Figure 1 shows XPS spectra of MoSi₂ and (Mo_0.9Cr_0.1)Si₂ silicide feedstock at the Si_2p, Mo_3d and Cr_3d. The area ratios of Si_2p peaks at SiO₂ and MoSi₂ are 2.9 and 1.0 for the MoSi₂ and (Mo_0.9Cr_0.1)Si₂ powder surfaces, respectively (Figure 1(a)). The area ratios of Mo_3d peaks at Mo-oxides and MoSi₂ are 1.7 and 1.0 in the MoSi₂ and (Mo_0.9Cr_0.1)Si₂ powder surfaces, respectively (Figure 1(b)). The Cr_2p peak of Cr₂O₃ was found in the (Mo_0.9Cr_0.1)Si₂ powder surface, and the area ratio of Cr₂O₃ to Cr-silicide is 1.4 (Figure 1(c)). It is apparent that the formation of silica and Mo-oxides on the (Mo_0.9Cr_0.1)Si₂ powder surface are substantially impeded by the formation of Cr-oxides. The oxygen in the sintered bulk is believed to be introduced from the native oxide on the stocking powders. According to the chemical analysis, the oxygen contents were 2.16 and 1.13 wt% in the MoSi₂ and (Mo_0.9Cr_0.1)Si₂ sintered bulks. Meanwhile, substantially decreasing of SiO₂ particles was observed in the sintered (Mo_0.9Cr_0.1)Si₂ and composite bulks (see the following section). It was consistently testified that Cr-addition can efficiently reduce silica in molybdenum disilicide. It is widely agreed that low silica at grain boundaries is appreciable to the creep resistance of silicide composites [5]. Thus, Cr-alloying is an alternative approach to reduce the SiO₂ contents of silicide composites, which could contribute to creeping property at elevated temperature.

Theoretical density of MoSi₂ and (Mo_0.9Cr_0.1)Si₂ silicides are 6.25 and 6.10 g/cm³, respectively, measured from XRD. Theoretical densities of MoSi₂ + 15vol%ZrO₂ and (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ are 6.16 and 6.04 g/cm³, calculated from a linear combination rule. The sintering density of different materials is shown in Table 1. High sintering density over 98% T.D. was obtained in both Cr-alloyed silicide and composite. Improved sinterability is directly related to the existence of (Cr, Mo)₅Si₃ type silicide as secondary phase presenting (see the following

section). The melting point of CrSi₂ is 1551˚C, nearly 470˚C lower than that of MoSi₂. According to a quasibinary section phase diagram of MoSi₂-CrSi₂, a peritectic reaction exists at 7 at%Cr at 1529˚C, taking place as L + Mo_1?xCr_xSi₂ Û Mo_yCr_1?ySi₂ [6]. A liquid phase might appear at the sintering temperature, which may greatly assist mass transport and densification process during sintering at 1600˚C - 1620˚C.

The XRD analysis shows that the (Mo_0.9Cr_0.1)Si₂ consists of a single-phase of tetragonal C11_b structure. The Cr solid solubility in the (Mo_0.9Cr_0.1)Si₂ is 2.5 - 2.6 at% by EDS, which is lower than the nominal composition of 3.3 at%. Typical microstructures of different materials are shown in Figure 2. The amount of SiO₂ particles in (Mo_0.9Cr_0.1)Si₂ was reduced to 1.4 vol% compared with 2.2 vol% in MoSi₂, measured from selected dense areas with imaging process. A small amount of Mo₅Si₃ was observed in MoS₂ (Figure 2(a)). In (Mo_0.9Cr_0.1)Si₂, the Mo₅Si₃ phase was replaced by a Cr-rich silicide with a composition of Cr:Mo:Si = 45:15:40 (at%) close to (Cr_0.75Mo_0.25)₅Si₃. The Cr-rich silicide appears as discrete grains and continuous phase along grain boundaries (Figure 2(b)). Grain growth is substantial in (Mo_0.9M_0.1)Si₂ by a factor of 3 of that in MoSi₂ (Table 2), which could be ascribed to the liquid phase sintering effect of the existence of the (Cr, Mo)₅Si₃ silicide with lower melting point. ZrO₂ particles were well-dispersed in both composites (Figure 3(c) and Figure 3(d)), and grain sizes in both composites were refined owing to the dispersion of ZrO₂ particles (Table 2). The Mo₅Si₃ grains are still visible in MoSi₂ + 15vol%ZrO₂ (Figure 2(c)), but the (Cr ,Mo)₅Si₃ is hardly recognized in (Mo_0.9M_0.1)Si₂ + 15vol%ZrO₂ composite from the contrast in SEM images (Figure 2(d)).

Mechanical properties are shown in Table 1 and Table 2. The average hardness values of the (Mo_0.9Cr_0.1)Si₂ matrix phase and (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ composite are 10% lower than their un-Cr alloyed counterparts (Table 1), which might result from the softening effect of the Cr-addition. The fracture toughness K_IC of the two composites are comparable, which are higher than the monolithic silicides in a factor of 1.4 - 1.6 (Table 2). The fracture toughness values from indentation fracture (K_c) and SEVBN (K_IC) methods are close in MoSi₂ + 15vol%ZrO₂. However, the K_c value is too high over K_IC in the (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ composite, which

is caused by the uncertainty in determination of the crack length. Figure 3 shows typical indentation cracks in the composites. The primary median cracks from the indent corners in MoSi₂ + 15vol%ZrO₂ are deflected (D), branched (Bra), bridged (Bri) by the dispersed ZrO₂ additives (Figure 3(a)), which dissipates and absorbs the energy for crack propagation. In contrast, multi-cracks are developed around the indent edges in (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ (Figure 3(b)). In such a case, the measured crack length is not reliable and inaccurate, which results in the overestimated K_c result. It is implied that the indentation fracture method is not suitable to describe the fracture toughness property of the composite (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂, therefore, a verification with a standard method has to be considered. The room temperature flexural strength s_f of the composites is lower than that of MoSi₂ (Table 2). The s_f of (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂is substantially reduced at 1200˚C compared with MoSi₂ + 15vol%ZrO₂. It is believed that interfacial fracture energy is changed between the alloyed silicide matrix and ZrO₂ particles [7], and grain boundaries can be weakened by the present of the (Cr, Mo)₅Si₃ phase in (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ at elevated temperatures.

Figure 4 shows time dependence of mass change after isothermal exposure at 1400˚C in air. All the samples present with parabolic kinetics in the steady oxidation stage, indicating the formation of protective oxide scales and diffusion controlled oxidation. Total weight gains in (Mo_0.9Cr_0.1)Si₂ and MoSi₂ are 1.7 and 3.6 mg/cm² after 1000 h exposure, inferring that the thickness of the SiO₂ layer on the Cr-alloyed silicide (Mo_0.9Cr_0.1)Si₂ is thinner than that on MoSi₂. The performance of (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ composite is similar to that of monolithic

MoSi₂. In contrast, an erratic weight loss occurred in un-alloyed composite MoSi₂ + 15vol%ZrO₂in the initial oxidation stage within 50 h. A parabolic weight-gain in MoSi₂ + 15vol%ZrO₂ is established when a continuous silica layer is formed. The rate constants in the steady oxidation stage (after 50 h) are 0.0014 and 0.0006 (mg²∙cm⁻⁴∙h⁻¹) in MoSi₂ + 15vol%ZrO₂ and (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂, respectively. Obviously, the oxidation rate is substantially decreased in the Cr-alloyed composite.

The oxide scales on both composites consist of tetragonal ZrSiO₄ and α-SiO₂ by XRD. Typical morphologies of the composite surfaces after 1000 h exposure are shown in Figure 5. The scale on MoSi₂ + 15vol%ZrO₂ surface contains large amounts of smaller zircon particles with a size of 1 - 2 mm (Figure 5(a)). In contrast, the surface of (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ is comprised zircon particles with a size of 3 - 5 mm (Figure 5(c)). Scale thickness is between 50 - 60 and 20 - 30 µm in MoSi₂ + 15vol%ZrO₂ and (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂, respectively. SEM cross sectional images show that Mo₅Si₃ grains are located at the interface between the base silicide and oxide scale of MoSi₂ + 15vol%ZrO₂ (Figure 5(b)). The fine zircon particle network in the scale of MoSi₂ + 15vol%ZrO₂ can provide fast diffusion paths for oxygen to the interface of scale-base silicide, which results in higher oxidation rate. In comparison, the oxide scale of (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ consists of nearly pure silica. Some large zircon particles mainly appear on the outermost surface of this composite. A few small discrete Cr₂O₃ grains were found at the interface (Figure 5(d)), which resulted from the outer diffusion and oxidation of Cr from the alloyed base silicide after the formation of the protective silica scale.

It is known that the formation of protective silica on MoSi₂ at high temperature surface is implemented by selective oxidation of Si: MoSi₂ + O₂ ® MoO₃(g) + SiO₂ at high pO₂; and MoSi₂ + O₂(g) ® Mo₅Si₃ + SiO₂, at low pO₂. In the latter case, and silica can form at surface and Mo₅Si₃ silicide usually form at the interface between scale and base silicide as oxygen diffuses through the surface oxide layer. In presence of ZrO₂ additives, SiO₂ will react with ZrO₂ to form zircon: SiO₂ +ZrO₂ = ZrSiO₄, thus, the resultant zircon particles are retained in

the oxide scale. The improved oxidation behaviour of the Cr- alloyed composite can be attributed to the chemistry change at the as received surfaces sintered in reduced atmosphere. It was reported that as-sintered surface of PLSed MoSi₂ + 15vol%ZrO₂ is comprised with a silicon depleted silicide (Mo-Zr-Si) [8]. The substantial weight loss presenting in initial oxidation occurred in this composite was ascribed to the rapid oxidation of the high Zr-content silicide in the outermost surface. In comparison, the PLS sintered (Mo_0.9Cr_0.1)Si₂ + 15vol%ZrO₂ surface consists of Cr-rich silicide with a low Zr content. It is believed that this silicide in the as-sintered surface could assists the formation of SiO₂ scale with a favourable microstructure (showed in Figure 5(d)) at relative lower temperature. The outer diffusion of Cr can also promote the mobility and diffusion of Si at high temperature in the Cr-alloyed composite. The investigation details will be published in a later paper.