Journal of the Korean Physical Society, Vol. 52, No. 5, May 2008, pp.

14831486

Magnetic Anisotropies in Cobalt-Nickel Ferrites (Nix Co1x Fe2 O4 )

Binghai

Department of Materials Science & Engineering, National University of Singapore, Singapore 119260
Zhili

Liu,

Jun

Ding,

Jiabao

Yi and Jianhua Yin

School of Materials Science & Engineering, National University of University, Singapore 119260
(Received 16 July 2007) This work presents a detailed study on the various magnetic anisotropies of Nix Co1x Fe2 O4 prepared by using a mechanochemical process. The results indicated that the magnetic anisotropy of Nix Co1x Fe2 O4 was strongly correlated with the Co2+ concentration. The Ni2+ substitution readily led to a decrease in the magnetocrystalline anisotropy and coercivity. While a large uniaxial anisotropy was induced in CoFe2 O4 by magneto-annealing, the uniaxial anisotropy decreased with the Ni concentration in Nix Co1x Fe2 O4 and disappeared in the pure NiFe2 O4 . A short-time mechanical milling was found to signicantly increase the coercivity in Nix Co1x Fe2 O4 samples and high coercivities of 5.1 kOe and 2.8 kOe were achieved in the CoFe2 O4 and the Ni0:5 Co0:5 Fe2 O4 samples respectively. The microstructure and magnetic studies indicated that the milling-induced high coercivity could be ascribed to milling-induced stress anisotropy and to the pinning eects of defects.

Dong

PACS numbers: 74.25.Ha, 75.30.Gw, 75.50.Gg, 75.50.Vv Keywords: Magnetic anisotropy, Stress anisotropy, Mechanical milling
I. INTRODUCTION

Cobalt ferrite (CoFe2 O4 ) is an important magnetic oxide for its potential in various hard magnetic applications, such as magnetic data storage and magnetooptical and magneto-electric devices [1{4]. Its spinel structure consists of two sublattices with intrinsic magnetic hardness stemming from the incompletely quenched angular momentum of Co2+ ions [5]. The magnetic properties of CoFe2 O4 show a strong dependence on its magnetic anisotropies, not only its intrinsic magnetocrystalline anisotropy but also its induced anisotropies. Large uniaxial anisotropy is induced in both singlecrystal [6] and nanostructured [7] CoFe2 O4 materials with the magneto-annealing processes. Large stress anisotropy is also reported to make signicant contributions to the high coercivities achieved in both CoFe2 O4 thin-lm (12.5 kOe) [8] and powdered (5.1 kOe) [9] materials. As all these induced high anisotropies in CoFe2 O4 occurred due to its large intrinsic magnetocrystalline anisotropy, it would be of interest to systematically examine the induced anisotropies in Ni2+ substituted cobalt ferrite (Nix Co1x Fe2 O4 ). Mechanical alloying has been reported to be a powerful tool for the study of nanostructured materials [10]. The purpose of
E-mail: msedingj@nus.edu.sg; Fax: +65-67763604

this work is to study the eects of Ni2+ substitution on the various magnetic anisotropies of Nix Co1x Fe2 O4 . Detailed magnetic studies and structure analyses are presented in order to understand the magnetic properties associated with the various magnetic anisotropies of Nix Co1x Fe2 O4 materials, i.e. the magnetocrystalline anisotropy, the magnetoannealing-induced anisotropy and the stress anisotropy.
II. EXPERIMENTS

-1483-

Nix Co1x Fe2 O4 (x from 0 to 1) powders were prepared by using a mechanochemical process in a Spex 8000 miller with metal Co, Ni and -Fe2 O3 as the starting materials. The ball-to-powder weight ratio was xed at 12 : 1. The as-milled powders were further annealed at 000  C for 2 hours with a slow cooling process at a rate of 1  C/min [11, 12]. To study the magneto-annealing-induced anisotropy, we annealed the Nix Co1x Fe2 O4 samples at 300  C for 10 hours under an applied eld of 10 kOe using a high-temperature vibration sample magnetometer (VSM, lakeshore). For studying stress-induced anisotropies, we mechanically milled the Ni0:5 Co0:5 Fe2 O4 sample for dierent periods of time by using a Spex 8000 miller with a ball to powder weight ratio of 12 : 1. X-ray diraction

-1484-

Journal of the Korean Physical Society, Vol. 52, No. 5, May 2008

Fig. 1. (a) XRD pattern and (b) bright-eld TEM image (inserted: selected area electron diraction) of Ni0:5 Co0:5 Fe2 O4 annealed at 1000  C for 2 hours.

Fig. 2. (a) Saturation magnetization (MS ) and Curie temperature and (b) coercivities (HC ) and magnetocrystalline anisotropies (K1 ) of Nix Co1x Fe2 O4 samples after thermal annealing at 1000  C.

(XRD, Bruker Cu K), transmission electron microscopy (TEM, JEOL JEM 3010) and Fe57 Mssbauer speco troscopy (FAST ComTec) were used for phase and structure analyses of the Nix Co1x Fe2 O4 samples. Based on the XRD analysis, we analyzed the micorstrain in the milled Ni0:5 Co0:5 Fe2 O4 samples by using the Williamson-Hall plot method [9]. The magnetic properties of the Nix Co1x Fe2 O4 samples were studied using a vibration sample magnetometer (VSM, lakeshore 4500).
III. RESULTS AND DIDCUSSION

Both the XRD and the TEM analyses indicated that the Nix Co1x Fe2 O4 samples annealed at 1000  C were all of the well-crystallized single spinel phase. Figure 1(a) and (b) show, respectively, a typical XRD pattern and a bright-eld TEM image for the annealed Ni0:5 Co0:5 Fe2 O4 sample. As the bright-eld TEM image shows, the annealing at 1000  C led to a hard agglomeration and sintering of submicrometer-size grains. Both the bright-eld and the dark-eld TEM analyses indicated a low strain level in the annealed samples, as evidenced by the uniform contrast inside the grains and

along the grain boundaries. The dark-eld TEM analysis indicated that the average grain size of the annealed Nix Co1x Fe2 O4 samples was around 200 nm, consistent with the results of the XRD analysis. The Mssbauer o analysis indicated that all the annealed samples had a site distributions of FesiteA [(Co1x Nix )Fe]siteB O4 . The Curie temperature analysis indicated only a single transition from the ferromagnetic state to paramagnetic state for each sample. With increasing Ni concentration (x value), the Curie temperature TC continuously increased from 516  C (x = 0, CoFe2 O4 ) to 581  C (x = 1, NiFe2 O4 ) (Figure 2(a)). The measured TC of CoFe2 O4 and NiFe2 O4 agreed well with those reported [13]. Both the saturation magnetization (Ms, Figure 2(a)) and the coercivity (HC , Figure 2(b)) of Nix Co1x Fe2 O4 decreased with increases in Ni concentration (x value). Based on the law of approach to saturation, magnetocrystalline anisotropy constants (K1 ) of Nix Co1x Fe2 O4 samples were estimated, as shown in Figure 2(b). It is clear that the Ni substitution led to a continuous decrease in K1 . In terms of the similar microstructure features of Nix Co1x Fe2 O4 samples, as indicated by the TEM and he XRD analyses, the decrease in HC should be ascribed to the decrease in K1 with Ni

Magnetic Anisotropies in Cobalt-Nickel Ferrites (Nix Co1x Fe2 O4 ) { Binghai Liu et

al.

-1485-

Fig. 3. (a) Hysteresis loop of Ni0:5 Co0:5 Fe2 O4 samples measured in the directions perpendicular and parallel to the eld of magnetoannealing (inserted: induced uniaxial anisotropies of Nix Co1x Fe2 O4 after magnetic annealing at 300  C under 10 kOe), (b) milling time-dependent MS and HC , (c) microstrain and anisotropy (K ) of milled Ni0:5 Co0:5 Fe2 O4 samples at room-temperature.

substitution. These results conrmed that the magnetocrystalline anisotropy of Nix Co1x Fe2 O4 strongly depended on the Co2+ concentration. It is well known that a uniaxial anisotropy can be induced in cobalt ferrites by magneto-annealing. The as-induced anisotropy has been ascribed to the uniaxial spin-lattice interaction energy that couples the spins of Co2+ ions and the axis of trigonal symmetry of B sites [5]. In this study, a large uniaxial anisotropy (Kinduced ) of 2.5 104 J/m3 was induced in CoFe2 O4 after annealing at 300  C for 10 hr under 10 kOe, as shown in Figure 3(a). The Kinduced was calculated based on the dierence in the loops measured in the directions parallel and perpendicular to the magnetic eld for the magnetic annealing (Figure 3(a)) [7]. Figure 3(a) shows that, with increasing Ni concentration, it was clear that the Kinduced of Nix Co1x Fe2 O4 decreased rapidly. No appreciable Kinduced could be observed when x value was larger than 0.5, as evidenced by the small dierence in the loops measured in the two directions. These results conrmed that the magneto-annealing induced anisotropy was closely related to Co2+ ions and the changes in the Co2+ concentration readily led to changes in the annealing-induced anisotropy. Recently, we reported that a short-time mechanical milling could lead to a high coercivity of 5.1 kOe in large-grained CoFe2 O4 powders, which was ascribed to milling-induced large stress anistropy and to the pinning eects of defects [9,14]. In order to check whether mechanical milling could induce high coercivities in Nix Co1x Fe2 O4 samples, Ni0:5 Co0:5 Fe2 O4 sample annealed at 1000  C was selected for a detailed study, as it still had a relatively large magnetocrystalline anisotropy (Figure 2(b)). The Ni0:5 Co0:5 Fe2 O4 sample was milled for dierent periods of time. A short-time milling was found to induce a large increase in HC and the maximum

HC , as achieved in the Ni0:5 Co0:5 Fe2 O4 sample, was 2.80 kOe after milling for 1 hour, corresponding to 496 % HC increase (from initial 0.47 kOe before milling). Figure 3(b) and (c) show the milling-time-dependent HC , microstrain and anisotropy (K ) of milled Ni0:5 Co0:5 Fe2 O4 samples. With the progress of milling, HC increased rapidly, accompanied by a rapid increase in both the micorstrain (estimated from Williamson-Hall plots) and the magnetic anisotropies K (estimated by using the law of approach of saturation). Further milling led to a decrease in the HC , the microstrain and the K, which nally stabilized at certain values after long-time milling. The changing tendencies of the HC , the microstrain and the K were similar to those of the milled CoFe2 O4 sample [9]. As indicated by Figure 3(b) and (c), the highest HC corresponded to the highest microstrain and the highest magnetic anisotropy, which implies that the high HC in the sample milled for 1 hr is closely related to the millinginduced high-level microstrain. The anisotropy analysis revealed an $64 % increase in the K (from 1.96 105 to 3.21 105 J/m3 ) for this sample. Such an anisotropy enhancement can be ascribed to a milling-induced large strain, i.e., to stress anisotropy. In order to study the coercivity mechanism behind the milling-induced high coercivity in the Ni0:5 Co0:5 Fe2 O4 samples, we examined the eld-dependent magnetization/demagnetization behaviors of the Ni0:5 Co0:5 Fe2 O4 sample milled for 1 h. Figure 4(a) shows the minor loops and the eld-dependent magnetization behaviors of the sample at room temperature (296 K). As seen from Figure 4(a), the coercive force H increased slowly at low applied elds and increased rapidly when the applied eld became larger than the saturation coercivity (2.8 kOe), i.e., Happlied /HC;max = 1. Such a demagnetization behavior is generally an indication of the domainwall pinning mechanism [15,16]. As is well known, me-

-1486-

Journal of the Korean Physical Society, Vol. 52, No. 5, May 2008

Fig. 4. (a) Field-dependent coercivity and the minor loops of the Ni0:5 Co0:5 Fe2 O4 samples milled for 1 h at 296 K (HC;max the saturation coercivity measured at 30 kOe, Happlied -the applied magnetic eld) and (b) TEM images of the Ni0:5 Co0:5 Fe2 O4 sample milled for 1 h.

chanical milling is a high-energy deformation process; thus, it is possible that milling-induced defects could be responsible for the domain-wall pinning coercivity mechanism. Therefore, we present a detailed TEM study on the Ni0:5 Co0:5 Fe2 O4 sample milled for 1 h. Both the bright-eld and the dark-eld TEM analyses revealed the formation of bountiful shear bands in this sample (as circled in Figure 4(b)). Under high-resolution TEM, a high density of dislocation-type defects was observed in these shear bands, together with strong lattice distortion. This indicates that a highly strained defective microstructure was induced by mechanical milling. These defects and the associated highly strained areas may act as active pinning sites for domain wall movement during the reversal magnetization process [9, 14]. Therefore, stress anisotropy and the pinning eects of the defective microstructures may account for the milling-induced high coercivity in the Ni0:5 Co0:5 Fe2 O4 sample, as they do for the milled CoFe2 O4 samples [9,14]. In addition, the milling-induced highest strain in Ni0:5 Co0:5 Fe2 O4 ($0.9 %) was of the same order as that in the milled CoFe2 O4 ($1.03 %). However, the milled Ni0:5 Co0:5 Fe2 O4 showed a lower milling-induced highest HC (2.8 kOe) when compared with that of CoFe2 O4 (5.1 kOe). This may be ascribed to the fact that Ni2+ has almost entirely quenched orbital momentum and thus a very small contribution to the spin-orbit coupling, as well as a smaller magnetostrcition. Therefore, Ni2+ substitution resulted in a decrease in magnetostriction CoFe2 O4 and, thus, a smaller milling-induced stress anisotropy and HC .
IV. CONCLUSION

induced anisotropies, high coercivities were achieved by milling the large-grained CoFe2 O4 and Ni0:5 Co0:5 Fe2 O4 samples. The milling-induced high coercivities could be ascribed to the stress anisotropy and to the pinning effects, both of which were created by the high-energy milling.
REFERENCES

Ni substitution in Nix Co1x Fe2 O4 materials resulted in appreciable decreases in both the magnetocrystalline anisotropy and the magnetoannealing-induced anisotropy, accompanied by decreases in the magnetic coercivities. While low coercivities were associated with the magnetocrystalline and the magneto-annealing-

[1] J. Ding, Y. J. Chen, Y. Shi and S. Wang, Appl. Phys. Lett. 77, 3621 (2000). [2] Y. Suzuki, R. B. van Dover, E. M. Gyorgy, J. M. Philips, V. Korenivski, D. J. Werder, C. H. Chen, R. J. Cava, J. J. Krajewski, W. F. Peck and K. B. Do, Appl. Phys. Lett. 68, 714 (1996). [3] J. Ding, T. Reynold, W. F. Miao, P. G. McCormick and R. Street, Appl. Phys. Lett. 65, 7074 (1994). [4] B. Zhou, Y. W. Zhang, Y. J. Yu, C. S. Liao, C. H. Yan, L. Y. Chen and S. Y. Wang, Phys. Rev. B 68, 24426 (2003). [5] W. Palmer, Phys. Rev. 120, 342 (1960). [6] R. F. Penoyer and L. R. Bickford, Jr., Phys. Rev. 108, 271 (1957). [7] Y. C. Wang, J. Ding, J. H. Yin, B. H. Liu, J. B. Yi and S. Yu, J. Appl. Phys. 98, 124306 (2005). [8] J. H. Yin, J. Ding, B. H. Liu, X. S. Miao and J. S. Chen, Appl. Phys. Lett. 88, 162502 (2006). [9] B. H. Liu and J. Ding, Appl. Phys. Lett. 88, 42506 (2006). [10] J. Ding, Y. Liu, P. G. McCormick and R. Street, J. Appl. Phy. 75, 1032 (1994). [11] J. Ding, P. G. McCormick and R. Street, J. Magn. Magn. Mater. 171, 309 (1997). [12] J. Ding, D. Maurice, W. F. Miao, P. G. McCormick and R. Street, J. Magn. Magn. Mat. 150, 417 (1995). [13] D. J. Craik, Magnetic Oxide (Wiley, London-New York, 1975). [14] B. H. Liu, J. Ding, Z. L. Dong, C. B. Boothroyd, J. H. Yin and J. B. Yi, Phys. Rev. B 74, 184427 (2006). [15] H. W. Zhang, C. B. Rong, J. Zhang, S. Y. Zhang and B. G. Shen, Phys. Rev. B 66, 184436 (2002). [16] H. Kronmller, Phys. Stat. Sol. (b) 144, 385 (1987). u