Influence of particle size on core loss in dust cores, for high frequency application
A paper from Tomotsuna Kamijo, Hirofumi Hojo, Hiroyuki Mitani and Shinya Arima reported on a study of the influence of particle size on core losses in soft magnetic composite dust cores.
In an initial set of experiments, an atomised high purity iron powder 300NH was sieved into particle size fractions with mean particle size ranging from 30 µm to 85 µm. Powder samples from each of these fractions were then processed into soft magnetic composite cores with a density of 7.0 g/cm3 and core losses were then measured, over a range of frequencies, at 0.1T magnetic flux density.
As shown in Fig. 1, total core loss was almost independent of powder particle size at 1 and 5 kHz frequencies, but, at higher frequencies, finer mean particle sizes are effective in driving down losses.
The total core losses are sub-divided between eddy current losses and hysteresis losses in Figs. 2 and 3 respectively.
Again, finer mean particles sizes are seen to be effective in reducing eddy current losses.
In earlier reported studies with higher magnetisation conditions (1.3T), coarser powder sizes were observed to reduce coercive force and hence reduce hysteresis losses and, therefore, there was an optimum particle size for minimum total core losses. At the lower magnetisation conditions (0.1T) reported here, however, the movement distance of a magnetic domain wall is estimated to be much shorter and therefore particle size has minimal influence on hysteresis losses.
In terms of magnetic properties, therefore, it would be concluded that the finer the particle size, the better.
However, particle size choice is also influenced by other considerations, such as compressibility, apparent density and flow rate, all of which deteriorate at finer particle sizes (Fig. 4). The authors have therefore concluded that a mean particle size of around 50µm probably represents the optimum compromise between low core loss and acceptable processibility.
It can be seen from Figs 2 and 3 that generally hysteresis loss accounts for more than 80% of total core loss. A need to find means of reducing hysteresis loss was therefore identified.
To achieve this objective, a second batch of powder samples was processed into soft magnetic composite cores and assessed, in terms of magnetic properties, in a similar manner to the initial batch.
For these samples, the 300NH powder was annealed three times at 970°C for 1.5 hours and was then sieved into fractions with mean particle sizes ranging from 30µm to 71 µm. These samples were then processed, using a combination of warm compaction and die wall lubrication, to achieve a density level of around 7.4 g/cm3.
This processing route was found to reduce hysteresis loss by around 50%, compared with the initial powder batches (Fig. 5). The two sets of powder batches were observed to deliver almost identical results in terms of the relationship between eddy current loss and mean particle size.
Overall, the authors have therefore concluded that the triple annealed powder with around 50 µm mean particle size processed into a 7.4 g/cm3 soft magnetic composite is the most appropriate magnetic powder product and that the development of this outstanding low core loss product is expected to stimulate an expansion of the market penetration for SMCs.
Influence of oxidisation depth on strength of soft magnetic iron core for electric motors
A second paper, from Mamoru Hosakawa, Mikako Takeda and Watanu Urushihara, turned attention to another issue that can limit the use of PM soft magnetic cores – the mechanical strength of the iron core, which, if not adequate, can lead to damage or breakage during the production of complex shaped cores.
It has been observed that, during the process of forming the insulating layers between individual powder particles during the production of an iron core, voids and other defects can be created that can act as trigger points for fracture initiation. Therefore finding means of reducing these voids, while maintaining insulation between powder particles, has been deemed critical to increasing mechanical strength of the core.
The authors have investigated the development of a thermal process route to form iron oxide layers to fill the voids between powder particles and therefore enhance transverse rupture strength of the core.
An effective thermal treatment, in this context, was found to comprise lubricant decomposition at 400°C for 2 hours followed by a treatment at 550°C for 30 minutes (Fig. 6). A cross-sectional SEM image of a sample, subjected to this thermal treatment, is shown in Fig. 7. The sample was oxidised to a depth of 0.4 mm from the surface. The decomposition of the lubricant was effective in promoting oxidation to a greater depth within the sample, resulting in increased strength (Fig. 5).
However, oxidation was only observed in the region of the sample surface and it was surmised that strength could be further improved if oxidation depth could be increased. The use of an oxygen release additive agent combined with the lubricant decomposition during the thermal process was therefore investigated. Mannitol or lithium peroxide was added at a level of 0.1 wt% and an improvement of transverse rupture strength to 61-76 MPa was observed. The relationship between oxidation depth and transverse rupture strength is shown in Fig. 8.
Following these studies, the authors are proposing to develop the strengthening technology for consolidated soft magnetic PM products.
 From paper: ‘Influence of particle size on core loss in dust cores, for high frequency application’, Tomotsuna Kamijo, Hirofumi Hojo, Hiroyuki Mitani, Shinya Arima, Kobe Steel, Ltd., Japan. Presented at the PM2012 World PM Congress, Yokohama, October 15-18.
 From paper: ‘Influence of oxidisation depth on strength of soft magnetic iron core for electric motors’, Mamoru Hosakawa, Mikako Takeda and Watanu Urushihara, Kobe Steel, Ltd., Japan. Presented at the PM2012 World PM Congress, Yokohama, October 15-18.
PM2012 World Congress Proceedings
Papers presented at the PM2012 World Congress will be published by the JPMA/JSP&PM in the Congress Proceedings, available in early 2013. Further information will be posted on the conference website.