The current work demonstrates that electron backscatter diffraction (EBSD) is a powerful and versatile characterization technique for investigating soft magnetic materials. The properties of soft magnets, e.g., magnetic losses strongly depend on the materials chemical composition and microstructure, including grain size and shape, texture, degree of plastic deformation and elastic strain. In electrical sheet stacks for e-motor applications, the quality of the machined edges/surfaces of each individual sheet is of special interest. Using EBSD, the influence of the punching process on the microstructure at the cutting edge is quantitatively assessed by evaluating the crystallographic misorientation distribution of the deformed grains. Using an industrial punching process, the maximum affected deformation depth is determined to be 200 - 300 μm. In the case of laser cutting, the affected deformation depth is determined to be approximately zero. Reliability and detection limits of the developed EBSD approach are evaluated on non-affected sample regions and model samples containing different indentation test bodies. A second application case is the investigation of the recrystallization process during the annealing step of soft magnetic composites (SMC) toroids produced by powder metallurgy as a function of compaction pressure, annealing parameters and powder particle size. With increasing pressure and temperature, the recrystallized area fraction (e.g., grains with crystallographic misorientations < 3°) increases from 71 % (200 MPa, 800°C) to 90% (800 MPa, 800°C). Recrystallization of the compacted powder material starts at the particle boundaries or areas with existing plastic deformation. The progress of recrystallization is visualized as a function of time and of different particle to grain size distributions. Here, large particles with coarse internal grain structures show a favorable recrystallization behavior which results in large bulk permeability of up to 600 – 700 and lower amount of residual misorientations (>3°).
I. INTRODUCTION
High-efficiency-electric-motors with tailored speed-torque characteristic are essential for the sustainable success of electric vehicles. Such electric traction motors require high-performance soft magnetic core materials with high induction, low magnetic losses at frequencies up to 1500 Hz, cost efficiency and large-scale-production potential. State-of-the-art electrical steels fit these requirements and are produced by the well-established cold/warm rolling technology followed by a heat treatment.1 However, realization of frequencies > 1500 Hz demand even thinner sheets (< 0.2 mm) and alloys with increased Si-content, which poses a significant challenge for rolling and shaping processes.2 Here, one focus is on optimizing the punching process and understand its impact on the materials magnetic performance. Alternatively, high-frequency applications can be realized by soft magnetic composites (SMC) which allow a high degree of freedom in the design and construction of electric motors.3 SMCs are produced by conventional powder metallurgy processes: water-sprayed iron particles (contain subgrain structure4) with an electrically insulating coating are compacted (p < 1000 MPa) and subsequently annealed at temperatures just below the thermal stability of the insulating coating (T < 600°C).5 As with electrical steel, the annealing step is essential to attain lower hysteresis losses and higher permeability due to relaxation and recrystallization of the strained material.6 Alternative coating systems like MgO and multilayers with improved thermal stability7 allow higher annealing temperatures which in turn facilitates a favourable adjustment of the SMC microstructure.4 This represents a promising way for the development of SMCs that can compete with electrical steels in respect to hysteresis losses and permeability. The isotropic behaviour of the grain orientation in non-oriented electrical steel and SMC material is illustrated by electron backscatter diffraction (EBSD) technique (Fig. 1). In addition, a significant difference in grain size between the two materials is observed. While the coarse structure of the electrical steels enables high permeability and low hysteresis losses, the fine grain substructure (in the following called grains) of the soft magnetic composites (SMCs) is considered to negatively influence the hysteresis losses.8
Inverse pole figure (IPF) of the two different soft magnetic materials show the structure orientation of typical a) conventional electrical steel and b) SMC microstructure at same magnification, 100x, EBSD.
Inverse pole figure (IPF) of the two different soft magnetic materials show the structure orientation of typical a) conventional electrical steel and b) SMC microstructure at same magnification, 100x, EBSD.
However, the coarse grain structure of electrical steel is the reason that the material reacts very sensitively to stress and strain introduced during the punching process. Here, large grains easily transfer energy over an extended distance resulting in rather broad zone of “disturbed grains”.9 For SMC materials, the inner grain structure of the particles is a required property to reach high density during compaction. On the other hand, the recrystallization process is inhibited and a part of the structure remains in a deformed condition. This fact is noticeable at the diffuse grains, with a gradient in colour on Fig. 1b) that indicates non-recrystallized and deformed areas. EBSD-analysis can be a powerful tool to visualize and quantify the content of deformation and recrystallized structure.10–13 This enables the establishment of structure-property relations and thus the possibility for specific material design. E.g., this was successfully demonstrated for Ni-based materials and stainless steels in Ref. 10. A qualitative assessment of the influence of the punching tool condition (wear conditions) on resulting microstructure and magnetic performance is presented in Ref. 14. The present paper focuses on the development of a reliable EBSD-based approach to detect and quantify local grain misorientations introduced by mechanical load during a punching process of electrical steels. The developed approach is then deployed for the analysis of SMC materials to quantify the degree of recrystallization during annealing which in turn is compared with magnetic performance.
II. EXPERIMENTAL PROCEDURES
For the microstructure investigations, polished microsections (cross-sections) have been prepared by metallographic techniques (Struers RotoPol-31). Special attention was paid to eliminate scratches and material break-outs during the polishing process. By combined scanning electron microscopy (SEM, ZEISS Sigma 300VP) and electron backscatter diffraction (EBSD, EDAX Tident), the orientation of the structure was analyzed and existing misorientations were determined (OIMA 7.3). Using optical microscopy in polarized mode (Kerr microscopy, ZEISS AxioImager Z2.m), the magnetic domain patterns were visualized. For the investigations, commercially available electrical steel and SMC material were used. In the case of the SMC material the powder was compacted to a toroidal geometry (inner/outer diameter of 40/30 mm, height of 5 mm) using different pressure steps (200, 400 and 800 MPa) and annealing parameters (temperature 400°C – 800°) under inert gas atmosphere. In addition, the SMC powder was sieved to obtain fractions between 50 to > 400 μm (50 μm steps) and to determine the influence of particle size distributions on recrystallization activity. For better comparability, all SMC samples have been investigated in the same orientation of the cross section with respect to the pressing direction. The method of EBSD can be utilised for the quantification of the pattern quality (image quality, IQ) and the orientation distribution (inverse pole figure, IPF) of structure details. Measuring points with the same orientation can be related to a grain, which allows the determination of size and alignment of all the grains of an image. If the local misorientations within a grain are determined, plastic deformations can be rated. There exist several possibilities for EBSD data analysis. The reference orientation deviation (ROD) reveals the highest misorientation compared to a pre-defined reference (or averaged orientation within a grain). The kernel average misorientation (KAM) compares neighbouring measuring points concerning their orientation and highlights finest local deviations. In addition, crystal lattice defects such as high dislocation densities or grain boundaries lead to diffuse diffraction lines which can be described in terms of the pattern contrast or the image quality. Therefore, also the pattern quality can illustrate the presence of stresses like elastic tension. Only grains that are completely visible in the image and with a size of greater than 10 pixel were selected for quantification. For all detected individual grains, the parameters Feret (max.) (largest extension), Feret (min.) (shortest extension) and area were determined. All the data were cleaned-up before the analysis with the neighbour orientation correlation (NOC) approach. For each magnification the step size and measurement speed were adapted. For magnetic analysis of the SMC fractions, the toroids were coiled with copper wire with 50 windings at primary and secondary side. The measurements were performed using the testing system MPG 200D of Brockhaus Messtechnik GmbH. The magnetic measurements in DC (cumulative) were conducted between 50 and 5000 A/m.
III. RESULTS AND DISCUSSION
A. Non-affected material state and model experiment
The investigation of the misorientations of electrical steel samples in the non-affected state is illustrated in Fig. 2a) and b). The pictures show the cross-sectional and plane view on the metallographic structure with ROD. The majority of the grains (∼90%) are green coloured corresponding to a very low misorientation angle below 1°. The small amount of yellow (∼8%) and red (∼2%) highlighted grains with higher misorientation angles have the origin mainly in metallographic preparation process or other unfavourable structure peculiarities and cannot be avoided completely. There were no misorientations >5° detected.
Overview of the non-affected state of electrical steel material in ROD illustration with a) cross-sectional view on a multilayer of the sheet sample and b) the plane view on the analyzed microstructure.
Overview of the non-affected state of electrical steel material in ROD illustration with a) cross-sectional view on a multilayer of the sheet sample and b) the plane view on the analyzed microstructure.
To investigate the influence of deformation and to define a reference sample, indentation experiments have been performed on grain-oriented electrical steels (Fig. 3). The result in Fig. 3a) shows examples of Vickers and Brinell deformations on the cross-sections of electrical steel sheets using the tool KAM. The higher magnification in Fig. 3b) illustrates the effect of plastic deformations in close vicinity to a Vickers indentation on the misorientation measurements. Besides the clear observation of the plastic deformations, the shear bands and some scratches become visible. In contrast, the tool ROD in Fig. 3c) visualizes the extent of detectable misorientations around the indentation impact zone. From this, it becomes obvious that misorientations can be detected up to 50 μm distance from the center. The largest deformations are evaluated and amount to ∼10% > 10° and 6% > 5°. Performing Kerr-microscopy at the same position, the resulting effect on the domain structure becomes visible. It has to be noted that the domain structure reflects plastic and elastic tensions in the affected area and a strong distortion in the area beyond deformations.
Model experiment with indentation impacts on the cross-section of grain oriented electrical steel (GO) with (a) a survey of some indentation impacts (Vickers and Brinell) and (b) higher magnification of a Vickers indentation both using KAM, (c) the same position using ROD and (d) Kerr-image of the surrounding of a Vickers indentation.
Model experiment with indentation impacts on the cross-section of grain oriented electrical steel (GO) with (a) a survey of some indentation impacts (Vickers and Brinell) and (b) higher magnification of a Vickers indentation both using KAM, (c) the same position using ROD and (d) Kerr-image of the surrounding of a Vickers indentation.
B. Depth of cutting edge, electrical steel
The influence of the cutting method on the deformed area in electrical steels is investigated. Fig. 4 shows examples for edges processed under three different conditions. The images have been taken in the cross-sectional view vertical to the tool movement. The first sample shown in Fig. 4a) was roughly separated using a blunt object. The EBSD analysis reveals a large amount of misorientations (∼25% >5°). The sample shown in Fig. 4b) was processed using an industrial punching tool. The result shows a detectable deformation of depth of around 150-200 μm and large amount of misorientations (∼8% >5°). The last sample presented in Fig. 4c) shows the result of a laser cutting process with some small areas of misorientations present (but without detectable change in orientation, 0% >5°). It has to be noted, that the edge itself for some samples (see, e.g., Fig. 4b) cannot be analyzed because of the weaker measured signal due to the high plastic deformations.
Comparison of the influence of different cutting methods on the deformation of the edge of electrical steel sheets: (a) rough cutting using a blunt object, b) industrial punching tool and c) laser cutting. The images have been taken in the cross-sectional view vertical to the tool movement using ROD.
Comparison of the influence of different cutting methods on the deformation of the edge of electrical steel sheets: (a) rough cutting using a blunt object, b) industrial punching tool and c) laser cutting. The images have been taken in the cross-sectional view vertical to the tool movement using ROD.
The deformed area obtained by means of an industrial punching tool is illustrated in Fig. 5 in the plane view, i.e. in direction of the tool movement.
Comparison of two different positions in the electrical steel sample edge with (a) slim strip structure and b) standard cutting edge. The images have been taken in the plane view in direction of the tool movement using ROD.
Comparison of two different positions in the electrical steel sample edge with (a) slim strip structure and b) standard cutting edge. The images have been taken in the plane view in direction of the tool movement using ROD.
In Fig. 5a) the influence of the industrial punching process on both sides of a slim material strip with some deformed regions is demonstrated. Fig. 5b) in contrast shows a position with no slim geometry. Both examples are characterized by a maximum deformed region of around 200 – 300 μm in depth. The slim strip structure of Fig. 5a) shows a higher deformation depth concentrated at positions with larger grains. It has to be noted that the quantified depth of deformation is somewhat larger for ROD compared to KAM. This is due to the fact that ROD considers a total grain and KAM considers neighbouring measuring points.
C. Compaction and recrystallization investigation, SMC materials
Compared to electrical steels the situation for SMC materials concerning deformed cutting edges is completely different. In Fig. 6 a section of a compacted (800 MPa) toroid in green condition manufactured by using a powder fraction (< 50 μm) of SMC powder is presented. Accordingly, after near-net-shape fabrication based on powder metallurgy technology no detectable additional deformation exists from the tool used. Instead, misorientations occur in all the iron particles. Here, the misorientations are concentrated on the particle boundaries.
Section of a compacted (800 MPa) toroid in green condition manufactured by using a powder fraction (< 50 μm) of SMC: a) inverse pole figure of orientation and b) peaks of misorientation with KAM.
Section of a compacted (800 MPa) toroid in green condition manufactured by using a powder fraction (< 50 μm) of SMC: a) inverse pole figure of orientation and b) peaks of misorientation with KAM.
Using different pressures (200, 400 and 800 MPa) and annealing temperatures (400, 600°C and 800°C) during manufacturing of industrial SMC powder material, the recrystallization behavior can be modelled (Fig. 7). For the green condition, the number of well-oriented grains with misorientations <3° (hereafter defined as recrystallized structure) decreases with increasing pressure (1a-c). The situation is very similar after additional annealing at 400°C for 20 h, i.e. the content of recrystallized structure is nearly the same (2a-c). For a temperature of 600°C the recrystallization starts taking place in the structure which is indicated by the green colored regions (3a-c). For non-conventional annealing at a temperature of 800°C, the recrystallized area with misorientations <3° is further increased and reaches values between 71% (200 MPa) and 90% (800 MPa), respectively (4a-c). The behavior of misorientations is further summarized in Fig. 8a). Starting from the green condition compacted using 800 MPa, an average misorientation of around 2° is present. By annealing at 800°C for 20 h, 70% of the measured points show misorientations <1°. The formation of new grains by recrystallization is illustrated in the cumulated area-weighted frequency in Fig. 8b). During annealing, the coarse grain size distribution (d50 = 115 μm, 50 = 50% of grains possess a size lower this value)) of the green condition changes and the SMC structure reaches a finer average grain size of around d50 = 50 μm.
Systematic variation of compaction pressure and annealing temperature to determine the amount of recrystallized structure (green, < 3° misorientation) of an industrial SMC material using ROD.
Systematic variation of compaction pressure and annealing temperature to determine the amount of recrystallized structure (green, < 3° misorientation) of an industrial SMC material using ROD.
Systematic variation of compaction pressure and annealing temperature to determine a) the amount of misoriented structure for an industrial SMC material using KAM (<5°) and b) the cumulated area-weighted frequency of a green condition sample and annealed sample.
Systematic variation of compaction pressure and annealing temperature to determine a) the amount of misoriented structure for an industrial SMC material using KAM (<5°) and b) the cumulated area-weighted frequency of a green condition sample and annealed sample.
In addition, the evolution of recrystallization depends on the deformation behaviour of the material and is a result of the particle to grain size distribution of the material. Larger particles seem to become easier plastically deformed in comparison to small particles, assuming that the grain structure inside the particles is the same.
By direct comparison of sieved SMC-powders with different particle-/grain size distributions (process-related variations are due to water spray process) in the toroids as shown in Fig. 9, the progress of recrystallization can be illustrated. The advantage of larger grains is to get flat and more deformed grains which initiates higher recrystallization activity over time. Fig. 10 illustrates the increasing permeability of toroids made of different sieved SMC powder fractions. The particle size distribution from 100 μm to 150 μm presents the most coarse grain distribution and highest increase in permeability (+75%). The largest particles > 400 μm and 350 – 400 μm reach the next highest values (+40%) but have obviously finer grains. With an ideal composition of particle size and tailored grain structure (large particles and coarse grain) the magnetic performance of the material can be increased systematically. EBSD analysis, therefore, can be a powerful tool to develop the microstructure of this next generation of SMC materials.
Comparison of the evolution of recrystallization of sieved SMC materials compacted with 800 MPa and annealed in different time steps for the fractions 1a)-1c) > 400 μm and 2b)-2c) 200 – 250 μm using ROD (<3°).
Comparison of the evolution of recrystallization of sieved SMC materials compacted with 800 MPa and annealed in different time steps for the fractions 1a)-1c) > 400 μm and 2b)-2c) 200 – 250 μm using ROD (<3°).
Permeability as a function of field for sieved SMC materials for the a) green condition, b) sample annealed at 400°C and c) sample annealed at 600°C.
Permeability as a function of field for sieved SMC materials for the a) green condition, b) sample annealed at 400°C and c) sample annealed at 600°C.
IV. CONCLUSION
In this paper the influence of different cutting methods on the deformation depth at the edge of electrical steels and the recrystallization behaviour of SMC materials during annealing have been investigated. For microstructure analysis, advanced EBSD tools have been used to evaluate misorientation and quantify the influence of deformations. In the case of electrical steels the depth of deformation was determined to be 200 – 300 μm when an industrial punching tool was used whereas for laser cutting the depth of deformation was determined to be zero. Using unaffected reference samples and model samples containing different indentation test bodies, the performance and limits of EBSD analysis have been demonstrated. In the case of SMC materials, from the evaluation of recrystallization it has been found, that the compacted powder particles showed the highest deformation at the particle boundaries. During different annealing procedures of different time and temperature the amount of recrystallized grains/areas increase (71 % (200 MPa, 800°C) to 90% (800 MPa, 800°C)). Larger particles show a stronger deformation behavior than smaller particles. By choosing powder fractions of larger particle sizes and a coarse grain structure for the production of SMC toroids, the permeability of SMC materials could be increased. From the results it can be concluded, that EBSD is a powerful and versatile characterization technique for investigating soft magnetic materials.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the German BMBF for funding.