Characterization of ferrites synthesized by mechanical alloying and soft chemistry

Characterization of ferrites synthesized by mechanical alloying and soft chemistry

NanoStmctared Materials, Vol. 12, pp. 641-644, 1999 Elsevier Science Ltd Q 1999 Acta Metallurgica Inc. Printed in the USA. All rights resewed 09659773...

302KB Sizes 0 Downloads 96 Views

NanoStmctared Materials, Vol. 12, pp. 641-644, 1999 Elsevier Science Ltd Q 1999 Acta Metallurgica Inc. Printed in the USA. All rights resewed 09659773/99/$+x front matter

Pergamon

PI1 so9f5.9773(99)00206-8

CHARACTERIZATION OF FERRITES SYTVTHESIZEDBY MECHANICAL ALLOYING AND SOFT CHEMISTRY N.

Millet’,

S. Begin Colid, P. Perriat’, R. Welteti, B. Malaman3

G. te

Caik*,

(1) L.R.R.S., U.M.R. 5613, BP 400, 21011 Dijon Cedex, FRANCE. (2) L.S.G.2M., U.M.R. 7584, Ecole des Mines, 54042 Nancy Cedex, FRANCE. (3) L.C.S.M., U.M.R. 7555, Universitb H.Poincarb, 54506 Vandoeuvre les Nancy, Abstract --For the purpose of comparing nanostructured materials prepared by dflerent synthesis m’ethodr, nanocrystalline Fe-based spinels were synthesized using two d@erent routes: soJi’ chemistry and high-energy ball milling. The as-prepared powders were characterized notably by thermogravimetric analyses and “Fe MZissbauer spectrometry. 01999 Acta Metallurgica Inc. INTRODUCTION

In the spine1 structure of the titanomagnetite, Fez.sTio 504, investigated in the present work, tetrahedral (A) and octahedral (B) sites are occupied by the cations (1). When the fenite contains some substitutional M cations, the cation distribution is not as simple as it is in magnetite at room and at low temperatures where all the Fe2+ions are on B sites. In the case of titanium ferrite, the distribution of cations has been extensively studied (2,3,4,5). All the authors have found the Ti4+ cations residing on octahedral sites but the location of the Fe” cations is still controversial even when results are derived from the same technique. Moreover a change of the cation to anion ratio is, possible in the spine1 phase leading to a deviation 6 from stoech:iometry : (Fe2.sTi&s04. In any case the dispute may originate f?om differences in several experimental conditions and control parameters such as temperature, grain size, deviation from stoichiometry, method of synthesis... To compare samples prepared by different methods, we have selected two routes, soft chemistry and mechanosynthesis, which lead to powder particles with nanometer-sized grains. Among the numerous methods of soft chemistry, the precipitation has been chosen (6,7). The second synthesis route is mechanosyntbesis (8,9,10) which is a high energy ball milling process. The structure and properties of nanocrystalline ferrites are characterized mainly by Mbssbauer spectrometry and by studing their oxidation behavior.

EXPERIMENTAL

DETAILS

The processes of soft chemistry are described more precisely elsewhere (11). It involves two steps : precipitation in an aqueous solution followed by thermal annealing under 641

642

FOURTH INTERNATIONAL CONFERENCEON NANOSTRUCTURED MATERIALS

a reducing mixture of NJHJH~O gases. The difFerem treatments eliminate remaining impurities and allow to reach a stoichiometric state (S=O) (12). Continuous grinding is performed in a planetary ball mill (Fritsch Pulverisette 7) with a powder to ball weight ratio R=1/40. The grinding tools are made of steel (Fe-13%Cr), the volume of the vial is approximately 45 cm3 and seven balls of diameter 0 = 13 mm am used. The powders mixture (Fe, a-Fez03, TiOz) in stoichiometric proportions and balls ate introduced in vials and sealed in a glove box under argon atmosphere. The resulting powders are all characterized by 57Fe Miissbauer spectrometry with a constant acceleration spectrometer and a s7CoRh source. Their oxidation behavior is followed by thermogravimetry (SERARAM TAG 24).

RESULTS

AND DISCUSSION

The mechanosynthesized spine1 phase appears after lh30 of grinding and all peaks on the XRD pattern are assigned to the spine1 phase after 4 hours (10). Ground powders consist of aggregates of nanometric crystallites. The average size of aggregates is approximately 100-200 nm and the size of small crystallites in aggregates is about 15 nm as also confirmed by transmission electron microscopy. In contrast, powders prepared by soft chemistry are very well dispersed and homogeneous with a grain size of about 18 mn. The oxidation behavior of both powders has been studied by thermogravimetry. The DTG curve displays one single peak for the powder obtained by soft chemistry (figure la) which can be attributed to the oxidation of Fe’+ only located in octahedral sites. For mechanosynthesized powders (figure lb), two peaks are observed which means that some Fe’+ cations are in A sites. The oxidation in the spine1 phase leads to a weight gain of 4.1% for the powder obtained by soft chemistry and of 4.64% for the ball-milled powder. As the theoretical weight gain for a stoechiometric ferrite, 5.27%, is larger than the previous values, it means that there is already a significant oxidation of the Fe”cations at ambient temperature. From such studies, we deduce that 6, in (Fe2,5Tio.s)l_s04,is about 0.040 and 0.022 in ” soft chemistry ” powders and in ground powders respectively. (4

Temperature (“C) Figure 1 : DTG oxydation curves dArn/dt = f(T) of Fe,,Ti,,.,O, synthesized chemistry, b) mechanosynthesis (4 hours, R=1/40, under argon)

by a) soft

FOURTH INTERNATIONAL CONFERENCEON NANOSTRUCTURED MATERIALS

643

This difference may be related to the agglomeration state of ground powders which limits the oxidation process at ambient temperature. The reactivity of the ground powder is smaller than that of the sa.mple obtained by soft chemistry route. Some Mijssbauer spectra of the titanomagnetites synthesized by mechanical alloying and by soft chemistry are shown in &ure 2. Both series of spectra are very different and very complex. Some information can nevertheless be extracted from undisputed features of the observed spectra. The spectrum of the spine1 phase synthesized by soft chemistry (figure 2a) corresponds to the Miissbauer spectrum expected for a ferrimagnetic polycrystalline material with the superposition of more than two magnetic hyperfiie sextets (13-14). Clear differences exist however with the spectrum published in litterature for a titanoferrite of the same composition (14). In particular, the hyperfine field of the outer sextet is much larger in our case H = 486 kG than it is for the titanomagnetite Fez.sTio5O4investigated by Melzer et al. H =447 kG (14). The outer sextet that we observe here cannot be associated with a titanomagnetite whose Ti content is x=0.50 even when taking into account the existence of deviations Ii-om stoechiometry. From the Ti content dependence of the hyperfime field, it is indeed mainly associated with a titanoferrite whose Ti content is less than at most x=0.10. The aforementioned Miissbauer characteristics are therefore accounted for by chemical composition variations inside each particle. The heterogeneities are then rather related to kinetic effects linked to partial oxidation of the iron cations occcuring when nanoparticles are studied in room conditions. Therefore whereas near the surface there are mainly iron Fe3’ cations, in the bulk of the particle the Fe” and titanium cations are predominant. The measured field H is indeed consistent with a FezOj surface layer. Even if other Fe ions environments with a distribution of iron and titanium neighboring atoms contribute also to the observed spectrum, it is mainly in agreement with particles with a Ti-rich core and a Fe-rich shell. Miissbauer spectra of the mechanosynthesized titanomagnetite (figure 2b) are similarly difficult to analyze. The shape of the spectrum recorded at room temperature with four external lines, with a characteristic field of about 420 kG, and an intense asymmetric central doublet suggests that superparamagnetic relaxation takes place in the sample (15-16). The assumption of superparamagnetic relaxation, which is consistent with an average crystallite size of 15 nm, is confirmed by spectra recorded at 200 and 100 K (11). They show a progressive increase of the magnetically split part of the spectra when the temperature decreases. The latter increase is related to a distribution of grain sizes, classilcally broad in ground powders, which results in a distribution of blocking temperatures.

Figure 2 : 57Feroom-temperature Miissbauer spectra of Fez 5Tio.504synthesized a) by soft chemistry, b) by high-energy ball milling.

644

FOURTH INTERNATIONAL CONFERENCEON NANOSTRUCTURED MATERIALS

Further, grain growth is induced by heat treatments under an argon atmosphere of ground powders at 1100K. The resulting Miissbauer spectrum is then similar to that of polycrystalline ferrites confirming the effect of grain size in nanoferrites [ 111.The outer hyperfine field (H= 448 kG) corresponds to a titanomagnetite with a titanium content close to 0.5. This result confirms that in mechanosynthesized powders, the heterogneities of composition are not as important as they are in those obtained by soft chemistry.

CONCLUSIONS In ferrites obtained by mechanosynthesis and by soft chemistry, the average crystallite size is about 15 nm, but whereas the ball-milled powders consist of aggregates, those obtained by soft chemistry are formed from isolated nanoparticles. Comparison of both ferrites show that oxidation phenomena occur thus at the surface of grains in the soft chemistry powders and lead to materials with a higher deviation from oxygen stoichiometry 6 = 0.040 in (Fe* 5Ti0.5)1_804 than those formed from mechanical alloying 6 = 0.022. Mossbauer spectra agree with the previous conclusions on the differences in oxidation behaviour and in crystallite size distribution. They demonstrate that heterogeneities of titanium content exist in soft chemistry powders while a superparamagnetic behaviour is evidenced in mechanosynthesized fenites at room temperature.

ACKNOWLEDGMENTS The authors would like to thank P. Delcroix, Dr. F. Bernard and Dr. D. Aymes fbr their help in the experiments, Dr. B. Gillot for helpful discussions.

REFERENCES (1) (2) (3) (4)

E.J. Verwey, E.L. Heilman, J.Chem.Phys., 1947, 15, 174. S. Akimoto, .J. Geomagn. Geoelec., 1954,61, 1. L. NCel, A&. Phys., 19554, 191. R. Chevalier, J. Bolfa, S. Mathiew, Bull. Sot. Franc Min. Cris., 1955, 78, 307. (5) A. Trestman, S.E. Dorris, S. Kumarakrishnan, J. Am. Ceram. Sot., 1983, 66, 829. (6) J.P. Jolivet, De la solution a l’oxyde, InterEditions/CNRS Editions, Paris, (1994). (7) A. Rousset, F. Chassagneux, J. Paris, J. Mater. Sci., 1986, 21, 3 111. (8) C. C. Koch, Annu. Rev. Mater. Sci., 1989, 19, 121. (9) V. V. Boldyrev, N. Z. Lyakhov, Pavlyukhin, E. V. Boldyreva, E. Y. Ivanov, E. G. Avvakumov, Sov. Sci. Rev. B. Chem., 1990,14, 105. (10) J. J. De Barbadillo, Key Engineering Materials, 1993, 77-78, 187. (11) N. Millot, S. Begin, P. Perriat, G. Le CatSr,J. Solid State Chem, 1998 in press. (12) D. Aymes, N. Millot, V. Nivoix, P. Perriat, Solid State Ionics, 1997, 101-103, 261. (13) H. Tanaka, M. Kono, J. Geomagn. Geoelect., 1987,39,463. (14) K. Melzer, Z. Simsa, M. Lukaslak, J. Suwalski, Cryst. Res. Technol., 1987, 22, 132. (15) V. U. Patil, R. G. Kulkarni, Solid State Comm., 1979, 31, 551. (16) E. Tronc, J. P. Jolivet, Materials Science Forum, 1997, 235-238, 659.