A. Factors affecting magnetic properties

Both material structure and composition control the magnetic properties of Fe–Si sheet. Aspects of the material structure considered most important are the grain size (d) and crystallographic texture, the combination of which predominantly influences the magnetic permeability (µ), coercivity (H _c), and hysteresis loss (W) properties. Perfectly isolating the independent effects of grain size and texture is difficult but studies indicate that increased grain size increases the permeability while decreasing the coercivity and hysteresis loss. ^{Reference Herzer3–Reference Shiozaki and Kurosaki5} The mechanism of improved properties at larger grain sizes is attributed to a reduction in grain boundary area, which eases the motion of magnetic domains during microscale magnetization processes. ^{Reference Cullity and Graham6}

Crystallographic texture, by virtue of the intrinsic magnetic anisotropy of the α-BCC structure in Fe–Si sheet, also influences these same properties. The main objective in developing textures is the influence on the orientation of the easy magnetization directions, along 〈001〉. Perhaps the best example illustrating this is the two available commercial rolled Fe–Si sheet structures (Fig. 1). In the case of the NGO sheet, the texture is weak with a (near) random orientation of the 〈001〉 directions both in and out of the sheet plane [Fig. 1(a)], resulting in a low degree of anisotropy in the magnetic properties; the performance here is nearly independent of the applied field (H) direction. The NGO sheets are thus most commonly utilized in electric motor applications, where the field is rotating continuously. In contrast, the GO sheets [Fig. 1(b)] possess a strong Goss texture from secondary recrystallization, wherein one of the 〈001〉 directions lies along the sheet length (rolling direction, RD), leading to near ideal magnetic properties in this direction (i.e., high permeability with low coercivity and hysteresis loss). GO sheets are primarily used in power distribution transformers, where the easy magnetization direction can be configured mostly parallel to the magnetic flux direction. The GO sheets also exhibit a significantly larger grain size (d > 4 mm), by virtue of the secondary recrystallization, compared to NGO sheets (d ∼ 150 μm), although this characteristic is considered to contribute less than the texture to benefiting magnetic properties. As a whole, the essential goal of texture development in Fe–Si sheet is to reduce the misorientation of the 〈001〉 directions with respect to the applied H field.

FIG. 1.

Schematics showing (a) non-grain-oriented (NGO) and (b) grain-oriented (GO) sheet structures. Superimposed are the orientations of select 〈001〉 directions, which are responsible for the vastly different sheet magnetic anisotropies.

In addition to structure, composition also influences the magnetic properties. Alloying with silicon significantly enhances the intrinsic magnetic properties of iron by increasing the permeability, while reducing coercivity, magnetostriction and core losses—particularly important in transformer applications. ^{Reference Chen7} Commercial rolling, however, is limited to producing Fe–Si sheet with a narrow silicon composition range (1–3 wt% Si), even though it is well-known that higher Si content alloys have superior magnetic properties. ^{Reference Haiji, Okada, Hiratani, Abe and Ninomiya8} In fact, Fe–Si alloy with 6.5 wt% Si content is considered the best for magnetic applications. But major decrease in workability arise in alloys containing >3.5 wt% Si. Such alloys crack readily during rolling preventing their cost-effective manufacturing.

Interstitial impurities are also an important aspect of the sheet composition. Interstitials have been documented mostly as having negative effects on magnetic properties, where the local elastic strains around the misfitting solute atoms disturb the motion of magnetic domains. Properties such as permeability, coercivity, and hysteresis loss are negatively impacted as a result. Therefore, it is most critical to minimize the content of carbon, sulfur, nitrogen, and oxygen impurities. ^{Reference Littmann1} Carbon is particularly important because, in addition to being an interstitial element, its presence in small amounts significantly increases the critical silicon content required to maintain the desired α-phase microstructure at elevated temperatures. Furthermore, C can cause precipitation of fine Fe₃C particles in service over time (aging), reducing the long-term efficiency of the Fe–Si sheets, in addition to decreasing the overall magnetic performance. ^{Reference Littmann1,Reference Yensen9} Therefore, C both in and out of solution with Fe–Si tends to reduce the magnetic performance. Similar effects have been documented for sulfide, oxide, and nitride inclusions. ^{Reference Benford10,Reference Leak and Leak11}

Thus grain size, texture, and composition are the three most critical elements that influence the magnetic properties of Fe–Si sheet. In commercial processing by rolling, a limited range of textures and (silicon) compositions are achievable in Fe–Si sheet. Recently, a simple shear deformation process—large strain extrusion machining (LSEM)—was successfully demonstrated for producing continuous Fe–Si sheet with controllable shear texture orientations. Furthermore, the process is also capable of producing sheet from high-silicon alloys, up to and including 6.5 wt% Si content, noted for their very low workability. The present paper presents magnetic property measurements from a 4 wt% Si, shear-textured Fe–Si sheet and compares the properties with those of rolled sheet.

B. Large strain extrusion machining (LSEM)

Details of the LSEM (shear) process have been described previously, but a brief summary of the principal process attributes is given here. Figure 2 shows a schematic of plane-strain LSEM in which a sharp, wedge-shaped cutting tool engages a rotating workpiece (WP) at a preset depth (t _o). The interaction between the tool and WP produces a chip extrudate (sheet) by intense shear along a narrow (∼100 µm thick) deformation zone, commonly idealized as a single shear plane [plane OA, Fig. 2(a)]. ^{Reference Efe, Moscoso, Trumble, Dale Compton and Chandrasekar12} The process of metal removal imposes large shear strains in the extrudate material. The confinement of material flow by a secondary die (constraint) placed above the primary tool, defines the geometry for the sheet with final thickness (t _c). This confinement of the deformation allows for extrusion of continuous sheet by suppressing localized flow (i.e., segmentation and shear banding), particularly important in metals of limited workability like Fe–Si. Additional details on the unique capability of LSEM to overcome the poor workability limitation of high-Si by enhancing hydrostatic pressure, and temperature via adiabatic heating, in the deformation zone can be found in Ref. [Reference Kustas, Sagapuram, Trumble and Chandrasekar13].

FIG. 2.

Details of plane-strain LSEM: (a) schematic, (b) side view of material flow during sheet formation with rake face normal (RFN) and chip flow direction (CFD) axes, and (c) example Fe–Si sheet.

The distance between the constraint and cutting tool is controlled a priori and defines the chip thickness ratio λ = t _c/t _o. The LSEM process is best described as simultaneous cutting-extrusion, with the extrudate sheet subjected to a large strain (>1) that is determined primarily by λ. A side-view of the material flow based on high-speed imaging that highlights the confined shear is shown in Fig. 2(b), and a section of a continuous Fe–Si sheet sample is shown in Fig. 2(c).

The strain and deformation path (texture) can be varied in this processing independent of the sheet thickness (t _c). The control parameters are λ and the tool rake angle, α, enabling a range of microstructures and predictable shear textures in Fe–Si sheet, which are not achievable by rolling. ^{Reference Kustas, Sagapuram, Trumble and Chandrasekar13}

The ability to process alloys of low workability by LSEM stems from the fact that hydrostatic pressures in the deformation zone can be adjusted to be much higher than in rolling, along with sufficiently high deformation temperatures due to confinement of the flow (in situ deformation-induced heating). ^{Reference Efe, Moscoso, Trumble, Dale Compton and Chandrasekar12,Reference Kustas, Sagapuram, Trumble and Chandrasekar13} These factors allow also for important control of the sheet microstructure by dynamic processes (e.g., recrystallization), which can be further enhanced by workpiece heating (T _o) prior to cutting. Dynamic recrystallization is particularly significant and useful in LSEM of Fe alloys, which traditionally require many passes of rolling at high temperatures to accomplish equivalent microstructural modifications. Thus, continuous sheet of high-silicon Fe alloy can be produced in a single deformation step with dynamically annealed structures, and a range of shear textures. This approach helps transcend the current limitations in rolling of high-silicon Fe sheet.

A. Specimen preparation

The starting Fe–Si material for the deformation processing experiments was a commercial hot-rolled plate from Scientific Alloys, Inc. with a nominal composition of 4 wt% Si. The full composition provided by the manufacturer was, in wt%: 3.83 Si, 0.028 C, 0.018 P, 0.015 S, 0.006 Al, balance Fe. For the shear-based LSEM process, a single experimental condition was analyzed at $\bar \varepsilon$ = 1.2, (λ = 1.5, α = 5°), T _o = 25 °C, and V _o = 0.5 m/s. The resultant sheet was 3 mm wide with a thickness (t _c) of 0.25 mm. A rolled sheet was also prepared from the same alloy by warm rolling at 300 °C to a similar strain as the LSEM condition. This sample had a width of 4.8 mm and thickness of 0.63 mm. Following the deformation processing, both the LSEM and rolled sheet specimens were given a recrystallization annealing treatment at T = 700 °C for 30 min to develop an equiaxed grain size d ∼25 µm. The resultant microstructures and textures of the two specimens after the annealing are shown in Fig. 3.

FIG. 3.

Microstructure and texture in the Fe–4 wt% Si sheet: (a) and (b): LSEM and rolled sheet microstructure, respectively; (c) and (d): crystallographic texture of the LSEM and rolled sheet, respectively. Note that the simple shear texture of LSEM is tilted with respect to the sample axes by 55°, unlike the rolling texture. The degree of tilt (orientation) of the simple shear texture can be controlled in the LSEM.

For both the LSEM and rolling conditions, the annealing treatment resulted in similar equiaxed grain size d = 20–30 µm [Figs. 3(a) and 3(b)]. The texture measurements [(101) pole figures, Figs. 3(c) and 3(d)] showed that each sample possesses fundamentally different crystallographic textures. The LSEM produces a simple shear texture, characterized by two partial inclined {110} and 〈111〉 fibers, highlighted on the (101) pole figure for reference. The two characteristic fibers represent rotations about the (110) crystallographic plane, which is parallel to the shear plane, and the 〈111〉 crystallographic direction, which is aligned with the shear direction. This inclined shear texture is unique compared to the measured γ (〈111〉//normal direction) and α (〈110〉//rolling direction) texture fibers of the rolled sheet [Fig. 3(d)]. Details of the shear textures from LSEM have been reported previously. ^{Reference Kustas, Sagapuram, Trumble and Chandrasekar13,Reference Kustas, Sagapuram, Chandrasekar and Trumble14} The angle by which these fibers are inclined relative to the sheet length (chip flow direction, CFD), designated as ϕʹ, can be measured directly on the pole figure; in this case ϕʹ = 55°. Such an orientation is consistent with the shear plane model predictions for the LSEM, where a near one-to-one correspondence between expected shear plane angle and measured texture orientation has been observed. ^{Reference Kustas, Sagapuram, Trumble and Chandrasekar13} Noteworthy is the presence of a shear texture despite a fully recrystallized microstructure, which is believed to result from continuous-type recrystallization mechanisms, different than the discontinuous recrystallization of rolled sheet. ^{Reference Kustas, Sagapuram, Trumble and Chandrasekar13,Reference Humphreys15,Reference Jazaeri and Humphreys16}

Magnetic properties were measured in the LSEM and rolled sheet after the annealing. They were also measured on a specimen from a non-grain-oriented (NGO) commercial sheet of thickness 0.27 mm and width 5 mm with trade name DI-MAX® HF-10X (NGO), supplied by AK Steel, Middletown, OH. The NGO sheet has a typical specification of 3.3 wt% Si with 0.9 wt% Al and exhibited a grain size d of 150 μm.

B. Magnetic properties

Magnetic properties of Fe–Si alloy sheet have been extensively characterized and methods to measure these properties have been thoroughly standardized. ^{Reference Fiorillo17} The most common method to characterize commercial rolled sheets is through the Epstein test, in which multiple sheet segments are stacked into a frame with overlapping corners. ^{Reference Fiorillo17,Reference Sievert and Ahlers18} This method, however, requires wider sheet (30 mm minimum) than is currently available by the laboratory LSEM processing. Consequentially, an alternative approach was taken for the smaller width, single sheet specimens.

Quasistatic hysteresis loops from single sheet specimens were measured at Magnet-Physics, Inc. in Fishers, IN using an IEC 60404-4 standard closed-circuit permeameter (Remagraph®C). Samples were lightly clamped within the measuring yokes and the polarization (J) was measured using J-compensated surrounding coils. In reporting the magnetization, polarization J was converted to induction B using $B = {\mu _{\rm{o}}}H + J,$ where ${\mu _{\rm{o}}}$ is the permeability of vacuum. Magnetic properties, including maximum relative permeability (µ_max), induction at H = 60 kA/m (B ₆₀), coercivity (H _c), and hysteresis loss (W) were determined from the measured hysteresis loops.

Sheet specimens with a minimum cross-sectional area of 0.3 mm² and length of ∼92 mm were analyzed. Prior to the measurements, the oxide scale resulting from the annealing was removed from the sheet surfaces by fine grinding using 600-800 grit SiC abrasive paper followed by pickling in a 1:5 nitric acid-ethanol solution for 3–5 min. For the NGO sheet, sections of bulk commercial sheets were cut and given a manufacturer-recommended stress relief anneal in a box furnace at 800 °C for 5 min prior to the oxide removal. Sample properties were gathered along the longitudinal directions, corresponding to the chip flow direction (CFD) for LSEM and rolling direction (RD) for rolling, respectively. Hysteresis loops were measured using an applied H field up to ±60 kA/m at room temperature. The rate of magnetization was controlled such that hysteresis loops were generated in ∼1 min so as to maintain a constant change in induction with time. The applied field was obtained using a C-shaped potential coil directly attached to the specimen surface. As an example, a B–H hysteresis loop measured for the commercial NGO electrical steel sheet from AK Steel is shown in Fig. 4.

FIG. 4.

Hysteresis loop for the AK Steel NGO sample: (left) portion of the right quadrant, (middle) full hysteresis loop, (right) 500 A/m zoomed hysteresis loop. Note, the loop shown here includes the virgin magnetization curve, from which maximum relative permeability is calculated.

The B–H hysteresis loop for the commercial NGO specimen exhibited a magnetically soft response, as expected of ferromagnetic materials like Fe–Si, which terminated at B ₆₀ = 1.92 T. High B values were obtained with a relatively small applied H field (left inset), indicative of high permeability and a narrow hysteresis loop width (right inset), yielding low coercivity (H _c) and hysteresis loss (W). Analysis revealed that µ_max was 10,900 for this condition, which is nearly identical to the value of 11,000 usually obtained from the Epstein frame for the same alloy sheet product. ^19,20 This was taken as confirming that the present characterization process yields accurate magnetic properties.