When we could find a new compound, single crystal preparation is very important for identifying the new compound, determining crystal structure and measuring physical properties. We have world top class techniques of growing single crystals of non-oxide refractory compounds like rare earth hexaborides, transition metal borides and carbides. Their melting temperatures are from 2500°C to higher than 3000°C.

Is it easy to grow single crystals of many new rare earth boron-rich borides found by us? We should investigate their phase relations. For example, YB₂₅ and YB₅₀, whose discovery triggered further searching work of rare earth boron-rich borides, decompose at high temperature without melting as shown in the Y-B binary phase diagram (below). In principle melt growth like floating zone (FZ) method cannot be applied to such decomposition compounds.

Single crystals of most of B₁₂ cluster boron-rich borides have been grown by high temperature metal solution growth method using Al, Cu and so on. However, the method could supply maximal several mm size crystals.

We looked for a breakthrough that makes it possible to grow single crystals of rare earth boron-rich borides by FZ method. The important point aimed at is a small difference between the decomposition temperature and the melting temperature. The difference is only 200°C for YB₂₅ and 250°C for YB₅₀, respectively. If we can reduce the melting temperature slightly, they may melt without decomposing. The solution is addition of Si that acts to reduce the melting temperature of borides. We tried the floating zone crystal growth of YB₂₅ and YB₅₀ by adding a small amount of Si.

The melting temperature of rare earth boron-rich borides is around 2000°C. RF induction heating is not suitable for them because they are semiconductors. We are using an image furnace. A schematic drawing of FZ crystal growth using a 4 Xenon lamp-ellipsoidal mirror image furnace is shown below. 4 pairs of Xenon lamp-ellipsoidal mirror are diagonally arranged in order to uniform temperature along the circumference of the crystal. Raising the temperature at the focus point of the Xenon lamp image forms the molten zone. A raw rod is fed into the upper part of the molten zone and a single crystal growing from the bottom part of the molten zone is pulled downwards.

Now, by the Si addition FZ method we could grow a single crystal of YB₄₄Si_1.0 that belongs to the same structure type as YB₅₀. On the other hand, the YB₂₅ phase was eliminated by the Si addition. FZ single crystal growth of REB₅₀Si and ScB₁₉Si successfully followed to YB₅₀Si. A photograph of a grown ScB₁₉Si_y is shown. The last half part (right hand side) is a single crystal because a seed crystal was not used.

It is necessary to confirm that the grown YB₄₄Si_1.0 crystal really belongs to the same structure type as YB₅₀. Powder XRD pattern of the single crystal YB₄₄Si_1.0 is compared with that of YB₅₀. Both show the same pattern, which confirmed that they belong the same structure type.

In the ternary Sc-B-C system, the Si addition method could be applied to the ScB₁₅C_0.8 phase. The chemical composition ScB_12.7C_0.62Si_0.08 (measured by EPMA) of the grown crystal is different from that synthesized by solid-state reaction method. The crystal has a face centered cubic structure. It is interesting that two new phases whose chemical compositions are very close to that of the cubic phase were found during the crystal growth investigation. One whose chemical composition is ScB_11.7C_0.6Si_0.04 has a hexagonal crystal structure and another whose chemical composition is ScB_12.8C_0.7Si_0.004 has an orthorhombic crystal structure, respectively.

Although single crystals of the silicon-rich REB_17.6Si_4.6 phases could be grown from the Si flux using the high temperature solution growth method, the FZ single crystal growth was unsuccessful. It seems to be difficult to keep a chemical composition in the Si flux within such a narrow molten zone region.