**4. Discussion**

#### *4.1. Influence of Processing Conditions on Microstructural Evolution*

Severe plastic deformation, especially at low temperatures, will promote the formation of supersaturated solid solutions. Diffusion, also taking place during the deformation process, will lead to segregation and phase formation, and as a result the resulting microstructure and elemental distribution depends on the strength of the one or the other process. As already displayed in Figure 2, the microstructure of the large sample made of Cu55Co45, resembles the one of high Co-content, deformed at high temperatures. For equiatomic composition however, there is quite a difference regarding Co agglomeration when comparing the sample processed with the large HPT and small HPT-samples. The following two paragraphs will discuss the influence of sample heating and strain rate on the evolving microstructure:

When deforming a sample with a large HPT tool, more volume is deformed and more strain energy ( τ γ dV) is transferred to heat during the whole experiment. Applied shear γ is proportional to *r* and integrating the infinitesimal volume V = 2π r h *d*r yields a r3-dependency of strain energy. Using typical values: the hardness value (~350 HV) gives an approximate shear strength of 500 MPa and 150 s per rotation result in a power of ~150 W for large HPT-samples, whereas one rotation taking 45 s for the small HPT-equipment results in a heating power of approximately 10 W. The increased energy released upon deforming a sample with larger HPT-equipment is counterbalanced by the decreased strain rate and by the increased diameter of the sample and the anvils, representing the samples "cooling finger". The latter point loses importance for long-lasting experiments (~10 h in total for the Cu55Co45 sample) as the anvils heat up and their cooling capability decreases. As a result, a large sample, which is nominally deformed at room temperature, heats up more easily, leading to accelerated diffusion and segregation processes.

The second important point that has to be taken into account regarding microstructural evolution is the lower applied strain rate for the large HPT-tool. The strain rate is proportional to angular velocity divided by thickness times the radius and for the conditions described above, smaller samples were subjected to an approximately 12 times higher strain rate compared to the large sample. During HPT, the tendency of Cu and Co to segregate by diffusional processes is counter-acted by severe shear deformation. For the large HPT-sample segregation processes are more pronounced and as a result the microstructure evolves, as can be seen in Figure 12. The microstructure is shown in tangential view for radii r = 5 mm, 10 mm, 15 mm and a clear increase in the size of dark regions is visible. Thus, segregation is promoted for large radii and an increased local temperature could therefore be deduced for regions of higher strain rate. Edalati et al. determined a temperature difference of several degrees between the center and the rim of the HPT-disc. This was done by FEM simulations for a sample of 10 mm diameter [45].

**Figure 12.** SEM—backscattered electron detection mode (BSE) micrograph of Cu55Co45, deformed with the large HPT tool at elevated temperatures. Micrographs were taken in tangential direction and for increasing radius (r = 5 mm, 10 mm, 15 mm) from left to right.

Regarding the results from synchrotron X-ray diffraction measurements (Figure 6), the low Co-containing sample (Cu81Co19) shows pure fcc phases for large radii. Diffraction lines from hcp-Co, still visible e.g., at a radius of 1 mm, vanish for higher applied strains. For high Co-contents (Cu22Co78), where deformation was possible to be performed at 300 ◦C, there are still hcp-Co particles remaining for large radii, as can be seen e.g., by a shoulder in the diffraction peak at q ~29 nm−<sup>1</sup> . Nevertheless, a solid solution of the two elements can be seen for both samples. The lines are not at their nominal position for pure elements, but slightly shifted to the right (from Cu towards fcc-Co, for low-Co concentration) or to the left (from Co towards Cu, for high Co concentration). The situation is different for the Cu55Co45 sample. Here, the diffractogram is dominated by very broad peaks—especially for higher q-values, which can be easily explained by overlapping peaks at (or close to) the original position of Cu and fcc-Co.
