**4. Discussion**

Summing up the results, the Al-6Ca-3Mn, Al-8Ca-2Mn and Al-8Ca-2Mn-1Ni alloys may be reasonably considered as a promising basis for developing novel alloys for special applications as an alternative to the hypereutectic Al-Si alloys. Al-Ca alloys show the same fraction of primary crystals as those in the FM180 alloy (~5 vol.%) but have more favourable phase transformation temperatures. As follows from the thermal analysis (Figure 10), the Al-6Ca-3Mn and Al-8Ca-2Mn alloys have a higher solidus temperature as compared to the FM180 alloy. This feature implies the opportunity for conducting the annealing of the ingots at a higher temperature and higher thermal stability in the exploiting conditions. The results obtained by direct and differential thermal analysis appeared to be virtually similar.

**Figure 10.** Results of the differential thermal analysis and direct thermal analysis on the: (**a**) Al-6Ca-3Mn; (**b**) Al-8Ca-2Mn; (**c**) FM180 alloys.

Since the Thermo-Calc software does not take into account the formation of ternary compounds in the Al-rich angle, the experimental Al-Ca-Mn-Ni liquidus projection should be different from the calculated one. Hence, in addition to the ternary Al10Ca(Mn,Ni)2 phase which appeared in the alloys containing more than 1%Ni, there is a ternary compound Al8Ca(Ni, Mn)2 based on the Al8CaNi2 phase. However, the Al9CaNi compound formed in the Al-Ca-Ni-La alloys is indicated in [12]. In the handbook [24], the presence of both Al8CaNi2 and Al9CaNi phases is alleged in the Al-rich angle of the Al-Ca-Ni ternary system. Having conducted the current study, we can propose the appearance of the Al-Ca-Mn-Ni diagram in the Al-rich angle (Figure 11). Nevertheless, the existence of two ternary compounds Al8CaNi2 and Al9CaNi requires experimental verification. According to the hypothetical view of the phase diagram, there are five primary crystallisation regions: Al4Ca+L; Al3Ni+L; Al6Mn+L; Al10CaMn2+L; Al8CaNi2+L. As is shown in the work [25,26], the T (Al16Mn3Ni) compound indicated in the diagram is present in the ternary Al-Mn-Ni system. Moreover, three invariant reactions including two peritectic L+Al3Ni→(Al)+Al6Mn+Al8CaNi2 (point P1); L+ Al6Mn→(Al)+Al10Mn2Ca+Al8CaNi2 (point P2) and one eutectic L→(Al)+Al4Ca+Al10Mn2Ca+Al8CaNi2 (point E) are drawn.

**Figure 11.** Hypothetical distribution of the solid phase fields in the Al-Ca-Mn-Ni system.

The Ca-bearing alloys are better in deformation than the Al-Si piston alloy due to the remarkable fineness of the eutectic structure. Poor processability of the Al-8Ca-2Mn- (2-4)Ni may be associated with a high number of the Al8CaNi2 phase. The latter, despite having a compact shape and small size of 40 μm, appears to act as a stress raiser causing brittle interfacial fracture (Figure 12). In contrast, the Al10Ca (Mn, Ni)2 crystals, likely less strong, break in a fragile manner simultaneously with the matrix that ensures fewer stresses. Notwithstanding the cracks that appeared at the very beginning of the rolling, the Al-8Ca-2Mn-(2-4)Ni alloys were successful in a way similar to the FM180 alloy. That is stipulated by the superfine ductile matrix surrounding the brittle Al8CaNi2 crystals.

**Figure 12.** Fracture surfaces of the hot-rolled samples: (**a**) Al-8Ca-2Mn-1Ni; (**b**) Al-8Ca-2Mn-2Ni.

Thus, we can reasonably conclude that, despite the promising natural composite structure of the Al-8Ca-2Mn-(2-4)Ni alloys, they are not feasible in hot deformation due to the presence of the high fraction of brittle compound. Therefore, the stiffness of the primary crystals should be comparable to that of the matrix (surrounding eutectic structure) that may be noted as a requirement to the structure of the new Al-Ca hypereutectic natural composites. With a noticeable difference in strength between matrix and primary crystals, even if the latter has a small size and uniform distribution within the bulk, they appear to become places where brittle deformation occurs. Summing up, the Al10Mn2Ca crystals mostly meet the foregoing requirements.
