THE EFFECT OF SILICA ON THE MICROSTRUCTURE OF MnZn FERRITES A. Giles, F. Westendorp To cite this version: A. Giles, F. Westendorp. THE EFFECT OF SILICA ON THE MICROSTRUCTURE OF MnZn FERRITES. Journal de Physique Colloques, 1977, 38 (C1), pp.c1-317-c1-320. <10.1051/jphyscol:1977165>. <jpa-00217025> HAL Id: jpa-00217025 https://hal.archives-ouvertes.fr/jpa-00217025 Submitted on 1 Jan 1977 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
JOURNAL DE PHYSIQUE Colloque Cl, supplkment au no 4, Tome 38, Aoril 1977, page Cl-317 THE EFFECT OF SILICA ON THE MICROSTRUCTURE OF MnZn FERRITES A. D. GILES Mullard Research Laboratories, Redhill, Surrey, U. K. F. F. WESTENDORP Mullard Magnetic Components, Southport, Lancs., U. K. RkurnB. - L'effet de la silice Si02 sur la microstructure d'un ferritc de manganese-zinc a et6 Ctudie aussi bien par addition de teneurs croissantes de poudre fine de silice - jusqu'a 0,75 Zen poids - que par addition de poudres de silice de differentes grosseurs. La position et la forme de la silice dans le materiau ceramique ont CtC Ctudikes en utilisant diverses techniques, en particulier en combinant I'attaque chimique avec la rnicroscopie Clectronique a balayage et I'analyse par microsonde. Abstract. - The effect of silica, SiO2, on the microstructure of a manganese-zinc ferro ferrite has been studied both by the addition of increasing amounts of fine silica flour, up to 0.75 0,{ by weight, and by the addition of different grades of silica. The position and form of the silica in the ceramic has been studied using a variety of techniques in particular combining etching with scanning electron rnicroscopy and electron microprobe analysis. 1. Introduction. - Silica is known to have a strong effect on the microstructure [l] of MnZn ferrites and promotes (( coarse crystallisation i. e. large grains with enclosed pores tend to form. Its effect is often explained by the presence of a grain boundary phase which is silica based and said to be liquid at the firing temperature [2, 31. Evidence for this boundary phase has been obtained using electron microprobe techniques and Auger spectroscopy on freshly fractured surfaces (by Verhoeven (unpublished)). Both these methods were thought to be rather indirect. In the electron microprobe method and at the silica levels of interest, any grain boundary phase is below the resolution of the instrument and in the Auger spectroscopy method the spatial resolution is such that only bulk grain boundary effects can be observed. In addition to its effect on microstructure, it is possible that silica could cause precipitates within grains resulting in a deterioration in the magnetic properties, both directly through domain wall pinning and indirectly through the system of internal stresses set up on cooling from the firing temperature. It was thought that if, by suitable heat treatment, one could alter the form or position of silica-rich regions one might improve the magnetic properties of ferrites with high silica contents. The need for a method to observe the position and form of the silica in this study is of prime importance. 2. Sample batches and preparation. - In a first set of experiments, batches of MnZn ferrite powder were prepared from pure raw materials with increasing additions of fine silica flour as shown in the table below : Batch 62 63 64 65 66 67 68 69 "/, w/w SiO2 addition 0 0.01 0.02 0.04 0.08 0.16 0.32 0.75 Rings from each of these batches were pressed and fired identically in a kiln with computer control of atmosphere and temperature. Successive sets of rings were fired under a variety of conditions. In a second experiment equal amounts of SiO, of different particle size ranging from fine flour to coarse sand were added to the ferrite powder prior to pressing. Rings pressed from these batches were fired at several temperatures. 3. Results. - 3.1 DENSITY. -. Figure 1 gives the sintered density as a function of silica addition for different firing temperatures. It is to be noted that batch 67 (0.16 %, SiO,) sinters to a normal density of 4.6 g ~rn-~ at 1 140 OC ; this is typical of the density normally achieved with undoped materials fired at 1 275 OC. 3.2 MAGNETIC PROPERTIES. - The materials investigated were high power ferrites for which the critical properties are saturation flux density and power loss. Figure 2 gives the power loss density, P,,,, as a function of SiO, addition for different firing temperatures. The losses for high silica batches are always high. At lower firing temperatures small additions of silica are seen to decrease the losses, but as the temperature rises the minimum in P,,, goes to lower silica content, the Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1977165
Cl-3 18 A. D. GILES AND F. F. WESTENDORP... 351. 3 0'2.- - 04 L 06. J 3 6 Sm02 odj~t,on ("i,) FIG. 1. - Sinter density as a function of Si02 addition and firing temperature. 62 64 66 68 70 Botcn number IiII~I 001 002 004 008 016 032 075 SIO, Add~t~on ("l0) FIG. 2. - Power loss density as a function of SiOz addition and firing temperature. minirlium loss found being 50 kw m-3 with the pure sample and high firing temperature. The saturation measured at 800 Am-' increases at first with silica content but falls off at the two highest silica levels in a manner corresponding to the decrease of the density values. 3.3 MICROSTRUCTURE. - Scanning electron micro- graphs at 275 OC FIG. 3. - Scanning electron micrographs of polished and lightly are shown in figure 3. At high firing temperatures the etched samples fired at 1 275 W.
THE EFFECT O F SILICA ON THE M ICROSTRUCTURE O F MnZn FERRITES general trend was that the microstructure changed from a normal fine-grained structure at low silica contents, through increasingly coarse crystallised microstructures at intermediate silica levels, and back to fine grained structures at the two highest silica levels. At the lowest firing temperatures no coarse crystals were observed in any of the batches, the first appearing in batch 67 (0.16 "/,io,) at 1 140 OC although the coarse grains in this case remained < 10 pm. The second experiment on the effect of the particle size of the silica addition confirmed the tendency towards coarse crystallization once the firing temperature was high enough for the silica to diffuse into the microstructure. At 1 275 O C the particle size appears to have no effect except in the casc of the coarse sand addition. These particles diffused into the rnicrostructure giving the three regions observed in the first experiment. Figure 4 shows, immediately round the FIG.4. - Optical micrograph of a sample with coarse sand addition fired a t 1 300 O C. hole where the silica grain once was. fine pore-free grains with thick grain boundaries. Further from the hole the grains become increasingly coarse-crystallised, initially a mixture of fine and coarse grains with some pores on the grain boundaries while further from the hole huge grains occur with few grain boundary pores. Beyond this region normal fine microstructure is observed. 4. Identification of second phases in the microstructures. - A combination of etching, scanning microscopy (SEM) and electron microprobe analysis were the main tools used in this study. The etching of polished and fractured surfaces in a dilute hydrofluoriclnitric acid revealed etch pits and particles in both the intermediate and high silica batches. By the technique of returning to an identified area before and after etching, etch pits were always found to originate from pores. The particles revealed by etching and any Cl-319 grain boundary phases present were too small (< 1 (pm)3) to identify by microprobe analysis except in a few cases in the highest silica batches where they proved to have a high silica content. By etching samples containing more than about 0.1 o/;i SiO, in hot concentrated hydrochloric acid it was possible to remove the ferrite grains to reveal a white skeleton retaining the overall shape of the original piece of ferrite (l). Microprobe analysis and chemical analysis of the skeleton show it to be almost pure silica with traces of the main ferrite constituents. Figure 5 shows an SEM picture of a sample etched for FIG.5. - Scanning electron micrograph of the silica skeleton emerging from a ferrite surface. one minute. The surface of the ferrite can be seen with an intricate structure of interconnecting silica strands emerging from it. The strands appear to be silica which has been trapped at the junction of three grains. Occasionally in samples with high silica content fine membranes can be seen and, particularly in coarse crystallised samples, circular rings of silica are seen resting on the ferrite surface. The grain structure in the silica is seen best using stereo pairs of micrographs. At the lower silica levels the strands are finer and cannot support themselves, giving a fine matted structure, figure 6. The study of the skeleton is more useful if we can see it emerging from the ferrite. We have done this by examining the same point on the surface after succes(l) After N. Meyer, N. V. Philips, Uden.
C1-320 A. D. GILES AND F. F. WESTENDORP were thicker and reached their final thickness of 0.2 -+ 0.4 pm, at the end of the l 275 OC soak when the grain size was 8-12 pm. No physical difference was observed between the samples quenched from the top temperature and those cooled normally, however x-ray diffraction analysis of the two skeletons showed the quenched skeleton to be amorphous while the slow-cooled skeleton was cc crystobalite. An amorphous or even liquid phase at 1 275 OC would be possible in the presence of impurities. On cooling a eutectic would form, the silica part of which is observcd as the skeleton. The formation of cc crystobalite is more difficult to explain. This does not usually form below 1 470 OC, however the conditions of pressure and the influence of the surrounding cubic ferrite crystals on the silica as it cools are unknown. FIG. 6. - Scanning electron micrograph of the collapsed silica skeleton of a sample containing 0.02 % SiOz addition. sive etching periods and indeed the strands appear at the intersection of grains while the unconnected silica rings appear from pores found within coarse crystallised grains. Pores on grain boundaries tend to result in a triangular silica structure. An attempt was made to find out at what temperature the silica appears at the grain boundary. This was done by quenching samples from different points on the 1 275 OC firing cycle and examining them by this technique. At 1 200OC the silica was not visible, at 1 250 O C fine wisps of < 0.1 pm could be seen on grains of 2-3 pm diameter, at 1 275 OC the strands 5. Conclusions.-The effect of silica on the microstructure of manganese zinc ferrites has been demonstrated. Intermediate levels of silica (0.1 -+ 0.15 :/, w/w) promote coarse crystallization and result in deterioration of magnetic properties except at low firing temperatures where some advantage is gained due to their higher sinterability. Samples with the highest silica levels exhibit a finer microstructure but very poor properties. This may be due to a thicker, stiffer grain boundary phase causing stresses and silica rich precipitates within the grains. It has been demonstrated that the silica forms an amorphous grain boundary phase during sintering and that this can be observed by etching in hot concentrated hydrochloric acid. Silica is also present inside grains around closed pores. References [I) BANDO, Y., et al., Proc. 1970 Int. Perm Magnet. Conf. N. Y. 4 (1971) 339-348. 121 KOOY, C., Sci. Ceram. 1 (1961) 21. [3] CARPAY, F. M. A. and Srurl~s, A. L., Sci. Ceram. 8 (1975) 23-38.