Magnetic properties of nanosized MnFeO   particles

 

M. Zheng³'*, X.C. Wu^b, B.S. Zou^b, Y.J. Wang³

 

' State Key Laboratory of Magnetism and Center of Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, P.O. Box

603, Beijing 100080, China ^ъ Institute of Physics, Chinese Academy of Sciences, Beijing 100080, China

Abstract

Nanosized MnFeO particles were prepared by chemical ultrasonic emulsion method. The as-prepared sample was found to be in amorphous state and showed spin-glass behavior at low temperature. The Curie temperature of the annealed sample is 160 K higher than that of the bulk material, which is thought to be due to finite-size scaling and also may be related to a nonequilibrium cation distribution. © 1998 Elsevier Science B.V. All rights reserved.

PACS: 61.46; 75.50.L

Keywords: Nanostructures; Chemical synthesis; Spin-glass behavior; Finite-size scaling

1. Introduction

Nanosized magnetic particles have attracted much interest not only due to unusual properties different from their corresponding bulk materials, but also from their promising technological ap­plications such as magnetic recording media, ferro-fluids, permanent magnets, etc. [1,2]. The general question is what happens to the bulk properties of a macroscopic body as one or more of its dimen­sions is reduced to atomic size. Of fundamental importance is the ferromagnetic systems or the spin-freezing temperature in spin-glass systems,

which in all studies has been found to be signifi­cantly lower than in the bulk. Some work has been done on Mn-ferrite particles, where a significant increase in T_c (relative to the bulk value) in par­ticles was found [3,4]. Some claimed that this effect was due to finite-size scaling [3,4] while others attributed it to the particle-size-dependent changes in the cation distribution [5-10]. In this article, we show a more detailed study of the magnetic proper­ties of MnFeO nanoparticles synthesized by ultrasonic emulsion method.

2. Experimental

Nanosized MnFeO particles were prepared in water-in-oil (W/O) microemulsions, which com­posed of water droplets surrounded by a layer of

surfactant molecules and dispersed in the oil phase. These water droplets can be used as nanopools to control the growth of particles. First, appropriate amounts of aqueous solutions of Mn²⁺ and Fe^{3 +} nitride salts were added into proper amounts of dodocal benzene sulfuric acid sodium salt (DBS) solution to generate mixed solution. Then appro­priate amounts of toluene were added into the above mixed solution and the W/O microemulsion was generated under continuous stirring. After 10 min stirring, the microemulsion was ultra-sonated for half an hour, then appropriate amounts of NaOH were added into it. The acoustic cavita-tion of ultrasound can make the particles disperse evenly. Finally, MnFeO sols were formed. By washing and distilling the sols, the sample of MnFeO nanoparticles coated with DBS was ob­tained (sample A). The fresh specimen was annealed at 400°C in vacuum with a pressure of 5 x 1СГ⁴ Pa for 1 h (sample B).

Structural characterization was performed by X-ray diffraction (XRD, Cu K radiation X " 1.5405 A). The size and shape of the particles were studied by transmission electron microscopy (TEM). A mean size of the spherical particles was about 5.6 nm for the annealed sample. Magneti­zation measurements were carried out by using extracting sample magnetometer in the temper­ature range from 3 to 300 K. A magnetic balance was used to determine the Curie temperature of the particles. All samples were kept in a vacuum of 1 x 10"⁴ Pa to resist oxidation during the heating process.

3. Results and discussion

Fig. 1a and Fig. 1b shows the XRD pattern for samples A and B. It indicates that the as-prepared sample (A) is in an amorphous state while the annealed sample (B) crystallizes well. The pattern of Fig. 1b fits Mn-ferrite well with the lattice con­stant almost the same as that of bulk MnFeO given by the ASTM card value (0.856 nm for our sample while 0.85 nm for bulk MnFeO). The Scherrer formula has also been used to determine the crystal size from the width of the XRD (3 1 1) line. The mean size of the particles is about 6.2 nm, which approximately agrees with that obtained by ТЕМ.

Fig. 2 shows the hysteresis loop of samples A and В measured at 1.5 and 300 K, respectively. At 1.5 K, sample A is ferromagnetic with a_s " 15.27 emu/g. But it changes to paramagnetic when the temper­ature reaches 300 K. The DC susceptibility of sample A measured in an applied field of Я " 5000 Oe is shown in Fig. 3, where the curves correspond to both the zero-field cooled (ZFC) and the field cooled (FC), respectively. Namely, in the case of ZFC operation, the sample was cooled first from 300 to 1.5 K without H, then measured with H, while for the case of FC the measurement was made on heating routing after the sample was cooled from 300 to 1.5 K in H. The DC-suscepti­bility curve shows spin-glass behavior. The spin-freezing temperature is about 38 K. To our knowledge, this is the first measurement in the nanostructured MnFeO system. The crystallized sample B is ferromagnetic both at low and room temperature. Its <r_s " 24.4 emu/g at room temper­ature is much lower than the bulk value (80 emu/g). The reduced saturation magnetization of MnFeO particles was also found by Chen and Tang et al. [3,9].

It was explained by a magnetic dead layer on the surface of particles. Assuming the thickness of the dead layer (t) is constant, the mag­netization of the particles can be expressed as

where cr_s(oo) " 80 emu/g corresponding to the bulk value at room temperature. Fitting the data with d " 5.6 nm in Eq. (1), we obtained the thickness of the dead layer which is about 0.6 nm. It is the same as what Chen et al. [9] found. So our data give a further support to the dead-layer theory.

Fig. 4a shows the curve of saturation magneti­zation versus temperature from 300 to 1173 K for annealed sample B. The magnetization decreases with increasing temperature and reaches zero at about 733 K, corresponding to the Curie temper­ature. When continuously heated, the magneti­zation begins to increase and reaches the highest value at 1050 K, then decreases to zero. As for the peak in the high-temperature range, it is due to the segregation of iron. Our sample is covered by a layer of surfactant, the carbonization of the sur­factant at high temperature can reduce the iron from the MnFeO nanoparticles. The sample of nanosized MnFe₂O₄ particles without surfactant covering did not show the peak in o-T curves as shown in Fig. 4b. We measured the high-angle XRD of the sample B after measuring the a-T, which was shown in Fig. 5. It fits to a-Fe well.

Fig. 4. Magnetization a versus temperature for annealed MnFeO particles sample (a) and MnFeO particles without surfactant covering (b).

The measured T_c is about 160 K higher than that of bulk material. Tang et al. [3] and Kulkarni et al. [4] also obtained 97 and 47 K rise in Curie temper­ature of the MnFeO nanoparticles for d " 7.5 and 12 nm, respectively. They described it by the finite-size-scaling formula [11] [T_c(d) - T_e(ao)]T;\ao) = + (d/d_oy\ where T_c(oo) " 573 K corresponds to the bulk Curie temperature, X is related to a correlation length exponent and d an order of the character­istic microscopic dimension, respectively. Accord­ing to the fitting given by Tang et al., X " 1.42 $ 0.13 and d " 2.0 nm are deduced. The mean diameter of particles in our MnFe₂O₄ sample is 5.6 nm, substituting d " 5.6 nm into Eq. (2) and using the above quantities of X and d, the Curie temperature of our sample should be about 152 К higher than the bulk materials. It is close to our experimental data, 160 K. Fig. 6 gives the plot of T_c versus the size of the particles, in which the data of Tang et al. [3], Kulkarni et al. [4] and our data are all put together. The solid line is fitted by Eq. (2) with X " 1.42 $ 0.13 and d " 2.0 nm. It seems our data can be fitted well by this formula. However, van der Zaag et al. [5,7] and Brabers [6] suggested that the increase in ordering temperature in this system is more likely to have its source in a redis­tribution of the iron and manganese cations over the tetrahedral (A) and octahedral (B) crystallo-graphic sites of the spinel structure. More recent research work done by different groups [7-10] provided more evidence on this point.

Fig. 5. X-ray-diffraction   pattern   of  annealed   sample   after measuring o-T.

Morrish et al. [8] measured Mossbauer spectra of MnFeO with various particle sizes and found the cation distribution on the A and B sites depends on the particle size. They thought the increased order­ing temperature is the result of an increased iron occupancy on the A sites. In addition, van der Zaag et al. [7] pointed out that oxidation of Mn²⁺ to References Mn³⁺ during the determination of T_c also plays a role. This has been further confirmed by Gillot et al. [10], where thermogravimetry and FTIR spec-troscopy studies have demonstrated that a change in the cation distribution takes place at x + 1.50 in MnFeO fine powders. As the annealing time of our sample is long enough to ensure an equili­brium cation distribution, the 160 K rise in T_c of the annealed MnFeO particles may also be caused by the cation distribution. Further studies should be done in future.

In summary, we prepared nanosized MnFeO particles by chemical ultrasonic emulsion method. The as-prepared sample shows spin-glass behavior at low temperature. The Curie temperature of the annealed sample is 160 K higher than that of the bulk material. The finite-size effect and the cation redistribution during heating process may both play a role in it.

Acknowledgements

This work was supported by National Nature and Science Foundation and partially by Chinese Academy of Sciences.

[1] M. Kishimoto, Y. Sakurai, T. Ajima, J. Appl. Phys. 76 (1994) 7506.

[2] J.L. Dorman, D. Fiorani (Eds.), Magnetic Properties of Fine Particles, North-Holland, Amsterdam, 1992.

[3] Z.X. Tang, C.M. Sorensen, K.J. Klabunde, G.C. Hadjipanayis, Phys. Rev. Lett. 67 (1991) 3602.

[4] G.U. Kulkarni, K.R. Kannan, T. Arunarkavalli, C.N.R. Rao, Phys. Rev. B 49 (1994) 724.

[5] P.J. van der Zaag, A. Noordermeer, M.T. Johnson, P.F. Bongers, Phys. Rev. Lett. 68 (1992) 3112.

[6] V.A.M. Brabers, Phys. Rev. Lett. 68 (1992) 3113.

[7] P.J. van der Zaag, V.A.M. Brabers, M.T. Johnson, A. Noordermeer, P.F. Bongers, Phys. Rev. B 51 (1995) 12009.

[8] A.H.  Morrish,  Z.W.  Li,  X.Z.  Zhou,  J.  Phys.  (Paris) С 1 (1997) 513.

[9] J.P. Chen, C.M. Sorensen, K.J. Klabune, G.C. Hadjipanayis, E. Devlin, A. Kostikas, Phys. Rev. B 54 (1996)

[10] B. Gillot, M. Laarj, S. Kacim, J. Mater. Chem. 7 (1997) 827.

[11] M.N. Barber, in: С Domb, J.L. Lebowitz (Eds.), Phase Transitions and Critical Phenomena, vol. 8, Academic Press, New York, 1983, p. 145.