IEEE TRANSACTIONS ON MAGNETICS, VOL. 38, NO. 5, SEPTEMBER 2002
1907
Magnetic Properties of Acicular Ultrafine Iron Particles Laudemir C. Varanda, Gerardo F. Goya, Maria P. Morales, Rodrigo F. C. Marques, Ricardo H. M. Godoi, Miguel Jafelicci, Jr., and Carlos J. Serna
Abstract—Magnetic properties of acicular ( 60 and 200 nm) iron particles, obtained by reduction of alumina-coated goethite particles, are reported. X-ray diffraction and Mössbauer spectroscopy showed that the particles consist of a -Fe core and a thin surface layer of maghemite. Magnetization data indicated an improvement of 28 in the saturation magnetization, coercive field, and squareness for particles with 60 nm. This magnetic property enhancement of the present particles, whose size is 40% smaller than those commercially available, could result in a similar decrease of the bit-size for higher density of magnetic media.
%
Index Terms—Coercivity, magnetic recording, magnetic variable measurements, metal ultrafine particles, passivation.
I. INTRODUCTION
T
HE introduction of high densities on magnetic recording media has meant a dramatic bit-size reduction and track density increase on the media surface. In order to enhance the density, recording media values as close as possible to the superparamagnetic limit, both particle size reduction and narrow size distribution are required [1]. However, the magnetic properties of the particles for magnetic recording, mainly coercivity and saturation magnetization, change with the particle size decreasing [2]. Thus, the signal-to-noise ratio and other questions about the stability of increasingly small written bits have been investigated [3]–[6]. Nowadays, advanced flexible media consist almost exclusively of metal particles (MP) based on acicular Fe or Fe alloys, with length on the order of 100 nm, and axial ratio 5 [1]–[6]. The only commercially significant process today for the production of the iron metal particles is the reduction of acicular oxyhydroxide particles followed by controlled particle surface oxidation to protect them [7], [8]. In this paper, monodispersed goethite particles (spindle type of 60-nm length and rodlike of 230-nm length), both with high axial ratios, were coated with alumina, reduced to iron, and stabilized by an oxide layer on the particle surface. The aim of this paper is to investigate the magnetic properties of the ultrafine Manuscript received February 19, 2002; revised May 28, 2002. The work of L. C. Varanda was supported by the Brazilian agency FAPESP. The work of G. F. Goya was supported by the Brazilian agencies FAPESP (01/02598-3) and CNPq (300569/00-9). L. C. Varanda, R. F. C. Marques, and M. Jafelicci, Jr. are with the Instituto de Química de Araraquara, UNESP, Araraquara 14801-970 SP, Brazil (e-mail:
[email protected]). G. F. Goya is with the Instituto de Física, University São Paulo, São Paulo 05315-970 SP, Brazil. M. P. Morales and C. J. Serna are with the Instituto de Ciencia de Materiales de Madrid, 28029 Madrid, Spain. R. H. M. Godoi is with the University of Antwerp, B-2610 Antwerp, Belgium. Digital Object Identifier 10.1109/TMAG.2002.802816.
iron particles 60 mn compared to those of larger particle size 200 nm . A decrease in the bit-size and an increase are expected in the recording density, proportional to the particle size decrease. II. EXPERIMENTAL PROCEDURE A new synthetic carbonate Fe(III)-based route, described earlier, was used in the preparation of the both goethite particle systems [9]. To prevent the particle sintering during the reduction, the goethite particles were coated with alumina, Al/(Al Fe [2]. Coated particles were dehydroxylated at 400 C for 3 h to carry out transformation to Al O -coated -Fe O , which were washed several times to eliminate impurities. Reduction of the Al O -coated -Fe O was carried out in a furnace by heating at 450 C for 10 h under hydrogen flow. After the end of the reduction process, the samples were cooled to room temperature under the hydrogen atmosphere. Finally, nitrogen gas was blown into a flask containing ethanol and the resulting stream was introduced in the furnace to provide slow oxidation on the particle surface. Morphology, particle size and size distribution were investigated by transmission electron microscopy (TEM) using a Phillips CM200 microscope. The samples were dispersed in toluene, and then a drop of this suspension was deposited onto a carbon-coated copper grid. The average particle size and standard deviation were calculated by counting around 200 particles. The present phases in the samples were identified by X-ray diffraction (XRD) using a Siemens D5000 diffractometer and Cu–K radiation, and the average crystallite size was calculated from the (110) iron reflection using the Scherrer equation [10]. Fe Mössbauer spectroscopy (MS) was performed between 4.2 K and 300 K, using a Co source contained in an Rh matrix and a conventional constant-accelerator spectrometer. All isomer shifts are given relative to that -Fe at room temperature. Magnetization measurements were and carried out in a commercial SQUID. Coercivity field values were obtained from the saturation magnetization hysteresis loops at room temperature with a maximum applied field of 50 kOe. III. RESULTS AND DISCUSSION The goethite precursor particles used were: 1) spindle shaped nm length and an axial ratio of 6 (identified as with nm length sample Fe60) and 2) rodlike shaped, with and an axial ratio of 10 (identified as sample Fe230), and both samples were monodispersed systems [9]. These particles were
0018-9464/02$17.00 © 2002 IEEE
1908
IEEE TRANSACTIONS ON MAGNETICS, VOL. 38, NO. 5, SEPTEMBER 2002
T
Fig. 2. Representative MS at = 300 and 4.2 K for reduced and nonreduced particles. Solid circles correspond to the experimental data, and solid lines to each component of the fitted spectrum.
Fig. 1. TEM images of the iron particle samples. (a) Fe60. (b) Fe230.
coated with a thin alumina layer in order to prevent sintering and acicular shape loss during the reduction process. Typical TEM images after the reduction in hydrogen flow at 450 C for 10 h are shown in Fig. 1. After reduction, the particle sizes, estimated by TEM for sample Fe60 [Fig. 1(a)] and and nm Fe230 [Fig. 1(b)] were, respectively, in length, showing a decrease between 10%–15% in the particle size during the reduction process. The XRD showed two phases: first, two reflections from the MP core identified as -Fe, and second, a small broad peak corresponding to the oxide passivation layer, assigned to a spinel structure as expected for MP [4]–[8], [11]. The average crystallite size obtained from (110) nm for sample Fe230 and nm reflection was for sample Fe60. These values, smaller than the size calculated from TEM, suggest that the iron core for sample Fe230 consist of about six crystallites, whereas for sample Fe60, it is formed by 3 crystal units. A room-temperature Mössbauer spectra (MS) of both precursor samples showed the coexistence of a central doublet and a magnetic sextet (Fig. 2). The central doublet is originated in relaxing and superparamagnetic (SPM) fractions of the sample, as confirmed by the disappearance of this doublet at 4.2 K in both samples. The broad lines of the magnetic sextet at room temperature indicate the existence of magnetic hyperfine field distribution due to poor crystallinity of the sample [12], in agreement with XRD results [9]. At 4.2 K, the hyperfine parameters showed the presence of goethite phase before reduction T, and by a magnetic sextet (hyperfine field, mm/s. A second sextet ( isomer shift, T and mm/s), corresponding to hematite phase was observed, amounting 10% of the total resonant
Fig. 3. Low field region of the hysteresis loops at room temperature for samples Fe60 and Fe230. The inset shows the magnetization curves up to = 50 kOe.
H
area. This small amount of hematite phase present in the samples was probably formed on the goethite particle surface during the synthesis. Due to the small particle size, poor crystallinity, and small amount, it was not possible to identify this phase by XRD [9]. K For reduced Fe60 and Fe230 samples, MS at K and shows the magnetic sextet with mm/s, indicating the presence of a -Fe phase. The central doublet observed in both spectra was assigned to the oxide phase in an SPM state. This behavior is consistent with the iron atoms associated in a very thin layer or as small particles. Thus, the presence of the doublet at room temperature can be identified with the surface passivation oxide layer. There is not evidence of any other component in the spectra. At 4.2 K, this cenT and tral doublet is magnetically split, having mm/s, close to the expected values for Fe in maghemite phase -Fe O . These results, in agreement with
VARANDA et al.: MAGNETIC PROPERTIES OF ACICULAR ULTRAFINE IRON PARTICLES
TABLE I MAGNETIC PARAMETERS: SATURATION MAGNETIZATION ( ), COERCIVE FIELD ( ), REMANENCE ( ) AND SQUARNESS (S) OBTAINED FROM THE HYSTERISIS LOOPS AT = 300 K
H
M
T
M
the small broad peak observed by XRD [11], indicate that the formation of a surface layer of maghemite yields the chemical stability observed in ultrafine iron particles. The volume of the passivation layer was estimated from MS to be 50% of the total particle volume, in agreement with the previous works on advanced particulate recording media [3]. The magnetization hysteresis loops at room temperature obtained for precursor samples presented the SPM behavior in agreement with the MS data. The magnetization curves for the reduced samples are shown in the Fig. 3, and the magnetic parameters obtained from hysteresis loops are presented in the Table I. The values obtained for magnetic parameters increase with the particle size decrease. Typical hysteresis loop for acicular fine iron particles was observed for sample Fe230. All magemu/g , netic parameters, saturation magnetization emu/g , and coercivity remanence magnetization kOe were expected for acicular iron partifield values cles having size larger than 100 nm in length [1]–[8]. are slightly lower than that reported for iron bulk (220 emu/g) due to the presence of alumina and iron oxide passivating layer, which is not contributing to the magnetization. In addition, spincanting effects due to the small particle size 200 nm may be values [13], [14]. also lowering the value probably occurs due For Fe60, the increase in the to decrease of the number and size of crystallites as a consequence of the particle size reduction. Apparently, these effects contribute to a closer behavior of the metallic iron bulk. Also, the squareness decreases as the crystallite increases suggesting incoherent mechanisms for rotation of the magnetic moment [15]. Accordingly, the lowest coercivity value was observed for sample Fe230 with the higher crystallite size (32 nm). Since the values are much less than those expected from shape present anisotropy (considering the axial ratios observed by TEM), it is likely that collective mechanism for incoherent domain rotation such as curling or fanning are operative. Considering Fe–Co particles, the particle size decreasing seems to improve magnetic properties, mainly coercive field [2], [3], [6], in agreement with these results.
1909
IV. CONCLUSION In this paper, an enhancement in the magnetic properties of ultrafine particles by decreasing the particle size below 100 nm was reported. A decrease in the iron particle size down to 60 nm leads to a significant improvement 28 in the main magnetic properties, such as saturation, coercive field, and squareness. Alloys with up 30% of the cobalt preparation can increase these improved magnetic properties, without promoting significant increase in the particle size. The significance of concurrently improving the magnetic parameters in the fine particles and obtaining the smallest practical size can hardly be overstated, as such allied phenomenon results in much lower bit sizes, thus increasing the recording density.
REFERENCES [1] K. O’Grady and H. Laidler, “The limits to magnetic recording—Media considerations,” J. Magn. Magn. Mater., vol. 200, pp. 616–633, 1999. [2] S. Hisano, K. Saito, S. Aizawa, K. Sano, K. Matsumoto, and K. Murata, “Ferromagnetic Metal Powder,” U.S. Patent 5.591.535, 1997. [3] M. P. Sharrock, “Recent advances in metal particulate recording media: Toward the ultimate particles,” IEEE Trans. Magn., vol. 36, pp. 2420–2425, Sept. 2000. [4] K. Okamoto, Y. Okazaki, N. Nagai, and S. Uedaira, “Advanced metal particles technologies for magnetic tapes,” J. Magn. Magn. Mater., vol. 155, pp. 60–66, 1996. [5] K. O’Grady, R. L. White, and P. J. Grundy, “Whither magnetic recording,” J. Magn. Magn. Mater., vol. 177–181, pp. 886–891, 1998. [6] S. Hisano and K. Saito, “Research and development of metal powder for magnetic recording,” J. Magn. Magn. Mater., vol. 190, pp. 371–381, 1998. [7] N. Sugita, M. Mackawa, Y. Ohta, K. Okinaka, and N. Nagai, “Advances in fine magnetic particles for high-density recording,” IEEE Trans. Magn., vol. 31, pp. 2854–2858, Nov. 1995. [8] M. Kishimoto, T. Nakazumi, N. Otani, and T. Sueyoshi, “Preparation and magnetic-properties of Fe–Co fine particles,” IEEE Trans. Magn., vol. 27, pp. 4645–4647, Nov. 1991. [9] L. C. Varanda, M. P. Morales, M. Jafelicci, Jr., and C. J. Serna, “Monodispersed spindle-type goethite nanoparticles from Fe(III) solutions,” J. Mater. Chem., submitted for publication. [10] L. D. Azároff, Elements of X-Ray Crystallography. New York: McGraw-Hill, 1968, pp. 549–552. [11] L. C. Varanda, M. Jafelicci Jr., and G. F. Goya, “Magnetic properties of spindle-type iron fine particles obtained from hematite,” J. Magn. Magn Mater., vol. 226–230, pp. 1933–1935, 2001. [12] R. M. Cornell and U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses. New York: VCH, 1996, pp. 146–154. [13] M. P. Morales, C. J. Serna, F. Bødker, and S. Mørup, “Spin canting due to structural disorder in maghemite,” J. Phys.: Condens. Matter, vol. 9, pp. 5461–5467, Jun. 1997. [14] F. T. Parker, F. E. Spada, T. J. Cox, and A. E. Berkowitz, “Spin-glass behavior of spinel iron-oxide particles on iron metal-particles recording media,” J. Appl. Phys., vol. 77, pp. 5853–5864, 1995. [15] C. D. Mee and E. D. Daniel, Magnetic Recording Technology, 2nd ed. New York: McGraw-Hill, 1995, pp. 3.4–3.27.
