Assessment of a universal nucleation and crystal growth mechanism in undercooled alloys with orthorhombicCrB-type structure

Kurzfassung

Electrostatic (ESL) and electromagnetic (EML) levitation techniques were utilized to investigate the nucleation and crystal growth in binary intermetallic melts, that solidify in CrB-type structure. The high purity alloying and subsequent contact free processing allowed the systems to stay liquid at substantial under coolings below the equilibrium melting temperature. The subsequent nucleation event was observed insitu and the microstructure of the solid sample was analyzed by optical and electron microscopy. EBSD analysis was used to obtain a better understanding through crystal-structure evaluation on the mechanisms at play. Binary, intermetallic (A50B50) alloys were chosen that are supposed to solidify inthe orthorhombic CrB-type crystal structure. The present work bases on a specialnucleation and growth mechanism, found by Kobold in Ni50Zr50. Based on these observations, the Hornfeck-Kobold-Kolbe growth model was developed, which connects a quasi crystalline (QC) core structure with a ten-fold twinned microstructure. According to the model, other n-fold symmetries are possible in CrB-structures alloys, based on their respective lattice parameters. This work assesses a part of these CrB-structured alloys in order to show the universal applicability of this model to orthorhombic systems. Based on the prototype system of NiZr, crucial features are stated, that are necessary for nucleation and growth based on the model. Besides the singular nucleation event and the icosahedral based nucleus structure, especially the orthorhombic unit cell of the CrB-structure (B33 phase) is of importance. It was shown, that the lattice parameter ratio a/b is responsible for the n-fold symmetry, since the twinning boundaries of the model, creating ten differently orientated grains around a pointsymmetric center, run along the diagonal of the unit cell. In this work, systems were chosen, based on the expected n-fold symmetry andit will focus on the investigations of the ten-fold (Ni50Hf50, Ni50Zr25Hf25), nine-fold(Ni50Gd50) and eight-fold (Ni50B50) systems. It will give an analysis on whether their behavior can be described by the proposed growth model and its crucial features.Ni50Hf50 and Ni50Zr25Hf25 were chosen, since Hf and Zr are very similar and almost completely miscible. Additionally, a ten-fold growth was expected. Both systems meet the expectations as most of the crucial features are observed. These systems clearly show a singular nucleation event, the growth front can be observed in situ and connected with the front of the NiZr system. In both systems, a single growth direction is present, with all grains orientated around one common [001] direction. InNiHf however, no singular growth structure could be identified in the microstructureanalysis. Instead, several smaller point-symmetric structures were found, that resemble the symmetric core structure found in NiZr. A solid-solid structural transition was found to disturb the optimal growth. Several nucleation modes are proposed i this work, and described, depending on the exact under cooling and which phase is stable upon nucleation. Only one mode, with the initial nucleation in the orthorhombicB33 phase, and two subsequent solid-solid transitions, can be connected tothe growth model. With the latent heat released during nucleation, the system transitions into the high temperature (HT) B2 phase. Upon cooling, all samples transition (back) into the B33 phase. An initial nucleation in the cubic B2 phaseresults in different features, such as multiple growth directions. Despite the twotransitions following the initial B33 nucleation (B33->B2->B33) the microstructurecould preserve the singular growth direction and some of the symmetric structuresthat can be connected to tertiary dendrites of the growth model. In Ni50Zr25Hf25, this effect is not that heavily pronounced. The solid-solid transition is present and itdisturbs the microstructure. However, a singular underlying growth structure can still be identified in the microstructure. Other crucial features are met, according to the model, just as in NiHf.Ni50Gd50 was the first system with a predicted uneven (nine-fold) symmetry. As the model is based on an icosahedral core structure, disturbances in a stable growth were anticipated. But, despite only small undercoolings reached, the system exhibitsa singular common growth direction with several symmetric structures andthe expected grain boundary angle. Only incomplete symmetric structures couldbe found, though. Due to the low undercoolings, modifications had to be made, to explain this microstructure in accordance to the proposed model. A heterogeneous growth front propagates through the melt and symmetric structures are formed regularly due to stacking faults or impurities. These nuclei can act as nucleation sites forthe model’s QC core structure. As these structures have to compete in growth with the general heterogeneous growth front, they take a cylindrical shape and cannot grow through the whole of the sample. This can also explain the deformations andmissing orientations found. With Ni50B50, only small under coolings could be reached. It was however possible,to connect this system to the proposed model. A singular growth, throughout the whole sample, with the expected eight-fold symmetry was observed. The twinning boundaries exhibit the expected grain boundary angle, but are highly distorted and shifted and it was shown, that smaller symmetries shift the atomic arrangements, which, in NiZr, extend almost perfectly along twinning boundaries and impose therefore only a small energy barrier. Other systems were investigated with expected symmetries from 8- to 11-fold. They are shortly described in this work and evaluated. Some systems are promising andshould be investigated further, others show no relation to the growth model, e.g.when they solidify in the B2 structure. The proposed nucleation and growth model, based on the NiZr prototype system,was finally assessed, based on presence of the defined crucial features in the investigated systems. It is concluded, that the model can hold as a universal characteristic in the liquid-solid phase transition of CrB-structured alloys. Especially,the grain boundary angle with the twinning boundary running through the diagonal of the unit cell and the directly dependent n-fold symmetry are highly preserved. The growth in a single direction is another crucial feature, that is observed in basically every case. Despite this, it is often difficult to observe the perfect growth structure, as it was found in NiZr. It was shown to be highly dependent on factors, such as solid-solid transitions (Ni50Hf50), inner structural tensions (Ni50B50) or disturbances during the formation of the core structure (Ni50Gd50). Additionally, modifications were implemented to adapt the model to a broad range of nucleation events