Metallothermic Reduction of MoO3 on Combustion Synthesis of Molybdenum Silicides/MgAl2O4 Composites

Combustion synthesis involving metallothermic reduction of MoO3 by dual reductants, Mg and Al, to enhance the reaction exothermicity was applied for the in situ production of Mo3Si–, Mo5Si3− and MoSi2–MgAl2O4 composites with a broad compositional range. Reduction of MoO3 by Mg and Al is highly exothermic and produces MgO and Al2O3 as precursors of MgAl2O4. Molybdenum silicides are synthesized from the reactions of Si with both reduced and elemental Mo. Experimental evidence indicated that the reaction proceeded as self-propagating high-temperature synthesis (SHS) and the increase in silicide content weakened the exothermicity of the overall reaction, and therefore, lowered combustion front temperature and velocity. The XRD analysis indicated that Mo3Si–, Mo5Si3– and MoSi2–MgAl2O4 composites were well produced with only trivial amounts of secondary silicides. Based on SEM and EDS examinations, the morphology of synthesized composites exhibited dense and connecting MgAl2O4 crystals and micro-sized silicide particles, which were distributed over or embedded in the large MgAl2O4 crystals.

As an alternative, metallothermic reduction reactions (MRRs) of metal oxides with Mg and Al as reducing agents produce MgO and Al 2 O 3 as precursors for the formation of MgAl 2 O 4 and such oxidation reactions are highly exothermic [13,14]. When combining Mg/Al-based MRRs with combustion synthesis, such a fabrication route is effective in producing MgAl 2 O 4 -containing composites. Moreover, the highly-exothermic MRRs render reduction-based combustion synthesis fit for self-propagating high-temperature synthesis (SHS). Many merits such as high energy efficiency, short reaction time, simplicity of operation and high-purity products have been recognized for the SHS process [15][16][17]. According to Horvitz and Gotman [18], reduction-based combustion synthesis using 2TiO 2 -Mg-4Al samples was performed to produce TiAl-Ti 3 Al-MgAl 2 O 4 composites. Omran et al. [19] conducted co-reduction of WO 3 [20] obtained MgAl 2 O 4 composites with MoSi 2 and Mo 5 Si 3 from co-reduction of SiO 2 and MoO 3 by Al in argon at a pressure of 5 MPa. The high Ar pressure was to suppress the volatilization of MoO 3 . Recently, Radishevskaya et al. [21] synthesized MgAl 2 O 4 by the SHS method using the reactant mixtures consisting of MgO and Al 2 O 3 , along with Al as the fuel, Mg(NO 3 ) 2 ·H 2 O as the oxidizer, and NaCl as the mineralizer. Results indicated that NaCl of 1 wt.% contributed to the completion of the formation of MgAl 2 O 4 and mechanical activation of the green mixture for 60 s facilitated the production of MgAl 2 O 4 without oxide impurities.
By using Mg and Al simultaneously as dual reductants, this work aims at investigating the in situ production of MgAl 2 O 4 -containing molybdenum silicide (Mo 3 Si, Mo 5 Si 3 , and MoSi 2 ) composites by the SHS process with reducing stages. That is, a solid-state combustion reaction involves the synthesis of MgAl 2 O 4 from the metallothermic reduction of MoO 3 and the formation of molybdenum silicides from elemental interactions between Mo and Si. Three different silicide phases were produced and their influence on reaction exothermicity and combustion wave kinetics was explored. Compositional and microstructural analyses were performed on the final composites. Moreover, some products were selected for Vickers hardness and fracture toughness measurements.

Materials and Methods
The raw materials utilized by this study include MoO 3 13 12 where stoichiometric coefficients x, y, and z are associated with the quantities of Mo and Si powders in the green mixtures, and also represent the molar proportion of silicide phase to MgAl 2 O 4 . The same composition of metallothermic reagents of 4/3MoO 3 + Mg + 2Al is adopted in R(1) and R(2), but R(3) has a different metallothermic mixture of 13/12MoO 3 + Mg + 3/2Al because R(3) comprises pre-added Al 2 O 3 . Because of metallothermic reduction of MoO 3 , the source of Mo for the formation of molybdenum silicides (Mo 3 Si, Mo 5 Si 3 , and MoSi 2 ) from R(1), R(2), and R(3) included both reduced and elemental Mo. It has been realized that magnesiothermic and aluminothermic reductions of MoO 3 are highly exothermic and have an adiabatic temperature (T ad ) exceeding 4200 K [22], which plays an important role in facilitating self-sustaining combustion for R(1), R(2) and R (3). When compared with the reduction of MoO 3 by Mg and Al, the formation reactions of Mo 3 Si, Mo 5 Si 3 and MoSi 2 are much less energetic. Among three molybdenum silicides, MoSi 2 is the most exothermic phase to form [23], and therefore, Al 2 O 3 at one-quarter of the required amount was added in the starting mixture to regulate the degree of violence of combustion.
Experimental ranges of x, y, and z conducted in this study were determined based on the reaction exothermicity of R(1), R(2) and R(3), which was assessed by computing T ad as a function of stoichiometric coefficients according to the following energy balance equation [24,25] with thermochemical data taken from [23].
∆H r + T ad 298 ∑ n j C p P j dT + ∑ 298−T ad n j L P j = 0 (4) where ∆H r is the enthalpy of reaction at 298 K, n j is the stoichiometric coefficient, C p and L are the specific heat and latent heat, respectively, and P j refers to the product.
The SHS experiment was performed in a windowed combustion chamber filled with high-purity argon (99.99%) at 0.2 MPa. Reactant powders were dry mixed and then were uniaxially pressed to form cylindrical test specimens with 12 mm in height, 7 mm in diameter, and 55% in the relative density. In this work, a cylindrical bottle partially filled with the raw materials and alumina (Al 2 O 3 ) grinding balls rotated about the longitudinal axis of a tumbler ball mill machine for 8 h to fully blend the reactant powders. The size of the alumina ball is 5 mm in diameter. The ball mill operated at 90 rpm. Because Al 2 O 3 is one of the precursors to form MgAl 2 O 4 , no contamination from grinding balls was detected.
The combustion wave propagation velocity (V f ) was determined from the time series of recorded combustion videos. The combustion temperature was measured by a 125 µm bead-sized thermocouple with an alloy composition of Pt/Pt-13%Rh. Details of the experimental setup were previously reported [25,26]. Phase components of the synthesized products were identified by an X-ray diffractometer with CuK α radiation (Bruker D2 Phaser, Billerica, MA, USA). Analyses of scanning electron microscopy (SEM) (Hitachi S3000H, Tokyo, Japan) and energy dispersive spectroscopy (EDS) were performed to examine the fracture surface microstructure and composition ratio of elements of the final products.
Measurement of Vickers hardness and fracture toughness of the products was performed [27]. For such measurements, only selected experiments under stoichiometric coefficients of x = y = z = 2 were carried out by placing the sample compact in a stainlesssteel mold. Densification of the product was conducted by a hydraulic compressor. Upon the completion of the SHS reaction, the burned sample was rapidly pressed when the product was still hot and plastic, which was held for about 15 s. The product density after compression reached about 93-95% of theoretical density and then the product surface was polished for the measurement. Microhardness was measured with a Buehler Micromet microhardness tester at a load of 1000 g and a dwelling time of 10 s. Five indentations were made to obtain the average values of the indentation imprint and crack length measurements.
In this study, Vickers hardness (H v ) was calculated from the applied load (P) and the average diagonal impression length (d) in the equation below [28,29]. The fracture toughness (K IC ) was determined by the indentation method using the following equation proposed by Evans and Charles [29]. where a is the half of the average length of two diagonals of the indentation and c the radial crack length measured from the center of the indentation.

Combustion Exothermicity of Reactions
Calculated values of T ad of R(1), R(2) and R(3) as a function of their respective stoichiometric coefficients are presented in Figure 1 in order to evaluate combustion exothermicity. A significant decrease in T ad with increasing silicide content is observed for all three synthesis reactions, mainly because the formation of molybdenum silicides is much less exothermic than the metallothermic reduction of MoO 3 . As revealed in Figure 1, the value of T ad associated with the formation of Mo 3 Si-MgAl 2 O 4 composites from R(1) decreases considerably from 3964 • C to 2415 • C as the coefficient x increases from 1 to 5. On account of a large heat capacity for Mo 5 Si 3 , R(2) is the weakest exothermic reaction and shows a decrease in T ad from 3475 • C at y = 1 to 2162 • C at y = 5. In spite of the dilution effect of pre-added Al 2 O 3 on combustion, R(3) intended for the synthesis of MoSi 2 -MgAl 2 O 4 composites is still very energetic with T ad ranging from 3840 • C to 2745 • C. Figure 1 indicates that R(3) has the highest T ad except for the case of z = 1. According to the analysis of combustion exothermicity, R(1) and R(3) were conducted in this study with the experimental variables of x = 2-5 and z = 2-5, respectively, and R(2) with y = 1-4 was carried out. Reactions with x = 1 and z = 1 were avoided, since the resulting combustion was often violent enough to melt down the powder compact and led to incomplete phase conversion. radial crack length measured from the center of the indentation.

Combustion Exothermicity of Reactions
Calculated values of Tad of R(1), R(2) and R(3) as a function of their resp chiometric coefficients are presented in Figure 1 in order to evaluate combust micity. A significant decrease in Tad with increasing silicide content is obse three synthesis reactions, mainly because the formation of molybdenum silici less exothermic than the metallothermic reduction of MoO3. As revealed in F value of Tad associated with the formation of Mo3Si-MgAl2O4 composites fr creases considerably from 3964 °C to 2415 °C as the coefficient x increases fro account of a large heat capacity for Mo5Si3, R (2) is the weakest exothermic r shows a decrease in Tad from 3475 °C at y = 1 to 2162 °C at y = 5. In spite of effect of pre-added Al2O3 on combustion, R(3) intended for the synthesi MgAl2O4 composites is still very energetic with Tad ranging from 3840 °C to 274 1 indicates that R(3) has the highest Tad except for the case of z = 1. According ysis of combustion exothermicity, R(1) and R(3) were conducted in this stu experimental variables of x = 2-5 and z = 2-5, respectively, and R(2) with carried out. Reactions with x = 1 and z = 1 were avoided, since the resulting was often violent enough to melt down the powder compact and led to incom conversion.

Combustion Wave Velocity and Temperature
A typical sequence of recorded combustion images from R(1) with x = 3 i in Figure 2, showing a stable and self-sustaining combustion process. A disti tion front allowed the propagation velocity to be determined. Variations of

Combustion Wave Velocity and Temperature
A typical sequence of recorded combustion images from R(1) with x = 3 is illustrated in Figure 2, showing a stable and self-sustaining combustion process. A distinct combustion front allowed the propagation velocity to be determined. Variations of combustion wave velocities of R(1), R(2) and R(3) with the molar ratio of silicide to MgAl 2 O 4 are presented in Figure 3. A declining trend consistent with the adiabatic combustion temperature was observed. This can be explained by the fact that the combustion wave propagation rate is essentially governed by layer-by-layer heat transfer from the thin combustion zone to the unreacted region, and therefore, is subject to the reaction front temperature. Specifically, Figure 3 points out a decrease in V f from 5.9 to 2.9 mm/s for R(1) with x from 2 to 5. For the similar range of stoichiometry of z = 2-5, R(3) has a faster combustion wave with V f ranging from 6.7 to 4.3 mm/s. On the other hand, the combustion front of R(2) has a slower speed and its V f decreases from 5.9 mm/s at y = 1 to 2.7 mm/s at y = 4. ent followed by a rapid cooling rate is characteristic of the temperature profil reaction. The highest value is considered as the combustion front temperature parison of Tc among three synthesis reactions in Figure 4a indicates that R(3) h est Tc of 1637 °C (z = 2), R(2) has the lowest 1442 °C (y = 2), and R(1) is in-betw °C (x = 2). A similar ranking of Tc can be seen in Figure 4b, which is associa synthesis of composites with a molar ratio silicide/MgAl2O4 equal to 4. Whe with Tc shown in Figure 4a, lower values of Tc = 1330 °C, 1103 °C and 1470 °C a in Figure 4b for R(1), R(2) and R(3), respectively. This confirms the decrease exothermicity with an increasing fraction of silicide formed in the composite.  parison of Tc among three synthesis reactions in Figure 4a indicates that R(3) h est Tc of 1637 °C (z = 2), R(2) has the lowest 1442 °C (y = 2), and R(1) is in-betw °C (x = 2). A similar ranking of Tc can be seen in Figure 4b, which is associat synthesis of composites with a molar ratio silicide/MgAl2O4 equal to 4. When with Tc shown in Figure 4a, lower values of Tc = 1330 °C, 1103 °C and 1470 °C a in Figure 4b for R(1), R(2) and R(3), respectively. This confirms the decrease exothermicity with an increasing fraction of silicide formed in the composite.   Figure 4a,b depict combustion temperature profiles measured from R(1), R(2) and R(3) under equal stoichiometric coefficients of 2 and 4, respectively. A steep rising gradient followed by a rapid cooling rate is characteristic of the temperature profile of the SHS reaction. The highest value is considered as the combustion front temperature (T c ). A comparison of T c among three synthesis reactions in Figure 4a indicates that R(3) has the highest T c of 1637 • C (z = 2), R(2) has the lowest 1442 • C (y = 2), and R(1) is in-between at 1574 • C (x = 2). A similar ranking of T c can be seen in Figure 4b, which is associated with the synthesis of composites with a molar ratio silicide/MgAl 2 O 4 equal to 4. When compared with T c shown in Figure 4a, lower values of T c = 1330 • C, 1103 • C and 1470 • C are observed in Figure 4b for R(1), R(2) and R(3), respectively. This confirms the decrease in reaction exothermicity with an increasing fraction of silicide formed in the composite.

Composition and Microstructure Analyses of SHS-derived Products
The XRD spectrum graphs of final products synthesized from R(1) with x = 2 and 4 are plotted in Figure 5a,b, respectively. Besides MgAl2O4, two silicide compounds were detected with Mo3Si the dominant and Mo5Si3 the minor. Because of the presence of Mo5Si3, there was a small amount of elemental Mo left in the end product. It should be noted that the production of MgAl2O4 justifies a combination reaction between in situ formed Al2O3 and MgO from the metallothermic reduction of MoO3 by dual reductants. Phase constituents associated with the products of R(2) are identified in Figure 6a,b, indicative of the Mo5Si3-MgAl2O4 composites with a trivial amount of Mo3Si. Because Mo5Si3 has a homogeneity range from 37.5 to 40 at% Si [20], no remnant Si was found in the Mo5Si3-MgAl2O4 products even containing some Mo3Si. Figure 7a,b shows the XRD spectra of the MoSi2-MgAl2O4 composites produced from R(3) with z = 2 and 4, respectively. It should be pointed out that MoSi2 formed from R(3) is α-MoSi2 (the low-temperature phase). This was due to the fact that the reaction temperature of R(3) was below 1900 °C [22], the phase transition temperature from α-MoSi2 to the high-temperature phase of β-MoSi2. As revealed in Figure 7a,b, there are small amounts of Mo5Si3 and Si in the as-synthesized MoSi2-MgAl2O4 composites.
When compared with the work of Zaki et al. [20], MgAl2O4 composites with MoSi2 and Mo5Si3 were produced from MoO3, SiO2, Al and MgO powder mixtures by self-sustaining combustion. They indicated the presence of small amounts of Mo5Si3, Al2SiO5 and free Si in the synthesized MgAl2O4-MoSi2 composites. The impurity Al2SiO5 was formed via a combination reaction of Al2O3 with SiO2. Moreover, the increase in MgO led to the formation of the other impurity Mg2SiO4 which was produced from the reaction between MgO and SiO2. Therefore, it is believed that the formation of Al2SiO5 and Mg2SiO4 could be due to incomplete reduction of SiO2, since these two phases were not found in the products of the present study. On the other hand, Zaki et al. [20] obtained MgAl2O4-Mo5Si3 composites without impurities and secondary silicides, on account of a larger heat release from combustion and a lesser amount of SiO2 contained in the sample.
In the work of Radishevskaya et al. [21], MgO and Al2O3 were added into a combustible mixture composed of Al, Mg(NO3)2⋅H2O and NaCl to produce MgAl2O4 through the SHS scheme. Results showed that the pre-added MgO and Al2O3 failed to be fully combined into MgAl2O4 unless mechanical activation of initial components in a planetary mill was conducted. In contrast, MgO and Al2O3 were not detected in the final composites of R(1), R(2) and R(3). This could be because these two precursors of MgAl2O4 were in situ produced from metallothermic reduction reactions in the present study.

Composition and Microstructure Analyses of SHS-Derived Products
The XRD spectrum graphs of final products synthesized from R(1) with x = 2 and 4 are plotted in Figure 5a Figure 7a,b shows the XRD spectra of the MoSi 2 -MgAl 2 O 4 composites produced from R(3) with z = 2 and 4, respectively. It should be pointed out that MoSi 2 formed from R(3) is α-MoSi 2 (the low-temperature phase). This was due to the fact that the reaction temperature of R(3) was below 1900 • C [22], the phase transition temperature from α-MoSi 2 to the high-temperature phase of β-MoSi 2 . As revealed in Figure 7a  When compared with the work of Zaki et al. [20], MgAl 2 O 4 composites with MoSi 2 and Mo 5 Si 3 were produced from MoO 3 , SiO 2 , Al and MgO powder mixtures by selfsustaining combustion. They indicated the presence of small amounts of Mo 5 Si 3 , Al 2 SiO 5 and free Si in the synthesized MgAl 2 O 4 -MoSi 2 composites. The impurity Al 2 SiO 5 was formed via a combination reaction of Al 2 O 3 with SiO 2 . Moreover, the increase in MgO led to the formation of the other impurity Mg 2 SiO 4 which was produced from the reaction between MgO and SiO 2 . Therefore, it is believed that the formation of Al 2 SiO 5 and Mg 2 SiO 4 could be due to incomplete reduction of SiO 2 , since these two phases were not found in the products of the present study. On the other hand, Zaki et al. [20] obtained MgAl 2 O 4 -Mo 5 Si 3 composites without impurities and secondary silicides, on account of a larger heat release from combustion and a lesser amount of SiO 2 contained in the sample.
In the work of Radishevskaya et al. [21], MgO and Al 2 O 3 were added into a combustible mixture composed of Al, Mg(NO 3 ) 2 ·H 2 O and NaCl to produce MgAl 2 O 4 through the SHS scheme. Results showed that the pre-added MgO and Al 2 O 3 failed to be fully combined into MgAl 2 O 4 unless mechanical activation of initial components in a planetary mill was conducted. In contrast, MgO and Al 2 O 3 were not detected in the final composites of R(1), R(2) and R(3). This could be because these two precursors of MgAl 2 O 4 were in situ produced from metallothermic reduction reactions in the present study.
For the Mo 3 Si-MgAl 2 O 4 composite of R(1) with x = 3 illustrated in Figure 8, the SEM image shows the fracture surface microstructure and EDS spectra provide the atomic ratios of constitution elements.

Conclusions
The in situ fabrication of Mo 3 Si-, Mo 5 Si 3 -and MoSi 2 -MgAl 2 O 4 composites was investigated by the SHS process integrating metallothermic reduction of MoO 3 with combustion synthesis. Mg and Al were simultaneously used as dual reductants to produce MgO