Synthesis of Submicron-Sized TiB₂ Powders by Reaction of TiC, B₄C, and Ca in Molten CaCl₂

Abstract

Submicron-sized TiB₂ powders (300 nm–1 μm) were prepared by the reaction of TiC, B₄C, and Ca assisted by molten CaCl₂. The optimal reaction procedure (1200 °C and 25 wt.% CaCl₂ + 25 wt.% Ca) was obtained by exploring the effects of the boronization reaction temperature and the addition of an amount of CaCl₂. It was found that the introduction of CaCl₂ not only promoted the reaction but also effectively inhibited the volatilization of excess Ca. Furthermore, SEM images of the products showed that the morphology and particle size of TiB₂ were inherited from the carbothermal reduction product TiC, which was dominated by the “template/growth” mechanism. The process of the boronization reaction was that B atoms migrated from B₄C and replaced the C atoms in the lattice of TiC.

1. Introduction

Recently, diborides have attracted significant attention in research in a multitude of areas. Among them, titanium boride (TiB₂) has great potential in the fields of impact-resistant armor, cutting tools, conductive coatings, and cathode materials because of its excellent properties, including a low density (4.52 g/cm³), an outstanding elastic modulus (525 GPa), a high melting point (2980 °C), a high hardness (25–34 GPa), and remarkable electrical conductivities (6.25 × 10⁵ S/cm) [1,2,3]. The key to achieving the extensive applications of TiB₂ material is the synthesis of powders. Nowadays, several preparation methods have been proposed to fabricate TiB₂ powders, including carbo/borothermal reduction [4], self-propagating high-temperature synthesis (SHS) [5], the sol–gel method [6], and the molten salt-assisted method [7]. However, each method has its own unique limitations or shortcomings in the preparation process. For example, although the carbo/borothermal reduction method exhibits the advantages of low raw material costs and easy realization of large-scale production, it requires a relatively high temperature as well as a long reaction time, which inevitably leads to the grain growth of the TiB₂ powders [8]. In addition, due to the strongly exothermic process of SHS, the safety of the reaction system cannot be ignored [9]. Although the sol–gel method can effectively prepare ultra-fine and uniformly distributed powder materials, the complexity of the process and the effective removal of residual carbon from the product are still problems restricting the wide application of this method [10]. Therefore, exploring a more economical, efficient, and safe preparation method to synthesize fine-grained TiB₂ powders has important scientific and practical value.

Combining the economic advantage of the carbo/borothermal reduction method, a method to prepare TiB₂ powders based on the reaction of TiC, B₄C, and Ca was proposed in our previous work [11]. TiO₂ was first reacted with C to synthesize TiC through carbothermal reduction (Equation (1)), and then the prepared TiC was reacted with B₄C and excess Ca to produce the target product TiB₂ (Equation (2)). Finally, the by-product CaC₂ and the residual Ca could be efficiently removed by a subsequent leaching process. This method cleverly solves the volatility problem of B₂O₃ in the high-temperature process, which ensures the precise addition of boron. Additionally, the calculated heat released during the entire reaction process shows that this method is mild and controllable [12]. However, although the addition of excess Ca can promote the reaction, excess Ca volatilizes at high temperatures and deposits in the sealing area of the crucible as well as the wall of the furnace tube, which seriously impedes the continuous production process. Moreover, the excess calcium releases more heat during the leaching process and leads to the oxidation of the TiB₂ powder. Thus, reducing the added amount of Ca is very necessary.

$${\mathrm{TiO}}_2+3\mathrm{C}=\mathrm{TiC}+2\mathrm{CO}$$

(1)

$${4\mathrm{TiC}+2\mathrm{B}}_4{\mathrm{C}+3\mathrm{Ca}=4\mathrm{TiB}}_2{+3\mathrm{CaC}}_2$$

(2)

Molten salt can accelerate the mass transfer process between the reactants, reduce the synthesis temperature and particle size of the product, and even influence the morphology of the product [13,14,15]. In this work, a molten salt-assisted approach was adopted, with the specific strategy of replacing the half mass of the calcium (Ca) in the original reaction system with calcium chloride (CaCl₂). The effect of the introduction of CaCl₂ on the target chemical reaction and the product properties was discussed. Moreover, the mechanisms of the boronization reaction were also analyzed.

2. Materials and Methods

Commercially available TiO₂ (purity > 99%, with a particle size of 100 nm, HushiLaboratoryReagent Co., Ltd., Shanghai, China), carbon black (purity > 98.5%, with a particle size of 24 nm, Mitsubishi Chemical Co., Ltd., Tokyo, Japan), B₄C (purity > 99%, with a particle size of 2–4 μm, Aladdin Reagent Co., Ltd., Shanghai, China), and calcium (purity > 99.5%, with a particle size of 1–5 mm, BeijingXingRongYuanTech Co.,Ltd, Beijing, China) were used as starting materials, and CaCl₂ (HushiLaboratoryReagent Co., Ltd., Shanghai, China) was added to form a molten-salt medium.

First, TiO₂ and C were mixed in the agate mortar for 30 min with the stoichiometric ratio (Equation (1)). Then, the mixture was charged into the graphite crucible and placed into a vertical furnace. The temperature of the carbon reduction reaction was set to 1500 °C and held for 4 h. After the reaction was completed, it was naturally cooled to room temperature in the furnace to obtain the target product titanium carbide (TiC). Subsequently, TiC and B₄C were ball-milled for 4 h in a steel jar with steel balls, and ethanol was used as the medium. After milling, the mixed slurry was dried and mixed with Ca and CaCl₂, and the mass ratio of TiC + B₄C, Ca, and CaCl₂ is provided in

Table 1. The specific reaction parameters of the boronization reactions.

Sample No.	Temperature (°C)	Mass Ratio (wt.%)
		TiC + B4C (Mixture of TiC and B4C)	Ca	CaCl2
1000-0.5Ca-0.5CaCl2	1000	50	25	25
1100-0.5Ca-0.5CaCl2	1100	50	25	25
1200-0.5Ca-0.5CaCl2	1200	50	25	25
1200-0.5Ca-0CaCl2	1200	66	33	0
1200-0.5Ca-1CaCl2	1200	40	20	40
1200-0.5Ca-0.5CaCl2-1/2B	1200	50 (molar ratio of TiC and B4C = 4:1)	25	25

. To investigate the effect of temperature on the reaction, the specific mass ratio of Ca to CaCl₂ was selected to be 1:1, and the total mass of Ca and CaCl₂ was equal to the total mass of TiC and B₄C (1000-0.5Ca-0.5CaCl₂, 1100-0.5Ca-0.5CaCl₂, and 1200-0.5Ca-0.5CaCl₂). To investigate the effect of the added amount of CaCl₂, the addition amount of Ca was half the total mass of TiC and B₄C, while the content of CaCl₂ was changed from 0, 0.5, and 1 times the total mass of TiC and B₄C (1200-0.5Ca-0CaCl₂, 1200-0.5Ca-0.5CaCl₂, and 1200-0.5Ca-1CaCl₂). The system was heated from room temperature to 1000–1200 °C at a rate of 10 °C min⁻¹ and held for 4 h in an argon atmosphere. Afterward, the as-synthesized products were leached with deionized water and HCl to remove the CaCl₂, as well as the residual Ca and CaC₂. Finally, the products were repeatedly rinsed with deionized water followed by drying in a vacuum oven at 90 °C. The process for the preparation of TiB₂ powders is illustrated in Figure 1.

The phase compositions of the powders were characterized by X-ray diffraction analysis (XRD, TTR III, Rigaku Corporation, Tokyo, Japan, Cu-Kα radiation, λ = 1.54178 Å). The products were scanned in the angle range (2θ) from 10 to 90°, and the scanning rate was 10 °C min⁻¹. An oxygen and nitrogen hydrogen analyzer (EMGA-830, HORIBA, Kyoto, Japan) was utilized to measure the residual oxygen content. The morphologies of the powders were evaluated using scanning electron microscopy (FE-SEM, ZEISS SUPRA 55, Oberkochen, Germany) and a transmission electron microscope (TEM, JEM-2100, Tokyo, Japan) equipped with energy-dispersive X-ray spectroscopy (EDX). The lattice parameters and proportions of the products were acquired by using the Full Prof suite software package (v.5, GSAS). The particle size distributions of the products were elevated by a laser diffraction particle size analyzer. The structure relaxation calculations were performed with the Vienna ab initio simulation package (VASP). The generalized gradient approximation (GGA) in the form of Perdew–Burke–Ernzerhof (PBE) was adopted. The plane wave cut-off energy was taken as 450 eV, and the k-points were set to 5 × 5 × 3. Each atom’s force was minimized to below 0.01 eV/Å through the comprehensive optimization of all structures.

3. Results and Discussions

3.1. Carbothermal Reduction

As mentioned above, in the carbo/borothermal reduction reaction system, TiO₂ reacted directly with B₄C, which could produce an intermediate product (B₂O₃). It is worth noting that B₂O₃ is significantly volatile at high temperatures, leading to the loss of the boron source [16]. However, in this work, the pre-deoxygenation of the feedstock was achieved through a carbothermal reduction pretreatment step to reduce the evaporation of boron and consumption of Ca in the next step. Figure 2a and Figure 2b show the equilibrium product distribution of the carbothermal reduction and the Gibbs free energy changes of Equations (1) and (3)–(6), respectively. From a thermodynamic point of view, TiO₂ could be fully converted to TiC when the temperature was higher than 1350 °C. Additionally, the carbothermal reduction did not directly proceed as expected according to Equation (1), while following the order TiO₂ > Ti₃O₅ > TiC [17]. In order to ensure the reaction could be completed, the temperature of the carbothermal reduction was determined to be 1500 °C, and the SEM images of the synthesized TiC powder at 1500 °C are shown in Figure 2c and Figure 2d. The presence of the significant agglomeration of TiC particles can be clearly observed in Figure 2c, which is quantified in the laser particle size analysis results in Figure 2g. The particle size of the prepared TiC agglomerates was mainly distributed in the range of 6–10 µm, with a D (50) and D (90) of 6.63 and 21.0 µm, respectively. However, the particle size of the individual TiC particles was mainly distributed in the range of 300–800 nm. Additionally, although no diffraction peaks of impurities were found in the XRD pattern (Figure 2f), the presence of residual carbon in localized areas is marked in the circle in Figure 2d. Meanwhile, the oxygen content of the prepared TiC powder was 3.132 wt.%, which may have originated from the oxide layer of the TiC surface and trace residual oxides. The residual carbon could be removed by the reaction of Ca and C, and the residual oxides could also be converted into TiB₂ during the boronization step [12].

$${3\mathrm{TiO}}_2{+\mathrm{C}=\mathrm{Ti}}_3{\mathrm{O}}_5+\mathrm{CO}$$

(3)

$${2\mathrm{Ti}}_3{\mathrm{O}}_5{+\mathrm{C}=3\mathrm{Ti}}_2{\mathrm{O}}_3+\mathrm{CO}$$

(4)

$${\mathrm{Ti}}_3{\mathrm{O}}_5+8\mathrm{C}=3\mathrm{TiC}+5\mathrm{CO}$$

(5)

$${\mathrm{Ti}}_2{\mathrm{O}}_3+5\mathrm{C}=2\mathrm{TiC}+3\mathrm{CO}$$

(6)

3.2. Effects of Temperature and CaCl₂ (Promotion of Thermodynamics as Well as Kinetics)

In this process, the carbothermal reduction was followed by a boronization reaction between the resulting product TiC and boron carbide. The effects of the temperature and the amount of added CaCl₂ on the synthesized TiB₂ powder will be discussed. The XRD patterns of the samples synthesized at different temperatures are shown in Figure 3a, where the addition amounts of CaCl₂ and Ca were both 25 wt.%. At 1000 °C, the XRD showed that TiC was the major phase, while TiB₂ and CaB₆ were minor phases, indicating that the boronization reaction was difficult to occur at this temperature. The presence of CaB₆ resulted from the reaction of Ca with B₄C (Equation (7)). No characteristic peak of B₄C was detected due to its weak diffraction peak intensity [18]. When the temperature was increased to 1100 °C, the main phase was changed to TiB₂, and diffraction peaks of TiC were also observed. The Rietveld-refined XRD results of the 1100-0.5Ca-0.5CaCl₂ products (Figure 3c) show that 77.2 wt.% TiB₂ was formed, and the residual content of TiC was 22.8 wt.%. The absence of CaB₆ may be due to the reaction between CaB₆ and TiC, producing TiB₂ and CaC₂, which was verified in a previous work [12]. With the temperature increase to 1200 °C, single-phase TiB₂ was obtained, indicating that the boronization reaction was completed. Thus, 1200 °C was necessary to ensure the reaction was completed. In addition, the XRD patterns of the products prepared at 1200 °C in the cases of different addition amounts of CaCl₂ are shown in Figure 3b. When no CaCl₂ was added, 80.89 wt.% TiB₂ was formed, but 19.11 wt.% TiC was also retained, as shown in Figure 3d. Increasing the addition amount of CaCl₂ to 25 wt.% and 40 wt.% produced no diffraction peaks of impurities except TiB₂. Thus, the addition of CaCl₂ is important to promote the boronization reaction. Although doubling the addition amount of Ca can achieve the same effect, it can also lead to significant calcium volatilization [12]. Figure 4 shows the saturated vapor pressure of Ca between 200 °C and 1200 °C based on Equations (8) and (9).

$${7\mathrm{Ca}+6\mathrm{B}}_4{\mathrm{C}=4\mathrm{CaB}}_6{+3\mathrm{CaC}}_2$$

(7)

$$\mathrm{Ca}\left(\mathrm{l}\right)=\mathrm{Ca}\left(\mathrm{g}\right)$$

(8)

$$\Delta G=-\mathrm{R}T\ln {\displaystyle \frac{p}{p*}}$$

(9)

where ΔG is the change in the Gibbs free energy of Equation (8); R is the molar gas constant (8.314); T is the temperature; p is the saturated vapor pressure of Ca; and p* is the standard atmospheric pressure. It can be observed that when the temperature was higher than the melting point of Ca (842 °C), the saturated vapor pressure of Ca rose sharply, and the excess Ca transformed into calcium vapor. Despite Ca being able to transfer via its vapor state, it was difficult to avoid its volatilization at high temperatures, and the volatilized calcium deposited on the lid of the sealed crucible, as shown in Figure 4. Deposited calcium on the crucible or furnace wall leads to an obstacle to the continuous production process and a safety risk. However, when 25 wt.% CaCl₂ was added, the volatilization of Ca was greatly reduced, and the deposited Ca was almost invisible on the sealed crucible. In fact, metal can react readily with various negatively charged ions in the molten salt, leading to the dissolution of the metal [19]. Suzuki et al. reported that Ca could be supplied as vapor and saturate in the molten CaCl₂ (about 6 mol% Ca at 1000 °C) [20]. The dissolution of Ca in CaCl₂ reduces the activity of Ca and inhibits its volatilization. Additionally, the introduction of CaCl₂ does not introduce new waste in the subsequent leaching process. In conclusion, CaCl₂ played a crucial role in this work and was not only able to provide a molten liquid-phase environment for chemical reactions, thereby effectively facilitating the transfer and exchange of substances, but also had the ability to dissolve part of the calcium, which significantly reduced the volatility of Ca at high temperatures.

3.3. Morphology and Particle Size of Prepared TiB₂ Powder

Figure 5 shows the SEM images of the reactants and products of the boronization reaction. Among them, Figure 5a and Figure 5b show the micro-morphologies of TiC and B₄C, respectively. Figure 5c presents the distributions of TiC and B₄C after ball-milling, where larger B₄C particles were surrounded by smaller TiC particles. Then, 25 wt.% Ca and 25 wt.% CaCl₂ were added manually due to the significantly different particle sizes. Subsequently, the morphological characteristics and particle size distribution of the resulting products after the boronization reaction at 1200 °C and the corresponding leaching step are shown in Figure 5d. TiB₂ had an irregular shape, with particle sizes mostly between 500 nm and 1 μm, as further demonstrated by the enlarged view of the agglomerates in Figure 5e. However, after increasing the CaCl₂ addition to 40 wt.%, the particle size and morphology did not change greatly, and it was found that the particle size prepared by this method was very close to that of the TiC. According to previous investigations, the molten-salt synthesis process always follows two mechanisms—the “template/growth” mechanism and the “dissolution/precipitation” mechanism—the selection of which is highly dependent on the solubility of the reactants in the molten-salt liquid-phase medium [21]. Liu et al. synthesized ultrafine TiB₂ powders by using TiO₂ powder, B powders, and NaCl/KCl. They found that TiB₂ would precipitate from the molten salt to nucleate and grow when its concentration reached the supersaturation condition in the molten-salt medium, which was dominated by the “dissolution/precipitation” mechanism [7]. Considering that both TiC and B₄C are covalent compounds with little solubility in CaCl₂, the TiB₂ products retain the morphology/size of the reactant TiC. The “template/growth” mechanism was expected to dominate the molten-salt route in this work [19], and the specific reaction mechanism will be discussed in the next section.

The representative microstructure of the TiB₂ particles obtained with 25 wt.% Ca and 25 wt.% CaCl₂ at 1200 °C was also characterized by TEM, as shown in Figure 6. The TEM images (Figure 6a) exhibit that the prepared TiB₂ powder was severely agglomerated. Additionally, the size of the single particles ranged from 300 nm to 1 μm, which is consistent with the results of the SEM images. The atomic number of B is small, which produces a low accuracy of identification. Thus, the EDS of Ti and B can only semi-quantitatively indicate their distributions. An oxide layer with a thickness of about 3 nm was clearly displayed on the surfaces of the particles (Figure 6c). The presence of an oxide layer on TiB₂ particles is a common phenomenon, which can reduce the specific surface energy of the powder and hinder the diffusion of atoms and, finally, affect the sintering densification process [7]. Additionally, the lattice spacing can be clearly observed in the HERTRM image of Figure 6d and was confirmed to be 0.2 nm and 0.261 nm, corresponding to the (101) planes and (100) planes of TiB₂, respectively. The selected-area electron diffraction (SAED) pattern shows a series of clear and well-organized diffraction spots, which confirms that the TiB₂ powder was well-crystalized.

3.4. Mechanism of Boronization Reaction

In this work, the boronization reaction can be described as a process in which B atoms leave B₄C through a mixture of molten-salt CaCl₂ and Ca and diffuse to the surface of TiC to produce TiB₂. TiC acts as a reactive template during the boronization process, and the morphology and particle size of TiB₂ is inherited from TiC. Interestingly, as shown in Figure 7a, several microcracks appeared in the products of 1100-0.5Ca-0.5CaCl₂, and the formation mechanism of these microcracks may have originated from the tensile stress effect of boron during its diffusion penetration into the TiC matrix. To explain this phenomenon, the microstress of TiB₂-1100-0.5Ca-0.5CaCl₂ and TiC were calculated using the Williamson–Hall method [22,23]:

$$\beta \mathrm{c}os\theta ={\displaystyle \frac{0.9\lambda }{D}}+\epsilon \left(4\sin \theta \right)$$

(10)

where D is the grain size of the particles, and $\epsilon$ is the internal strain.

The microstress of the TiC and 1100-0.5Ca-0.5CaCl₂ powders were calculated to be 2.04 × 10⁻⁶ and 2.34 × 10⁻⁶, respectively. It is obvious that the substitution of boron for C produced internal stress in the TiC powders. Furthermore, to verify the actual source of these microstresses, the first-principles calculations using the density functional theory (DFT) were performed. According to the report by Vekables et al., the only remaining “trace” concentration of boron in TiC was able to induce the precipitation of TiB₂ in the direction of the (111) crystal plane of the TiC lattice [24]. Thus, a model of a supercell containing 24 atoms was designed, consisting of three closely spaced layers. Each layer was composed of Ti and C atoms arranged according to the (111) crystal plane orientation. Such a design was intended to simulate the penetration of boron into TiC and to observe its possible effects on the structure of the TiC. Generally, there are two different strategies for doping boron atoms into TiC supercells. The first strategy is that all three B atoms directly replace the original C atoms in the supercell, as shown in Figure 7c(I). The second strategy is that two of the B atoms take the place of C atoms, while the remaining third B atom is specially placed in an interstitial position corresponding to the TiB₂ lattice, as shown in Figure 7c(II) [25,26]. The entry of boron atoms into the supercell causes lattice relaxation, and the simulation results of this process are displayed in Figure 7b. In the first case, after six iterations, three C atoms were replaced by three B atoms. It is observed that the equilibrium surface spacing between neighboring (111) Ti layers increased by 8% from 0.2498 nm to 0.2531 nm. In the second case, after eight iterations, the increase in the equilibrium surface spacing between the (111) Ti layers was more significant, reaching 10.5%, which may imply that the B atom at the interstitial position had a more drastic effect on the lattice structure of TiC. As a result, the increased spacing between the TiC (111) crystal planes promoted the diffusion of B and the formation of TiB₂. The spacing of TiB₂ was 28% larger than that of the TiC grain, resulting in large tensile stress and the cracking of the TiC grains, as well as the formation of new TiB₂ grains. Furthermore, the atomic radius of B was 85 pm, which was larger than that of C (77 pm); meanwhile, the molar volume of TiB₂ (15.41 cm³/mol) increased by about 26% relative to the molar volume of TiC (12.22 cm³/mol). A significant increase in the molar volume typically results in a volume expansion of the TiC lattice during the process of the boronization reaction.

Moreover, the effect of the varying amounts of the added B₄C on the composition of the final product is also explored in this section. Figure 8a clearly demonstrates the differences in the XRD patterns of the products obtained after the addition of half-stoichiometric (1/2-stoichiometric) versus full-stoichiometric B₄C, respectively. It is observed that the added amount of B₄C did not reach the full-stoichiometric ratio. In addition to the expected TiB₂ and TiC phases, the presence of diffraction peaks of an unknown phase was detected, which provides a new perspective for understanding the phase transformation mechanism in the boron penetration process. The binary TiC-TiB₂ diagram obtained from Factsage 8.3 using the database of SpMCBN [27] shows that Ti₃B₄ was stabilized as an intermediate product in the transformation of TiC into TiB₂ when the temperature was less than 1160 °C, and a further increase in temperature could promote the transformation of Ti₃B₄ into TiB₂. However, Ti₃B₄ (57 at.% B) is a peritectic product between orthorhombic and metal liquid TiB₂ at 2180 °C [28,29,30], which is difficult to generate under the presented conditions. Hence, in this work, we prefer to consider the observed phase as a prototypical TiB₂ structure originating from the TiC transformation process rather than Ti₃B₄. Figure 8c shows an SEM image of the 1200-0.5Ca-0.5CaCl₂-1/2B particles, and the morphology and particle size are consistent with the pure TiB₂ particles prepared in this work. However, the EDX spectrum of point A demonstrates the simultaneous presence of the elements Ti, C, and B, with a mass ratio of Ti, B, and C of 56:30:14. This result shows that the C atoms in TiC were not completely replaced by B, and a prototypical structure close to TiB₂ was formed. Although the lattice spacing of the edge particle (Figure 8(e₂)) was determined to be 0.313 nm, which corresponded to the (001) crystal plane spacing of TiB₂, the EDS mapping, as shown in Figure 8f, shows that the TiC particles contained traces of B and edge TiB₂ particles contained residual C. These phenomena suggest the diffusion of B into TiC to substitute for C.

4. Conclusions

In summary, using TiC, B₄C, Ca, and CaCl₂ as raw materials, submicron-sized TiB₂ powders were synthesized via the molten-salt method at 1200 °C. The conclusions are given below:

(1)

The introduction of CaCl₂ not only promoted the reaction but also reduced the volatilization of excess Ca.

(2)

The particle size and morphology of TiB₂ were inherited from TiC based on the “template/growth” mechanism, and the particle size of the prepared TiB₂ ranged from 300 nm to 1 μm.

(3)

Boronization was a process in which B atoms from B₄C diffused into the TiC lattice and gradually replaced the C atoms.