Optical micrographs of the as-cast microstructures of samples 1–5 ( f 1⁄4 Nd 2 Fe 14 B). 

Contexts in source publication

Context 1

... of micro-valves, micro- pumps, micro-positioners, etc., because the energy densities of these devices are often too low for the requirements of the application. Magnetic micro-actuators and systems, usually referred to as MAGMAS, are miniature devices that convert electrical energy into mechanical energy, making use of the fields produced by permanent magnets. Such devices could overcome the energy-density problem, provided magnets with thicknesses of 100–500 m m [1] and with properties comparable to bulk magnets can be produced economically. Thin rare-earth transition-metal (RE-TM) magnets can be produced by familiar techniques like pulsed-laser deposition (PLD) [2], sputtering [3], and molecular-beam epitaxy (MBE) [4]. All of these techniques are able to produce magnets with excellent properties; however, PLD and MBE are probably too expensive for any commercial realization in the area of Nd–Fe–B permanent magnets, but sputtering, plasma spraying [5] (which can deposit material at a practical rate of several m m per minute) and ion–plasma sputtering can produce sufficiently thick magnets with good magnetic properties in several hours or less. An excellent review of some routes for preparing thick-film magnets can be found in Dempsey et al. [6]. The most obvious method for producing small, 100–500- m m-thick magnets is to use a straightforward top-down approach and simply cut up large magnets. Unfortunately, though, high-quality RE-TM magnets can suffer surface degradation during micro-machining, resulting in a loss of coercivity and remanence [6], and Hinz et al. [7] report that the brittleness of Nd–Fe–B magnets ‘‘restricts the lower limit of thickness by machining the bulk specimens to some hundreds of micrometers’’, which eliminates the possibility of producing 100- m m-thick samples using such a process. The same authors, in their own experiments, produced Nd–Fe–B samples from Magnequench’s MQU-F powder with thicknesses in the range 300–600 m m using their one- stage ‘‘extreme’’ die-upsetting process. In comparison, their thinner samples, down to 150 m m, produced using a double die-upsetting process, had diminished coercivities. Other techniques that have been applied with varying degrees of success include screen printing [8], bonding, and injection molding. The ideal technique for a MAGMAS magnet would combine a simple, low-cost process with good magnetic properties. For this reason, we have looked at attempting to sinter materials with sufficient amounts of rare earth to overcome the problems of evaporation that are always associated with sintering small samples at temperatures above 1000 1 C. Whereas, previously reported sintering-type techniques have involved complex processing techniques like two-stage die upsetting while applying direct Joule heating to produce 200- m m-thick samples [9] or a special granulation process [10], we have chosen to apply a simple process and modify the starting material to produce high- coercivity RE-TM magnets with thicknesses down to 100 m m. The five alloys used in this series of experiments were produced by induction melting appropriate amounts of neodymium (99.9%), iron (99.9%) and standard Fe 2 B ferro-boron, and pouring each molten alloy into a copper book-type mold with dimensions of 40 mm  250 mm  400 mm. Each 15-kg melt was poured at 1530 1 C. Small samples from close to the edge of each ingot were mounted and prepared for metallographic examination. Approxi- mately 5 kg of each ingot was then hydrogen decrepitated at 1-bar overpressure and jet milled in argon to a mean particle size of 5 m m. The chemical compositions of the five alloys are shown in Table 1. The powders were stored under argon prior to pressing in parallel and perpendicular presses: the latter with a transverse magnetic field of 0.75 T. The amount of powder poured into the pressing die was varied so as to produce samples with thicknesses in the range 100 m m–5 mm. The pressure applied to form the green compacts was 390 MPa. The green compacts were heated in a vacuum of approximately 10 À 5 mbar at 5 1 C min À 1 to temperatures in the range 680–960 1 C for 1 h, after which they were allowed to cool in the furnace to room temperature. The initial cooling rate was limited to 10 1 C min À 1 . Each sintering experiment involved 3–5 pressed green samples that were placed in a tantalum sample tube, to prevent contact with the silica-glass furnace tube, and located precisely in the hot zone of the furnace. After the sintering treatment in the furnace the samples were cut with a diamond-wire cutter so as to have sides of approximately 250 m m  250 m m and then measured at room temperature using a vibrating-sample magnetometer (VSM) with a maximum field of 1.8 T to provide information on their magnetic properties. The chemical analyses of the phases in the as-cast alloys were carried out with a JEOL 840A scanning electron microscope. The microstructures of the five cast alloys are shown in Fig. 1. The various phases present in the alloys were analyzed using energy-dispersive X-ray diffraction and wavelength-dispersive X-ray diffraction. The phases are identified in the figure. It is clear that with increasing neodymium concentration in the as-cast alloys there is a pronounced increase in the amount of Nd-rich material, and due to the richness of the starting compositions in terms of neodymium all the alloys were in the two-phase region of the phase diagram, where f (Nd 2 Fe 14 B) is in equilibrium with the liquid; this is in contrast to conventional Nd–Fe–B magnet compositions, where a three-phase region exists and f is in equilibrium with the liquid and Z (Nd 1.1 Fe 4 B 4 ) [11]. The proportion of coarse particles of the primary f tetragonal phase is reduced with the extra neodymium, so that there is substantially more Nd-rich matrix already in alloy 2. Within this matrix there is also an increased share of binary eutectic (Nd+ f ). These same features are also clearly present in alloy 3, with 30.9% Nd. The higher neodymium concentration in the alloys 4 and 5 and the non-equilibrium solidification result in increasingly smaller amounts of f phase and separated solidification of the Nd-rich residual melt. The consequence of this is the presence of primary crystals of neodymium and a binary eutectic (Nd+ f ). Fig. 2 shows a graph of the effect of the sintering temperature on the intrinsic coercivity of the samples produced in the range 680–960 1 C. Each sample was produced from 0.8 g of material and pressed in a die with a diameter of 6 mm. The thicknesses of the samples that fully densified were approximately 5 mm. It is clear that the samples produced from alloys 1 and 2 had very poor properties, and that these compositions, which are relatively close to the classical Neomax composition of Nd 15 Fe 77 B 8 , require much higher temperatures to densify and develop a high coercivity. In contrast, the samples from alloys 3–5 exhibited excellent coercivities of 800–1000 kA m À 1 in the range 800–960 1 C. The plateau in the graph between 800 and 960 1 C is, however, a little deceptive: although the measured magnetic properties of the samples remained high, in terms of coercivity, the physical characteristics deteriorated rapidly with the samples beginning to deform and melt at temperatures above 840 1 C (alloy 5), 880 1 C (alloy 4) and 960 1 C (alloy 3). At lower temperatures, i.e., below 750–800 1 C, the coercivity of the samples from alloys 3 and 4 fell off rapidly as the materials failed to densify completely. Samples from alloy 5, however, retained a coercivity of 800 kA m À 1 , at least down to 680 1 C. The fall off in coercivity is not a direct consequence of the lack of density. However, the very high ratio of the surface area to the volume means that evaporation rates for the neodymium are high, much higher than with a conventional bulk magnet, and this tends to lead to a shortage of rare-earth-rich phase, which is required to encompass the individual grains and fully densify the sample. The consequence of this can be a porous structure with a less-than-perfect coating of the grains and a tendency to form small oxides and defects on the surfaces of the Nd 2 Fe 14 B grains, thus leading to a lower coercivity. The results of the measurements of the same samples’ remanent magnetization after magnetizing in a field of 1.8 T are shown in Fig. 3. The curves are similar to the coercivity curves in Fig. 2; this is not surprising as the same inability to densify at low temperatures that reduced the coercivity of the samples from alloys 3 and 4 also results in a decrease in the remanent magnetization of the samples. Fig. 4 shows the microstructures from samples of alloys 3 (Fig. 4a) and 4 (Fig. 4b) that were sintered at 800 1 C. There is clear evidence of porosity and a lack of full density in Fig. 4a. On the basis of the results shown in Figs. 2–4 we chose a sintering temperature of 800 1 C and produced samples from alloys 3–5 with thicknesses in the range 100 m m–5 mm. This thickness variation was achieved by altering the amount of powder added to the die between 20 mg and 0.8 g. The results of the measurements of the intrinsic coercivity for different sample thicknesses are shown in Fig. 5. It is clear that all three alloys can produce magnets with good properties when the samples are relatively thick, i.e., 1 mm or more; however, as the samples are made thinner, below 0.5 mm, the coercivities of the samples from alloy 3 decline rapidly, while those of alloy 4 remain constant, and those from alloy 5 were observed to increase slightly. Having determined that we required a composition in the range Nd 41.0 Fe 50.4 B 8.6 –Nd 51.0 Fe 40.7 B 8.3 in order to obtain high coercivities for 100- m m-thick samples, we used a perpendicular-alignment press and a field of 0.75 T to produce an aligned magnet from alloy 4 powder. The sample was pressed as the magnetic field was applied and then sintered in ...

Context 2

... microstructures of the five cast alloys are shown in Fig. 1. The various phases present in the alloys were analyzed using energy-dispersive X-ray diffraction and wavelength-dispersive X-ray diffraction. The phases are identified in the figure. It is clear that with increasing neodymium concentration in the as-cast alloys there is a pronounced increase in the amount of Nd-rich material, and due ...