**2. Materials and Methods**

Commercially available pure elemental powders were used as a starting material: Fe (MaTeck, Jülich, Germany, 99.9%, −100 + 200 mesh), Co (Goodfellow, Hamburg, Deutschland, 99.9% 50–150 μm), high purity Co (Alfa Aesar—Puratronic, Ward Hill, MA, USA, 99.998%, −22 mesh), and Cu (Alfa Aesar, Ward Hill, MA, USA, 99.9%, −170 + 400 mesh). The reason for using two grades of Co powder was the following: although the diameter of the particles of the less pure powder is rather large, the grains are small, meaning HPT deformation of high Co containing materials became difficult. For comparison, scanning electron micrographs of both types of Co powders are presented in Figure 1. Description of the other powders used is given in reference [40].

**Figure 1.** scanning electron microscopy (SEM) micrographs of Co particles. The main difference is the particle shape, the purity, and the resulting grain size. (**a**) Purity 99.9%. (**b**) Purity 99.998%.

For preventing oxidation, the powders are stored and prepared for further processing in an argon (Ar)-filled glove box. For the first processing step (pre-compaction of mixed powders in the HPT-tool), the powders are contained in a specially-made, Ar-filled compaction device [40]. After pre-compaction at a pressure of 5 GPa, the samples are deformed in air, at room temperature (RT) or at elevated temperatures, as described below (Table 1). Samples of 8 mm diameter and approximately 0.5 mm thickness were produced. To investigate the effects of varying process conditions on microstructure and the GMR, several samples of about 50/50 at% were made. They vary in sample size, processing temperature, and applied strain rate. To prove the feasibility of upscaling the HPT-process, one sample with approximately equiatomic composition was made with a different, larger HPT-tool. It is capable of applying a force of 4 MN; thus, samples of diameters of 30 mm (pmax = 5.6 GPa) and thicknesses of several millimeters can be deformed. The strain rate which was applied to the small specimens was ~0.6 s−1, whereas it was approximately 0.05 s−<sup>1</sup> for the large sample. In the latter case, the powder mixture was simply poured in the anvil's cavity and deformed under atmospheric conditions. According to Equation (2)

$$
\varepsilon\_{\rm vM.} = 2 \text{ } \pi \text{ n } \mathbf{r} \text{ / (t } \sqrt{3} \text{)}, \tag{2}
$$

the equivalent von-Mises-strain εv.M. applied to the material upon HPT is given by the radius r, the number of applied rotations n, and the thickness t. For samples of 8 mm diameter, this yields an applied strain of 2200 at a radius of 3 mm after 100 rotations, assuming a thickness of 0.5 mm. For the large sample (diameter 30 mm), a strain of 1300 was applied at a radius of 10 mm (rotations: 250, thickness: ~7 mm). Although the applied strain is smaller and the strain rate is considerably lower, the large sample heats up during HPT due to plastic deformation and a lack of heat dissipation [45]. Therefore, it was deformed at elevated temperature although no external heating device was used.


**Table 1.** Summary of investigated samples, varying in composition (energy dispersive X-ray -measurement) and processing parameters including HPT-tool size, processing temperature, number of rotations, and purity of the used Co-powder (where applicable).

To enable a thorough microstructural and magneto-resistive investigation, at least two small HPT samples were made out of each powder mixture, but only one large sample was made since several GMR specimens and specimens for microstructural analysis could be made from one sample. All relevant sample and process parameters are summarized in Table 1.

For detailed investigations of the microstructure scanning electron microscopy (SEM) and transmission electron microscopy (TEM) studies as well as synchrotron measurements were performed. Vickers hardness was measured to ensure there was a microstructurally saturated state of the sample. GMR-specimens were taken from this microstructurally saturated region. Hardness measurements were made along the radius of the HPT-disc in a tangential direction (Micromet 5104, Buehler, Lake Bluff, IL, USA). SEM micrographs using a backscattered electron detection mode (BSE) were acquired with a LEO 1525 (Zeiss, Oberkochen, Germany) that was further equipped with an electron backscatter

diffraction (EBSD) system (e- FlashFS, Bruker, Berlin, Germany) and an energy dispersive X-ray (EDX) system (XFlash 6-60, Bruker, Berlin, Germany). The conventional EBSD can easily be changed to a transmission EBSD geometry (Transmission Kikuchi Diffraction, TKD) [46,47]; thus, EBSD-TKD was performed for a chosen TEM specimen (Cu55Co45). The TEM specimen was made using conventional metallographic thinning methods and a final ion polishing step with grazing incidence. For analysing the results of EBSD-TKD and EDX measurements, the software package Esprit 2.1 from Bruker (Billerica, MA, USA) was used. The TEM/TKD—Specimen was taken at a radius of ~10 mm from the large HPT-disc, where the incident electron beam was parallel to the axial HPT-direction. Using an acceleration voltage of 30 kV and a step size of 5 nm to 9 nm (four scans in total), the orientation data was recorded. After scanning, a clean-up of the data was performed, where certain zero solutions are absorbed by the surrounding well-defined matrix. Grain boundaries were defined by having more than 15◦ misorientation between the two grains, and grains below the size of five pixels were rejected.

Using the same specimen, TEM micrographs were made with a JEOL-TEM (JEM2200FS, Tokyo, Japan). For phase analysis, transmissive synchrotron X-ray diffraction measurements were made at PETRA III, P07 at DESY, Hamburg, Germany (Deutsches Elektronen Synchrotron). A beam energy of 98.25 keV was used and recorded spectra were analyzed with the software PyFAI and compared with CeO2–standards. Magnetic properties were measured with a superconducting quantum interference device (SQUID, Quantum Design MPMS-XL-7, Darmstadt, Germany). For small HPT-discs, the specimens for SEM and synchrotron investigations were taken out in accordance with reference [40]. For the sample made with the large HPT-tool, the TEM study was performed in an axial view. For SQUID measurements, the applied field was parallel to the axial HPT-direction.

For GMR-measurements, long and thin specimens were cut out of the HPT-disc along a secant line. According to Equation (2), the center of the HPT-disc does not receive a lot of deformation; thus, it was avoided for resistivity measurements. The closest distance of the GMR-specimen to the former center of the HPT-disc was at least 1.5 to 2 mm. The length of the GMR-specimens was several millimeters; their thickness was reduced by grinding and polishing to approximately 200 μm in order to increase the voltage drop. We sought to make sure that the impact of heat during specimen preparation was kept as low as reasonably possible; no microstructural changes are expected to occur upon sample preparation. In literature, it was found that for measurements on thin films, the applied current and magnetic field are often within the film plane [48]. Following this, the radial-tangential-plane was identified as the "film plane" for small HPT-samples. In this case, the applied current has components in the radial and shear direction. For the large HPT-sample, the current was applied in an axial direction. The specimen was placed in a four-point resistance setup, which again is placed within the air gap of an electromagnet (Type B-E 30, Bruker, Karlsruhe, Germany). The air gap was fixed with 50 mm and the diameter of the conical pole pieces was 176 mm (max. field = 22.5 kOe) providing enough space within the homogeneous magnetic field for the four-point resistance setup. The voltage was measured using a multimeter (model 2000, Keithley, Cleveland, OH, USA), the current source was a sourcemeter (model 2400, Keithley, Cleveland, OH, USA), and the applied current was 500 mA to 800 mA. The strength and stability of the magnetic field was measured using a Gaussmeter (Model 475 DSP, Lakeshore, Westerville, OH, USA). The electromagnet's power supply, the current source, and the voltage and field measurements were connected for communication and data collection.
