarization, large density of grain boundaries, and little spin flip when tunneling for the nanocrystalline samples. The drop of the MR below 55 K is caused by spin glass transition with “frozen” disordered spins in the surface regions. Acknowledgment. This work was supported by the National Key Program for the Fundamental Research Development Plan of China (973 Project-2005CB623605) and National Science Foundation of China (NSFC, #10374044 and #10474037). References and Notes (1) Pickett, W. E.; Moodera, J. Phys. Today 2001, 54, 39.
Lu et al. (2) Ziese, M. Pept. Prog. Phys. 2003, 65, 143. (3) Dowben, P. A.; Skomski, R. J. Appl. Phys. 2004, 95, 7453. (4) Snyder, G. J.; Hiskes, R.; Dicarolis, S.; Beasley, M. R.; Geballe, T. H.; Phys. ReV. B 2002, 53, 14434. (5) Coey, J. M. D.; Venkatesan, M. J. Appl. Phys. 2002, 91, 8345. (6) Yanase, Y.; Siratori, K. J. Phys. Soc. Jpn. 1984, 53, 312. (7) Zhang, Z.; Satpathy, S. Phys. ReV. B 1991, 44, 13319. (8) Hu, G.; Suzuki, Y. Phys. ReV. Lett. 2002, 89, 276601. (9) Versluijs, J. J.; Bari, M. A.; Coey, J. M. D. Phys. ReV. Lett. 2001, 87, 026601. (10) Coey, J. M. D.; Berkowitz, A. E.; Balcells, L.; Putris, F. F.; Parker, F. T. J. Appl. Phys. 1998, 72, 734. (11) Eerenstein, W.; Palstra, T. T. M.; Saxena, S. S.; Hibma, T. Phys. ReV. Lett. 2002, 88, 247204. (12) Liu, K.; Zhao, L.; Klavins, P.; Osterloh, F. E.; Hiramatsu, H. J. Appl. Phys. 2003, 93, 7951. (13) Liu, H.; Jiang, E. Y.; Bai, H. L.; Zheng, R. K.; Wei, H. L.; Zhang, X. X. Appl. Phys. Lett. 2003, 83, 3531. (14) Gong, G. Q.; Gupta, A.; Xiao, G.; Qian, W.; Dravid, V. P. Phys. ReV. B 1997, 56, 9. (15) Zeng, H.; Black, C. T.; Sandstrom, R. L.; Rice, P. M.; Murray, C. B.; Sun, S. Phys. ReV. B 2006, 73, 020402. (16) Poddar, P.; Fried, T.; Markovich, G. Phys. ReV. B 2002, 65, 172405. (17) Matsuyama, K. J. Magn. Soc. Jpn. 2001, 25, 51. (18) Gong, G. Q.; Gupta, A.; Xiao, G.; Qian, W.; Dravid, V. P. Phys. ReV. B 1997, 56, 5096. (19) Tang, J.; Wang, K. Y.; Zhou, W. J. Appl. Phys. 2001, 89, 7690. (20) Kodama, R. H.; Berkowitz, A. E.; McNiff, E. J.; Foner, S., Jr. J. Appl. Phys. 1997, 81, 5552. (21) Oliver, S. A.; Handeh, H. H.; Ho, J. C. Phys. ReV. B 1999, 60, 3400. (22) Bhowmik, R. N.; Ranganathan, R.; Sarkar, S.; Bansal, C.; Nagarajan, R. Phys. ReV. B 2003, 68, 134433. (23) Stewart, S. J.; Mercader, R. C.; Vandenberghe, R. E.; Cernicchiaro, G.; Scorzelli, R. B. J. Appl. Phys. 2005, 97, 054304. (24) Coles, B. R.; Sarkissian, B. V. B.; Taylor, R. H. Philos. Magn. B 1978, 37, 489. (25) Bonetti, E.; Del Bianco, L.; Fiorani, D.; Rinaldi, D.; Caciuffo, R.; Hernando, A. Phys. ReV. Lett. 1999, 83, 2829. (26) Goya, G. F.; Berquo, T. S.; Fonseca, F. C.; Morales, M. P. J. Appl. Phys. 2003, 94, 3520. (27) Xiao, J. Q.; Jiang, J. S.; Chien, C. L. Phys. ReV. Lett. 1992, 68, 3749. (28) Ziese, M.; Hohne, R.; Semmelhack, H. C.; Rechentin, H.; Hong, N. H.; Esquinazi, P. Eur. Phys. J. B 2002, 28, 415. (29) Xiao, J. Q.; Jiang, J. S.; Chien, C. L. Phys. ReV. Lett. 1992, 68, 3749. (30) Inous, I.; Maekawa, S. Phys. ReV. B 1996, 53, R11927.
