Keywords

15.1 Introduction

Soft magnetic materials, such as Fe–Si, Fe–Ni, Fe–Al–Si, and Fe–Co to name some, are widely used in modern electric and electronic devices [1, 2]. They transform electrical to mechanical energy and are indispensable in applications that involve the conversion of the electric current waveform. Therefore, their application includes transformers, inductors, communication equipment such as audio transformers, recording heads and magnetic modulators and are widely used in ordinary appliances such as hair dryers, vacuum cleaners, washing machines, air conditioners, phone (wireless) chargers, and computer power supplies [3].

A soft magnetic material is characterized by very high magnetic permeability, low coercivity, and high magnetic saturation [4]. In practice, it is visualized on a hysteresis loop where the magnetic response shows a quasi-reversible ferromagnetic system of dipoles, whose response is nonlinear and depends on the dipoles’ rotation into the field direction applied [5]. The Fe-50% Ni alloy, also called permalloy, is considered a soft magnet with a relatively high magnetic saturation (~1.6 T), high relative permeability (10,000–55,000), and low eddy current losses [5, 6]. Although it has advantages from the point of view of magnetic properties over other alloys, it is more expensive than Fe-Si.

Currently, soft magnets are fabricated through a cold rolled strip process, and in the case of brittle alloys or hard-to-work, the manufacturing process could be powder injection moulding or powder metallurgy [7, 8]. However, these techniques are limited in terms of the shape of the part that can be produced and therefore the customization of the magnets. For this reason, additive manufacturing (AM) has been suggested as an alternative for producing parts with magnetic properties [7]. Fe-50% Ni alloy has been successfully manufactured by laser powder bed fusion (LPBF) [9, 10] as well as through direct energy deposition (DED) [11, 12]. The authors have demonstrated reaching lower magnetic saturation and permeability than the alloy fabricated by other techniques, due to the presence of porosity and other defects produced during the solidification. Nevertheless, it has been observed that by increasing the grain size and applying stress relief through heat treatment, magnetic properties, such as permeability, can be enhanced in samples fabricated using AM [13]. Haftlang et al. conducted a study on Fe–Ni alloy, investigating how its crystallographic texture can facilitate the alignment of magnetic domains, thereby capitalizing on crystallographic anisotropy [12]. Additionally, Zhang et al. who produce Fe-30% Ni by LPBF discuss the laser parameters handling to achieve low crystallite size (130 µm) and then get better magnetic properties [14]. However, the influence of important process parameters such as the laser speed and power on the magnetic properties is currently unknown.

This study aims to fabricate samples using an experimental design to statistically analyse the optimal combination of laser parameters to achieve optimum magnetic properties. The results are compared with the magnetic properties of samples obtained through conventional methods, to assess the feasibility of replacing them with additive manufacturing. The novelty of this work is in developing a new understanding of the laser parameters that positively impact the laser powder bed fusion of magnetic components made of Fe-50 wt% Ni alloy.

15.2 Methodology

15.2.1 Sample Fabrication

Commercial Fe-50 wt% Ni alloyed powders were purchased, which were fabricated using Nitrogen atomization, with a particle size < 30 μm. Cylindrical samples were fabricated and built on a stainless steel construction platform, using laser powder bed fusion (Concept Laser Mlab 200R Ge) under a nitrogen atmosphere. The samples were fabricated using a 200 W fibre laser with a spot size of ~75 µm varying the volumetric energy density (VED). Additionally, a meander scanning pattern was employed with a 67° rotation between the consecutive layers. The samples were cylinders of 8 mm in height and 6.4 mm in diameter following ASTM E9 [15]. As shown in Table 15.1, the fabrication followed the design of the experiment (DOE) with variables laser power (P) and laser speed (V), both on three levels, producing each variable set in triplicates. The layer thickness and hatching space were set to 30 μm and 70 μm, respectively. After the fabrication, the as-built samples (AB) were removed from the stainless steel platform using wire electric discharge machining (W-EDM), applying a constant voltage of 100 V and a current of 2.5 A, respectively.

Table 15.1 Samples fabricated through additive manufacturing and obtained from conventional technology (wrought)
Full size table

Additionally, an 8 mm in diameter and 50 cm in length Fe-50 wt% Ni bar (abbreviated AR—as-received) was purchased, which was cut with an abrasive disc to produce samples of 6 mm in height.

15.2.2 Microstructural and Density Characterization

The metallographic specimens were prepared by wet grinding using sandpaper from #400 to #4000 grit and then polished with colloidal silica suspension. The microstructure was revealed using an aqua regia, composed of 15 ml HCL and 5 ml HNO3, for 10 s. The surface images were produced with a scanning electron microscope (SEM) ZEISS 1540XB. The density was measured with the Archimedes method, whose values varied from 8.09 to 8.13 g·cm−3, reaching 98% of relative density.

The phase composition was determined by X-ray diffraction (XRD) using an X'Pert3 MRD XL and equipped with Cu Kα1 radiation source (λ = 1.54056 Å). The accelerating voltage and current were 40 kV and 40 mA, with a scan rate of 0.02°·s−1 in a scan range of 20–80°. The phases were identified using X’pert, and then, the Rietveld refinement was performed using MAUD software.

15.2.3 Magnetic Characterization

Next, the saturation magnetization taken at 80,000 A·m−1(B80k) and peak permeability (µr) of commercial and additively manufactured cylindrical samples were measured. The magnetizing current was provided by a bespoke hysteresis graph developed at Cardiff University [16] incorporating a Kepco bipolar DC power supply Lakeshore Gaussmeter and Fluxmeter, and a PC running a custom-designed LabVIEW virtual instrument. The sample was wound with a 10-turn coil and magnetized to a peak magnetic field strength of 80 kA·m−1 with a cycle time of approximately 5 min.

15.3 Results and Discussions

15.3.1 Morphological Characteristics

Figure 15.1 shows the Fe-50 wt% Ni powder alloy SEM image and its particle size distribution. The particles are mainly spherical, where the bigger particles have some satellites of small diameter, while the rest of the powder is composed of a large number of small particles. The histogram exhibits a log-normal distribution, where 90% of the particles are under 30 µm and many particles are about 5 µm in diameter.

Fig. 15.1
The S E M image of alloying powders with a mix of larger and smaller round-shaped particles and a histogram plot the number of particles and cumulative percentage versus particle diameter. It has a positively skewed distribution.

a SEM image of alloying powders and b histogram of particle size distribution

Full size image

Figure 15.2 shows the SEM image of an as-built sample, taken from the top surface, and its diffraction pattern. The SEM image shows the micrograph of the alloy whose grains are randomly oriented and with grain size distribution due to the fabrication method [17]. The zoom image shows the grain boundary and the structure of fine dendrites inside of the equiaxed grains observed commonly in metallic alloys fabricated through AM [18, 19]. It is important to note that small pores (approximately 5 μm) were observed in both the top view and cross-section. However, the density reached was on average 98.4% ± 0.5% as measured by the Archimedes principle.

Fig. 15.2
The S E M image of A B sample surface highlights the gain boundary in normal view and a zoomed in view and a graph plots counts versus 2 theta with three peak points at 111, 200, and 220.

a SEM image of AB sample surface and b the X-ray sample diffraction pattern collected with Cu Kα radiation

Full size image

The diffraction pattern shows peaks associated with the γ-FCC (face-centred cubic) structure with a space group Fm–3 m composed of Fe and Ni atoms. According to the Rietveld refinement, the cubic cell has a lattice parameter of 3.588 Å and suggests texture oriented in (200). F. Haftlang et al. [12] researched the Fe-50% Ni fabricated through direct energy deposition (DED) and the influence of the crystallographic orientation-dependent magnetic properties. They found that maximization of magnetic properties could be done when the grains are oriented on Cube {001} <100> [5].

15.3.2 Magnetic Properties

Figure 15.3 compares the complete hysteresis loops (B–H curves) and zoom of AR and some AB samples. The loops are typical of soft magnet materials, where the AR sample reaches saturation under a small external magnetic strength, reflecting high permeability (µr) characteristics. In fact, all AB samples exhibited lower permeability than AR samples due to the internal defects generated by LPBF, which hindered magnetic domain alignment at the same magnetic force applied. However, the magnetic saturation at 80,000 A·m−1 (B80 K) was similar between the AR and AB samples whose values were 1.71 ± 0.03 T for AR and a range of 1.56 ± 0.12 T to 1.73 ± 0.09 T for AB. It demonstrated that AB can reach similar values regarding magnetic saturation, whose values seem to depend on the alloy rather than other factors. Table 15.2 shows the principal data extracted from the hysteresis loop.

Fig. 15.3
A trend graph plots B over H, wherein a set of four trends representing A R, A B S3, A B S7, and A B S5 rises from negative 1.8 to 1.8 from the third quadrant to the first quadrant through the origin.

Magnetic hysteresis plot (B–H curves) of the magnetic field strength (H) and flux density (B) of AR and selected AB samples

Full size image
Table 15.2 Magnetic field saturation, permeability, and coercivity of the as-built and as-received samples
Full size table

Similar values have been reported for magnetic saturation [4, 8, 20] but with other fabrication techniques with permeabilities over 40,000 [12]. However, samples fabricated through AM can reach a permeability of 5,000 with a coercivity (Hc) of 100 A·m−1 [9] and a magnetic saturation of 1 T [11]. Nonetheless, the AB samples move in a wide range of Hc from 1,730.5 ± 173.8 to 455.5 ± 158.7 A·m−1 It is due to the presence of sample’s internal defects as porous, voids, residual stress, dislocation density and impurities [4, 19, 21] that impact negatively the magnetic properties. These internal defects are often part of the metallic samples manufactured by AM, which are characterized by having an average of 2% porosity when considered fully dense and the residual stress left by local heating and rapid solidification. Therefore, it is expected to achieve lower magnetic properties. However, studies have demonstrated that heat treatments that increase the grain size and allow residual stress relief could improve the magnetic properties of samples fabricated by AM [22, 23].

Figure 15.4 uses a contour graph to show the influences of the laser parameters (power and scan speed) on the magnetic properties such as the magnetic “saturation” (B80 k), the permeability (µr), and coercivity (H). The results show that the combination of low laser power and speed is possible to achieve the maximum B80k and µr and the lowest coercivity. The latter phenomenon can be attributed to the low laser speed, which enables the melting of the layers while concentrating the power in smaller increments.

Fig. 15.4
A three-part contour plot plots V versus P for A R sample 1.71, 159.3, and 559 for the H value ranging between 455 and 1,735.

Contour plot of the a magnetic field (B) collected at 80 k A·m−1, b permeability (µr), and c coercivity (H) influenced by laser power (P) and scanning speed (V)

Full size image

Figure 15.5 shows a scatter matrix that combines the magnetic characteristics and the laser parameters utilized in the construction of the samples, providing statistical evidence of their correlation using the Pearson coefficient (r). The VED is strongly influenced by the laser speed (V) rather than the laser power (P) since as the speed increases VED drops drastically. The matrix reveals an inverse relationship between P and µr, suggesting that excessive melting of the layers may lead to an increase in microstructural defects. As V increases, there is a corresponding increase in the sample density (ρ), which suggests that some magnetic properties can be attributed to the presence of porosity. This is evident in the case of B80k and H, which exhibit a slight positive correlation with density.

Fig. 15.5
A scatter matrix plots P, V, V E D, B subscript 80K, mu subscript r, H, and rho for the Pearson's value r equals 0, 0.19973, negative 0.96891, negative 0.43319, negative 0.08918, negative 0.04369, negative 0.8525, negative 0.18562, negative 0.03986, 0.46507, negative 0.30156, 0.21889, negative 0.31747, 0.52904, 0.22338, negative 0.07747, 0.63849, negative 0.69202, 0.45579, 0.19363, and 0.36107.

Scatter matrix to visualize bivariate relationships between combinations of laser power and speed, volumetric energy density (VED), sample density (ρ), and magnetic properties

Full size image

15.4 Conclusions

The comparison of the magnetic properties of samples manufactured conventionally and by LPBF, when varying the laser power and speed, indicates the following:

  • The B80k achieved for samples produced by LPBF was close to the commercial samples. This means that the magnetic saturation was somewhat unaffected by the type of manufacturing method employed.

  • The µr is greater in AR samples than AB due to the presence of many internal defects in the last, since this acts as a pinning mechanism for magnetic domain walls. This opens the opportunity to explore the effect of annealing to decrease or eliminate some of these defects to improve the magnetic properties.

  • The laser speed affects the volumetric energy, used to construct the AB samples, more than the laser power. Therefore, V influences the permeability, where at higher laser speed, the permeability decreases due to the appearance of more internal defects as porosity.

  • The best magnetic properties obtained using AM were achieved using a laser power of 190 W and a low speed of about 300 m·s−1, with total volumetric energy close to 300 J·mm−3.

  • The alloy produced through LPBF exhibits magnetic properties that are comparable to those of the same alloy obtained through conventional methods. Therefore, further research is needed to explore the potential of laser powder bed fusion and to manufacture Fe–Ni components.