**5. Conclusions**

Thus, recent studies have shown that SPD significantly a ffects the atomic structure and properties of amorphous alloys. The variation of microstructure resulting from SPD processing is closely related to processing parameters (amount of shear strain, temperature of processing, imposed pressure).

In a number of publications, it was shown that nanocrystallization in the amorphous phase occurs under SPD. We should note that under conventional schemes of deformation, nanocrystallization was observed in a small fraction of amorphous material in shear bands, whereas during HPT, it takes place throughout the entire volume of amorphous samples. It is interesting to note that during HPT of an amorphous alloy (for instance Nd–Fe-B), the amorphous phase is decomposed into the amorphous and crystalline phases of basic pure metals. In some amorphous alloys (for instance, MS Ti50Ni25Cu25), besides nanocrystallization during HPT processing other complex transformations of the structure were observed. Amorphous "clusters" become visible, and these "clusters" represent some kind of amorphous structure. This contrast could be the result of the existence in the amorphous phase of regions with reduced free volume and with enhanced free volume.

During the HPT processing of a number of other amorphous alloys, nanocrystallization was not observed, but a new amorphous structure could be formed as a result of HPT, depending on changes in the short-range order, the total amount, and the redistribution of free volume. Perhaps, structures produced by the SPD methods are in some aspects comparable to the nanoglass-type structures produced by IGC.

Correspondingly, as a result of the HPT processing of amorphous alloys, essential transformations occur in their properties, in particular mechanical properties. For instance, as a result of preliminary HPT processing, the fracture fractography changes. Nanoindentation studies have shown that HPT processing leads to a significant increase in the values of the strain rate sensitivity in comparison with the initial state. At the same time, the course of change of the elastic modulus in a Zr-based BMG depends on the temperature of the HPT processing (20 or 150 ◦C). In some cases, HPT leads to a decrease in the values of Young's modulus. The first work, indicating the emergence of tensile ductility in some BMGs (Zr65Al7.5Ni10Cu12.5Pd5) after HPT processing, has been published. The emergence of tensile ductility can be explained by the formation of a high density of nanoscale inhomogeneities in the amorphous state. High tensile strength, high hardness, and low elastic modulus provide the grea<sup>t</sup> potential of BMGs for various commercial applications; however, these applications are limited by the brittleness of amorphous materials. Thus, a decrease in the elastic modulus and an increase in tensile ductility via HPT processing can provide wide applications for amorphous alloys.

It has been also revealed that the process of crystallization that occurs during the annealing of the MS amorphous alloys subjected to HPT di ffers significantly from the crystallization of the non-deformed analogues. In the case of MS Nd-Fe-B alloys, this enabled producing higher magnetic properties via a combination of HPT processing and annealing than via annealing of non-deformed analogues. Therefore, the combination of HPT processing and annealing in some cases can lead to the formation of specific nanostructured states with improved functional properties. However, currently, many aspects of the nature of the transformation of the structure and properties of amorphous alloys subjected to HPT are still unclear and require further research.

A discrepancy between the experimentally observed and predicted shear strains has been detected. The actual strain is significantly smaller than the predicted one.

The authors proposed a new method, "accumulative HPT", to achieve high strain in hard materials, including BMGs. The study showed that the structure of the Zr-based BMG during accumulative HPT transforms much more significantly than in the case of conventional HPT with the same number of revolutions.

**Author Contributions:** Conceptualization, D.G.; writing, original draft preparation, D.G.; writing, review and editing, V.A. All authors read and agreed to the published version of the manuscript.

**Funding:** The work is funded by RFBR IND-a Research Project 19-58-45014. **Acknowledgments:** The authors are grateful for the productive collaboration of Ruslan Valiev, Aleksandr Glezer, Jing Tao Wang, Horst Hahn, Dmitry Louzguine-Luzgin, Aleksandr Aronin, Galina Abrosimova, Yulia Ivanisenko, Evgeniy Ubyivovk, Evgeniy Boltynjuk, Andrey Bazlov, Roman Sundeev, Anna Churakova, Askar Kilmametov, Almir Mullayanov, Alfred Sharafutdinov, Ilshat Sabirov, Julia Bazhenova, and many other people.

**Conflicts of Interest:** The authors declare no conflict of interest.
