Yttrium metal powder (purity > 99.9%, under 20 mesh; RARE METALLIC Co., Ltd.) supplied as paraffin slurry was separated by vacuum filtration and washed several times with hexane. Silicon powder (purity > 98.0%, under 100 mesh; Oxygen content = 1.03 wt%; Fuji film Wako Pure Chemical Corp.) was pulverized by using a rotary ball mill at 118 rpm for 8 hrs using 5 mmf Si₃N₄ grinding media to give the Si powder having mean particle size of 8.3 μm and oxygen content = 1.17 wt%. These powders were mixed with a composition of Y:Si = 1:2 in molar ratios, and uniaxially pressed at 30 MPa in a 15 mm diameter stainless-steel die. The powder compact was placed in a quartz tube (inner diameter = 10 mm and length = 10 cm), which was evacuated using a rotary pump, and then was sealed using a H₂-O₂ gas burner. The quartz tube was heat treated at 1100˚C for 5 hr using a muffle furnace (Model MPC10; ADVANTEC Co., Ltd.) using the heating rate of 10˚C/min.

The same powder compact was placed in a high purity BN crucible (N-1 grade; DENKA Company Ltd.), which was heat treated in a graphite susceptor at 1090˚C for 5 hrs or 1500˚C for 1 hr using a vacuum furnace with a graphite heating element (Model FVPHP-R-3: FUJIDENPA KOGYO Co., Ltd.). Heating and cooling rates were both 10˚C/min.

A Yttrium metal ingot (purity > 99.9%, 7 - 10 mm; RARE METALLIC Co., Ltd.) was surrounded by Si powder (Purity > 99.99%, Oxygen content = 0.081 wt%, 300 μm pass; KOJUNDO CHEMICAL LABORATORY Co., Ltd.) using the molar ratio of Y:Si = 1:2 in a high purity BN crucible (N-1 grade; DENKA Company Ltd.), which was heat treated in a graphite susceptor at 1600˚C for 30 min under Ar atmosphere of 0.1 MPa using a furnace with a graphite heating element. Heating and cooling rates were both 10˚C/min.

A small amount of pulverized YSi₂ powder was immersed to various solvents including purified water, methanol (>99.8%), ethanol (>99.5%), propanol (>99.7%), butanol (>99.5%), benzene (>99.5%), n-hexane (>96.0%), or acetonitrile (>99.5%) in a test tube, and was settled for 24 hrs at room temperature under the atmospheric condition observing every hour whether the formation of bubbles or white precipitates occurred or not.

Low purity YSi₂ powder synthesized at 1500˚C for 1 hr and high purity YSi₂ powder synthesized at 1600˚C for 30 min were ground for 12 hrs by using a rotary ball mill at 120 rpm in various solvents including 2-propanol, n-hexane, or acetonitrile using 5 mmf Si₃N₄ grinding media at room temperature under the atmospheric condition. After milling, each solvent was removed using a rotary evaporator, dried in an oven at 120˚C for 6 hrs, and sieved by passing through a 60 mesh nylon sieve. The oxygen content of each powder was determined by the hot-gas extraction method (EMGA-920, HORIBA Ltd.) using 20 mg powder sample encapsulated with 1.0 g pure iron in a Ni crucible.

Table 1 summarizes the fabrication conditions of Y-Si compounds with Y:Si = 1:2 composition indicating five powder samples as No. 1-5. Figure 2 shows the XRD patterns of powder samples No. 1-4. Since the XRD pattern of sample No. 5 was found to be identical to that of No. 4, Figure 2 lacks its XRD pattern. Table 2 summarizes the XRD identification results of Y-Si compounds prepared in this work. Table 3 and Table 4 compare powder XRD peak data obtained in this work having 1) hexagonal and 2) orthorhombic, respectively, crystal structures with those of Powder Diffraction File (PDF) database registered by the International Centre for Diffraction Data (ICDD). As shown in Table 2, Y-Si compounds with hexagonal structure were identified as α-YSi₂ or β-Y₃Si₅, because both phases are indistinguishable with their XRD patterns as shown in Table 3.

It was reported that α-YSi₂ and β-Y₃Si₅ had the identical crystal structure, P6/mmm, and β-Y₃Si₅ had defect structure: Si atoms occupy 83% of the 2(d) positions in the space group P6/mmm [13]. Also, the orthorhombic phase was identified as β-YSi₂, because it was reported that the low temperature phase α-Y₃Si₅ (Ortho.) was stable below 450˚C [9]. Apparently, only sample No.2 has the orthorhombic structure, and the other samples possess the hexagonal structure as shown in Table 3 and Table 4. Samples No. 1 and 2 were fabricated with low purity metal powders at virtually the same temperature around 1100˚C, at which the low temperature phase, α-YSi₂, should be stable according to the latest phase diagram [1]. However, sample No. 2 fabricated at 1090˚C in a graphite furnace under vacuum condition (continuously evacuated using a rotary pomp) possesses the orthorhombic structure, the high temperature phase stable above 1250˚C. It would be noteworthy that sample No. 3 that was fabricated at 1500˚C in the same conditions other than temperature as sample No. 2 possesses the hexagonal structure, the low temperature phase stable below 1250˚C. This apparent contradiction would be explained as follows; Samples No. 1-3 were fabricated using Yttrium powder, which should contain a significant amount of surface oxygen, although the oxygen content was not accounted for its purity, 99.9% that accounted only for the total metal impurity. Because sample No.1 was fabricated in a sealed quartz tube (closed system), the vaporization of SiO according to Equation (1) would be suppressed. This speculation would be supported by the fact that weak peaks of Y₂O₃ are observed only in the sample No.1, which also shows middle intensities of Si peaks suggesting that Equation (1) be suppressed. However, when heat-treated under vacuum (open system), the amount of Si and Y₂O₃ would be reduced according to Equations (1) and (2), respectively, resulting the formation of unstable phase with orthorhombic structure [4] as observed for sample No. 2. When heat-treated at higher temperature, 1500˚C, this unstable Y₂Si₃O_x phase would decompose to the stable phases, β-Y₃Si₅ and Y₂SiO₅ and/or Y₂Si₂O₇ as shown in Table 2.

Intensity: vs: very strong; w: weak; vw: very weak; tr: trace; ―: not detected. *1: It was reported that α-YSi₂ (No. 11-0596 in the PDF-ICDD) and β-YSi₂ (No. 11-0324 in the PDF-ICDD) were the low and high temperature phases, respectively, and the phase transformation occurred at 1250˚C [3]. *2: Since it was reported that the low temperature phase α-Y₃Si₅ (Ortho.) was stable below 450˚C [9], the high temperature phase β-Y₃Si₅ (No. 72-2173 in the PDF-ICDD) would be the stable phase in the current experimental condition. It was also reported that α-YSi₂ and β-Y₃Si₅ had the identical crystal structure, P6/mmm, and β-Y₃Si₅ had defect structure: Si atoms occupy 83% of the 2(d) positions in the space group P6/mmm [13].

When high purity metal powders were used, only diffraction peaks attributed to the hexagonal phase and Si are observed for sample No. 4 and almost no other oxide phases exist. A fact that the stoichiometric mixture of Y:Si = 1:2 resulted in the formation of Si suggests that the hexagonal phase would not be α-YSi₂ but β-Y₃Si₅, which would be further supported by the fact that the sample No. 4 was heat-treated at as high as 1600˚C, at which this mixture should form a homogeneous liquid phase according to the Y-Si phase diagram [1]. To confirm whether the mixture of β-Y₃Si₅ and Si would be the stable phases for the stoichiometric mixture of Y:Si = 1:2, sample No.4 was further heat-treated at 1100˚C for 12 hrs, which was specified as sample No. 5 in Table 1. This equilibration process did not change the diffraction peaks observed for sample No. 4 at all, supporting the hypothesis above mentioned. A recent study regarding the fabrication of YSi₂ by the spark plasma sintering at 1273 K under Ar atmosphere using high purity metals also demonstrated that the main peaks of all samples agreed well with the YSi_1.67 standard pattern (JCPDS 01-071-3919), representing the single phase of the AlB₂-type structure (P6/mmm) [12]. It would be interesting that the 2θ values of all diffraction peaks of the β-Y₃Si₅ hexagonal lattice shifted in the order of sample No. 1 > No. 3 > No. 4 as shown in Table 3, suggesting that their lattice constants would be enlarged with the amount of oxygen impurity.

As described in the introduction, although the latest Y-Si phase diagram shown in Figure 1 [1] demonstrated that the α- and β-YSi₂ phases were the stable phases for the stoichiometric composition of Y:Si = 1:2 adopting the experimental results clearly showing their existences [6] [7]. However, a detailed investigation of phase equilibria and crystal structures of the compounds taking the effects of impurity oxygen into account demonstrated that the formation of YSi₂ was not confirmed as a stable phase [4]. As a result, there have been several phase diagrams for the Y-Si system. The current experimental results suggest that the high temperature phase with the hexagonal structure, β-Y₃Si₅, would be the stable phase even for the stoichiometric composition of Y:Si = 1:2, and that the high temperature phase with the orthorhombic structure, β-YSi₂, would be the meta-stable phase under high oxygen impurity content. Further investigation would be necessary to clarify the phase diagram of Y-Si system.

Figure 3 and Figure 4 show the appearances of α type and β type YSi₂ powders (sample No. 2 and 3), respectively, immersed in purified water, methanol, ethanol, propanol, butanol, benzene, n-hexane, or acetonitrile (from the left end to the right end) settled for 0, 1, 2, 4, 8, 16 and 24 hrs at room temperature under the atmospheric condition. It was observed that the dispersibility increased in the order as follows, water < methanol < ethanol < 2-propanol > butanol ~ acetonitrile >> benzene ~ n-hexane, demonstrating that the best solvent is propanol for dispersing YSi₂ powders. There was almost no difference between α type and β type YSi₂ powders, but β type showed better dispersibility after long duration, which would be due to the difference of mean particle sizes of both powders, 17 μm and 10 μm, respectively. It should be noted that neither the formation of bubbles nor white precipitates was observed in all solvents even after 24 hrs for both powders, suggesting much superior chemical stability to solvents of YSi₂ powders than Yttrium metal and Alkaline Earth Silicides, for which the formation of bubbles and/or white precipitates were observed in all solvents immediately after dispersing these powders.

Table 5 summarizes the oxygen contents of YSi₂ powders with low and high purity, sample No. 3 and 4 respectively, before and after grinding in 2-propanol, n-hexane or acetonitrile at room temperature under the atmospheric condition. It was confirmed that the use of high purity metals resulted in reducing oxygen content by about 58% comparing to the low purity metal powders. After grinding these powders in a ball mill at 120 rpm for 12 hrs using 2-propanol, n-hexane, or acetonitrile as solvents, the oxygen contents increased significantly probably due to the mechano-chemical effect. Although those solvents without hydroxy group, n-hexane and acetonitrile, were expected not to increase the oxygen content, there was almost no difference for oxygen contents after milling in all solvents for powder sample No. 3 (low purity). However, for powder sample No. 4 (high purity), n-hexane gave the least oxygen content probably due to its less contents of oxides such as Y₂SiO₅ and/or Y₂Si₂O₇ as shown in Table 2.