Crystal Growth from Undercooled Melt

At the hypercooling limit, the whole melt grows below Tm [15], unlike the general case where the remaining melt grows isothermally at Tm- AThyp for, e.g., Ni, a closely packed metal, is 446 K, while it is 1,977K for Si because of the large enthalpy of fusion, AHf. Therefore, it is not realistic that the Si melt undercools beyond AThyp. 

In summary, compared with the single crystal ingot grown at a constant growth rate by the CZ method, the sample crystallized from the 

In the commercial application, the drop tube method, as mentioned above, is suitable for mass production. The detailed investigations, such as temperature measurement of each small droplet and in situ observation of microstructure formation are not easy to attain because each droplet is in free fall. Here, the levitation method, where an Si droplet with a diameter of mm can be levitated by electromagnetic force using an electro-magnetic levitator (EML), as shown in Fig. 8.5, is a powerful investigation technique because the controlled droplet position enables us to measure the surface temperature of the droplet by pyrometer and to observe the crystallization behavior in situ by a high-speed video camera (HSV) [16-18]. 

---

Fig. 5.3. Growth process of Si crystal from undercooled melt [16]. (a) Crystal is growing with faceted interface, (b)-(d) Faceted dendrite grows from a part of faceted interface. The direction of the growth of faceted dendrite is parallel to the 111 facet face on the interface [16]...

The process of crystallization from solution is generally understood in terms of three interdependent conditions supersaturation, nucleation and growth. For crystallization from the melt or the gas phase, supersaturation is simply interpreted as undercooling. Of the essence is a metastable phase, at chemical potential higher than that of the crystalline phase under the same conditions. The transition from the metastable state to the state of equilibrium between phases corresponds to nucleation. 

The presence of HOCP considerably slows down the melt crystallization process of PB-1. Therefore, the adopted values, lowered by increasing the HOCP fraction, provided similar rates of crystallization for pure PB-1 and blends. Previous calculations from the spherulite growth rate and from the overall kinetic rate constant showed that the number of nuclei per unit volume was similar for samples crystallized at equal undercoolings. Had we used a constant value of T, there would have... 

---