Preparation and Characterization of Magnesia-Based Powder Added with Transition Metal

The solid-phase reaction method for preparing forsterite (Mg 2 SiO 4 ) using MgO and SiO 2 powders has the disadvantages of high reaction temperature, long reaction time, and inhomogeneous reaction depending on the particle size of MgO. Therefore, MgO-based powders with a high reactivity were synthesized using a coprecipitation method with substitutional elements (Mn or Ni), and the effects of processing parameters on synthesizing MgO-based binary composition powders were investigated through the particle characteristics. The crystal structure continuously changed with the contents of the substitutional element, showing the same trend as the atomistic simulation results. The MgO-based powders showed higher reactivity than the conventional MgO powder, which could be confirmed in the particle characteristics, such as particle size and crystallinity, obtained in a short reaction time, and at a relatively low temperature. The optimum composition ratio in the binary composition powder for forming the Mg 2 SiO 4 depended on the type of substitutional element, and the reaction mechanism was identified based on the particle characteristics.


Introduction
Ceramic powders are synthesized using various methods, such as solid state reaction, solgel reaction, hydrothermal synthesis, and coprecipitation [1][2][3][4]. The solid state reaction method is the most widely used for producing polycrystalline ceramic powders from mixtures of solid starting materials [5,6]. There is no waste disposal problem in the solid state reaction method because there is no need for solvent in the reaction [5], while there is a disadvantage of expensive construction and its maintenance because it should be conducted at a high temperature of 1000-1500 C [6]. In addition, this process requires a long reaction time, and the uniformity in the reaction depends on the particle size and distribution [7].
Forsterite (Mg2SiO4) is employed as an insulator in tubes, substrate for resistors, substrate for integrated circuits, and an implant material, owing to its excellent properties, such as excellent insulation performance, low dielectric loss, high mechanical strength, excellent thermal stability, and large coefficient of thermal expansion [7][8][9][10]. Therefore, the various synthetic methods mentioned above have been studied for the successful synthesis of Mg2SiO4 [7][8][9][10][11][12][13][14][15][16]. Among these methods, the solid state reaction from mixtures of magnesia (MgO) and silica (SiO2) powders has been generally utilized for the simple synthesis of Mg2SiO4 without post-solvent and environmental problems. However, the solid state reaction between powders requires a high temperature of over 1500 C and longer reaction time during the reaction compared with other methods [1,5,6]. To improve the issues of high process temperature and long reaction time in the solid state reaction for preparing Mg2SiO4, various studies have been undertaken to increase the reactivity of MgO and SiO2 powders [12][13][14][15][16]. Ta-Wui Cheng et al. reported that the reaction temperature in the solid state method could be reduced to 1500 C by adding serpentine [12]. Lin Cheng et al. increased the surface activity by employing nanosized MgO and SiO2 powders prepared using a high-energy milling method, and reduced the heat-treatment temperature to 850 C [13]. Liugang Chen et al. synthesized Mg2SiO4 at 850 C using MgO precursors such as magnesium hydroxide (Mg(OH)2) or magnesium carbonate (MgCO3) [15,16].
MgO, which can be utilized for synthesizing Mg2SiO4 by the solid state reaction, has been prepared by several methods, such as coprecipitation, electrodeposition, sol-gel technique, hydrothermal, solvothermal, preparation using a bubbling setup, and microwave-assisted synthesis [17]. Among the various methods, the coprecipitation technique is a simple and cost-effective method to produce ceramic powder, including MgO, by adding a chemical precipitant to a cation-containing solution or by changing temperature and/or pressure. It is possible to synthesize powder under various conditions by adjusting the concentration, pH, mixing and stirring speed, and solution temperature [18][19][20][21]. In addition, coprecipitation can be an efficient manufacturing procedure to produce MgO-based binary powders with uniform properties across their surface, thereby controlling the physical and chemical properties of MgO-based binary powders and assisting the determination of the structure-activity relationships [22].
The three oxides, MgO, nickel oxide (NiO), and manganese oxide (MnO), play similar roles in minerals because the cations have the same charge and similar radii. Each combines to form an olivine structure (structure of Mg2SiO4), and undoubtedly there is complete miscibility between these three olivine members at high temperatures [23,24]. NiO and MnO are of particular interest in the solid solution research on MgO, having the same rock salt crystal structure as MgO, and all three compounds have approximately the same unit cell dimensions [25], confirming that the addition of MnO and NiO to MgO lowers the melting point of the solid solution and tends to decrease in temperature as the addition amount increases. Based on this concept, it is necessary to develop a new composition and/or a new MgO-based powder for preparing Mg2SiO4 at a lower temperature with a short reaction time and a uniform reactivity. Therefore, in this study, new MgO-based binary composition powders were prepared using the coprecipitation method, in which it is easy to control the various parameters mentioned above. The Ni and Mn as substitutional elements were employed to prepare the binary powders. The reactivity of MgO and binary powders with SiO2 was compared in preparing Mg2SiO4 through particle characteristics. The optimal composition ratio in the binary powders was induced with the type of substitutional element, and its reaction mechanism for forming Mg2SiO4 was proposed and discussed based on the particle characteristics, such as particle size and crystallinity. was used as a reaction catalyst. Each starting material was dissolved in distilled water to 0.5 M concentration and then mixed with mole ratios as shown in Table 1. In this study, two binary powders of Mg1-xNixO and Mg1-xMnxO were prepared with the composition ratios shown in Table 1.

Preparation of MgO-based binary powder
In addition, the single-composition powders of MgO, NiO, and MnO were synthesized as reference powders. Each precursor solution was mixed by stirring at 300 rpm for 30 min at room temperature, and then the pH of each mixed solution was adjusted to 9, 11, and 13 by dropwise addition of NaOH. In addition, the pH was fixed at 11 with controlling the concentration of NaOH as 0.1, 0.5, and 1 M to investigate the effect of catalytic concentration on the nuclei growth rate. Again, stirring at 300 rpm for 1 h was conducted for uniform dispersion of each metal hydroxide converted from each metal chloride (each precursor). The reacted and precipitated powder was washed five times with distilled water, dried at 100 C for 24 h, and then ground. Each powder was heat-treated at 600, 800, 1000, and 1200 C in an inert atmosphere for the Mg-Ni binary powder and in a reducing atmosphere for the Mg-Mn binary powder [26]. To investigate the effects of the pH value and the concentration of catalyst on the particle size and morphology, the heat treatment was conducted using a In addition, to evaluate the reactivity of the synthesized powder with SiO2, the pellet-type sample was prepared by mixing each powder (MgO, Mg-Ni binary powder, Mg-Mn binary powder) and SiO2 with a mole ratio of 2:1, followed by ball milling for 24 h, and then heat treatment was performed at 1200 C in a reducing atmosphere. The heat-treatment atmosphere for the reactivity test was decided based on a real application condition. The MgO-based powders for the reactivity test were prepared at 800 C in an inert atmosphere for the Mg-Ni binary powder and in a reducing atmosphere for the Mg-Mn binary powder, while the MgO powder heat-treated in the inert atmosphere was employed in the reactivity test.

Characterization
The phase analysis for the synthesized MgO-based binary powders and the Mg2SiO4 samples obtained from the reactivity test were characterized using X-ray diffraction (XRD; MiniFlex II; Rigaku, Tokyo, Japan) with Cu K radiation ( = 1.54060 Å) in the 2 range from 20 to 80 with a scan speed of 2/min. The average crystallite size (D) was calculated from the full width at half maximum (FWHM) of the XRD peak using the Scherrer equation

Atomistic simulation for crystal structural property in MgO-based powder
Two types of atomistic simulations were conducted. The first is a molecular dynamics (MD) model to simulate Young's modulus of MgO and doped MgO (i.e., Mg1-xNixO and Mg1-xMnxO, where x = 0.03, 0.06, and 0.12) using a simulated tensile test. The MD model uses the pairwise potential with partial charges proposed by Pedone et al. [28] and is implemented with customized codes in the LAMMPS MD simulation package [29]. This potential includes a Morse potential with an additional repulsive term and long-range Coulomb interaction. Compared with the commonly used BKS potential, a pairwise potential has the advantage of more stable convergence at a wide range of temperature.

Powder characteristics with the species and content of the substitutional element
The XRD patterns of the Mg1-xNixO (x = 0-1) powders synthesized at pH 11 are shown in Fig. 2 before and after heat treatment in an inert atmosphere (Ar atmosphere). As the temperature increased, the width of the peak narrowed, indicating that the phase was stable [27,30]. As the amount of Ni in the Mg1-xNixO composition was increased, the peak was slightly shifted to a higher angle. The peak should be shifted according to Bragg's law [31,32] because the ion radius of Ni (69 pm) is smaller than that of Mg (72 pm). Through this result, it was confirmed that the Mg-Ni binary powder was completely synthesized as a solid solution until the composition of Mg0.5Ni0.5O. Therefore, the Mg0.7Ni0.3O powder with an intermediate composition was used for investigating the particle size and morphology. increasing temperature, showing a bigger particle size in the NiO powder, while the particle size of Mg0.7Ni0.3O powder did not grown much with the temperature, compared with those of the MgO and NiO powders. In particular, the particle size of MgO powder was too small at 600 C, which was employed as the reference powder. The particle morphology of the MgO, Mg0.7Ni0.3O, and NiO particles became polygonal at 1000, 1200, and 800 C, respectively.
Based on the XRD analysis and particle size observation, the most suitable heat-treatment temperature was considered to be 800 C for preparing the Mg-Ni-based binary powder, showing a similar particle size of 0.09  0.04 m and crystallinity with the MgO powder of 0.08  0.04 m.
The XRD patterns of the Mg1-xMnxO (x = 0-1) powders synthesized at pH 11 are shown in Fig. 4 before and after heat treatment in a reducing atmosphere of 95%Ar-5%H2. Because Mn(OH)2 forms Mn3O4 during the drying process in an ambient atmosphere, the OHshould be removed during the heat treatment for synthesizing the Mg-Mn binary and MnO powders [32]. In the Mg-Mn binary system, the XRD peak was shifted to a lower angle with increasing Mn content because the ionic radius of the Mn(II) (83 pm) ion is larger than that of the Mg ion, according to Bragg's law [33,34]. When the amount of Mn in the Mg1-xMnxO composition was increased, an amorphous phase was detected in the as-synthesized powders as shown in Fig. 4(A). This phenomenon was also observed at a relatively low temperature of 600 C (Fig. 4(B)). The difference in the ionic radius between Mg and Mn is larger than that The grain-growth behaviors were different from those powders shown in Fig. 3 [34,35]. Based on the XRD analysis and particle size observation, the most suitable heat-treatment temperature was also considered to be 800 C for preparing the Mg-Mn-based binary powder, even though the particle size of the Mg0.7Mn0.3O powder (0.08  0.04 m) was slightly larger than that of the MgO powder (0.06  0.03 m) at 800 C in the reducing atmosphere, but the same as the MgO powder of 0.08  0.04 m prepared in the inert atmosphere.The crystallite size calculated from the XRD patterns using Scherrer's formula [36] is shown in Table 2 for each powder synthesized at 600, 800, and 1000 C. When the MgO powder was heat-treated in the inert atmosphere, there was almost no difference in the crystallite size. Similarly, the crystallinity was not significantly different for the binary and NiO powders synthesized in the inert atmosphere.
When the heat treatment was performed in the reducing atmosphere, the crystallite sizes for the MgO and binary powders were much smaller than those in the inert atmosphere till 800 C, and then increased significantly at 1000 C, except for the binary powder. The smaller crystallite size at the relatively low temperatures in the reducing atmosphere is seen as an effect of hydrogen [37].

Effects of pH and catalyst concentration
The effects of pH in the mixed solution on the powder characteristics, such as crystallinity, particle size, and particle morphology, were investigated for both binary powders with the Mg0.7M0.3O composition after heat treatment at 1200 C in an inert atmosphere for the Mg-Ni binary powder and in a reducing atmosphere for Mg-Mn binary powder. In the XRD analysis, there was no difference in crystallinity with the pH value, as shown in Fig. 6, without any peak shift in both compositions, indicating that both powders were completely synthesized as the solid solution with a binary composition. However, the particle size showed a tendency to decrease with increasing pH values in the microstructure (Fig. 7), showing a polygonal morphology in all cases. In the Mg-Ni binary powder, the particle size was not much changed after showing a small change between pH 9 and pH 11, while the Mg-Mn binary powder showed a transition point between pH 11 and pH 13 in the particle size.
The particle size of the Mg-Mn binary powders was larger than that of the Mg-Ni binary powders because the activation energy was low, as mentioned above [34,35]. In addition, the isoelectric points of Mg(OH)2, Ni(OH)2, and Mn3O4 would affect its nucleation behavior [38], which are pH 1012, pH 78, and pH 56, respectively, including the atmosphere for the heat treatment. Therefore, the agglomeration phenomenon would be hindered at pH 13 above the isoelectric point, resulting in refining the particle size in both cases.
The crystallinity, particle size, and particle morphology with changing the concentration of the catalyst to 0.1, 0.5, and 1.0 M were investigated for both binary powders of Mg0.7M0.3O composition. At that time, the pH value was fixed as 11. As shown in Fig. 8, there was no significant difference in the crystallinity with the concentration of catalyst.
However, as the concentration of catalyst increased, the particle size of both Mg1-xMxO powders became fine, showing the effect of the heat-treatment atmosphere (Fig. 9). The Mg0.7Mn0.3O showed a bigger grain size than Mg0.7Ni0.3O at all pH values. These phenomena shown in the particle size seem to be due to the faster growth rate than the nucleation rate at a relatively low concentration of catalyst [39]. In addition, it could be seen that the heattreatment atmosphere had a great influence on the grain growth behavior of the Mg0.7M0.3O powders synthesized by the coprecipitation method.

Simulation results with the species and content of the substitutional element
From the MD simulations, the calculated Young's modulus of MgO is 176 GPa, which is between the literature values of 130 GPa [40] and 300 GPa using a nanoindentation test [41].
In  [43]. Because the ionic radius difference between Mg 2+ and Ni 2+ is smaller than that between Mg 2+ and Mn 2+ , it is expected that the chemical strain in the Ni-doped MgO is less than that in the Mn-doped MgO.
Therefore, the reduction in Young's modulus in the Ni-doped MgO is smaller than the case of Mn.
Regarding the simulations for XRD patterns, they were performed with the content of the substitutional element as 1.5, 3, 6, and 12% in the Mg1-xNixO structure, which is shown in Fig. 10(A). The diffraction patterns of relative intensity against the angle of diffraction show a peak backward shift, as well as an increased intensity in the doped structures accordingly.
The Mg1-xMnxO structures with the content of the substitutional element as 1.5, 3, 6, and 12% were also obtained as shown in Fig. 10(B). The peaks were shifted to a greater degree than that for the Mg1-xNixO structures as shown in Fig. 10(A). It is shown that the ion radius of Mn is shifted to the left because that of Mn is larger than that of Mg, and Ni is shifted to the right because the atomic radius of Ni is smaller than that of Mg, as mentioned in Section 3.1.
The XRD patterns obtained with the simulation correspond quite well to those with experimental results shown in Figs. 2 and 4. Therefore, it can be confirmed that Mn and Ni, which are substitutional elements, were synthesized as a solid solution.

Reactivity with the species and content of the substitutional element
The XRD results before and after the reaction between Mg1-xMxO powders and SiO2 are shown in Figs. 11 and 12, with the species and content of the substitutional element. The MgO and amorphous SiO2 peaks were detected before heat treatment. When only MgO was used, Mg2SiO4 and unreacted SiO2 (cristobalite) were detected (Figs. 11(B-1) and 12(B-1)).
In the reaction between the Mg-Ni binary powders and SiO2, unreacted Ni was detected with Mg2SiO4 and unreacted SiO2 (Fig. 11(B-3)). As the content of Ni was decreased, the peak intensity of Ni was reduced without unreacted SiO2. However, in the case of Mg-Mn binary oxides, only Mg2SiO4 was synthesized by completely reacting with SiO2 ( Fig. 12(B-2) and 12(B-3)), independent of the content of Mn. This is because the electronegativity of Mn is higher than that of Ni, which is due to the strength of the ionic bonding force [44], which can be seen to correspond to the above simulation results. Therefore, MnO, which has a strong binding force, was not reduced to Mn, but NiO is considered to be precipitated as Ni because the bond with oxygen ions is broken.
The microstructure and element analysis results after the reactivity test for each composition are shown in Fig. 13. The Mg-Ni binary powders showed better reactivity than the MgO powder, enhancing the reactivity with increasing content of Ni. However, as the content of Ni was increased, the Ni element was precipitated as shown in Fig. 13(B) and 13(C), which was verified as Ni element in the element analysis and Fig. 11. On the other hand, in the Mg-Mn binary powders, as the content of Mn increased, the reaction layer around SiO2 became thick. As can be seen from the XRD results in Fig. 12, it was confirmed that Mn was not precipitated and fully dissolved in MgO, showing a core-shell structure.
Through this, it was found that the substitution by Mn element in MgO powder could make a stable binary composition and reaction layer rather than the Ni element for synthesizing Mg2SiO4.
Through the results above, it was possible to synthesize large quantities of powder with a uniform composition by the coprecipitation synthesis method, which is economically cheaper than the existing sol-gel process [45,46]. In addition, in the synthesis of Mg2SiO4, the MgObased binary powder with a substitutional element doped into MgO would react completely to obtain a uniform Mg2SiO4. Furthermore, the nonuniform reactivity, which is a problem of the conventional solid-phase reaction method [47], could be improved by controlling processing parameters in the coprecipitation process.

Conclusions
Mg1-xMxO (M = Ni and Mn) binary powders with improved reactivity were prepared by doping substitutional metal ions into MgO using the coprecipitation method. As the content of the substitutional element was increased, the diffraction patterns were shifted to higher and lower angles continuously in the Mg1-xNixO and Mg1-xMnxO powders, respectively, showing a greater degree in the Mg1-xMnxO powder, owing to the difference of ion radius. The most suitable heat-treatment temperature was considered to be 800 C for preparing the binary powder in both cases, based on the particle size and crystallinity, even though the particle size was dependent on the species of substitutional element and the Mn-doped binary powder showed a larger particle size than the Ni-doped one. The crystallinity and crystal size increased as the heat-treatment temperature increased, especially in the case of Mg1-xMnxO binary composition, showing an effect of the heat-treatment atmosphere. The pH value of slurry and the concentration of catalyst did not significantly affect crystallinity, but the particle size tended to be finer as the pH value increased, indicating again that the heat-     Table 1.   Table 1.